CHAPTER 11

Functional organization of motoneuron

pool and its inputs
Department of Physiology, Harvard Medical School,
ELWOOD HENNEMAN
Boston, Massachusetts
Department of Physiology, Duke University Medical
LORNE M. MENDELL
Center, Durham, North Carolina

CHAPTER CONTENTS Distribution of la excitation to motoneuron pools

Percentage of motoneurons receiving terminals from
Morphological Considerations single la-fibers
Columnar arrangement of motoneuron pool Factors influencing percentage of motoneurons receiving
Dimensions of a-motoneurons and distribution of cell size homonymous projections
Scaling of motoneurons Comparison of projections to homonymous and heteronymous
Initial segment of a-motor axons motoneurons
Axon collaterals of a-motoneurons Correlations between morphology and function
Recurrent inhibitory feedback from motoneurons Latency of EPSPs
Direct synaptic interconnections between spinal motoneurons Other examples of divergence in inputs to motoneurons
Species of motoneurons la-projections to motoneurons controlling other parts of
Terminals of motoneurons in muscle the body
Morphology of neuromuscular junctions Group II input from secondary endings in muscle spindles
Matching the properties of motoneurons and the muscle fibers Inhibitory inputs to motoneurons
they supply la inhibitory interneurons
Concluding comments Renshaw cells
Firing Patterns of Individual Motoneurons and Motor Units Group Ib input from Golgi tendon organs
Functional significance of size of motoneurons Monosynaptic input from descending pathways
Measurement of total output of motoneuron pools Topographic factors governing development of connections of
Critical firing levels of motoneurons la-fibers to motoneurons
Relation of critical firing level to axon diameter and Concluding comments
motoneuron size Nonuniformity of Motoneurons
Effects of inhibitory inputs on critical firing level and rank order Early classification of tonic and phasic types of motoneurons
during repetitive firing Significance of nonuniformity of muscle fibers
Recruitment of motor units in humans Motoneuron properties independent of size
Evidence regarding alternative patterns of recruitment Differential responses of motoneurons to injected currents
Evidence regarding voluntary selective control of motor units Influence of muscle on developing and mature motoneurons
Size principle in other species Evidence from human disease
Modulation of firing rate Concluding comments
Organization of Input to Motoneuron Pools How Size of Motoneurons Determines Their Susceptibility
Anatomical studies to Discharge
The motoneuron Properties of motoneurons that influence susceptibility
Anatomy of la-branches to motoneurons to discharge
la-synapses on motoneurons Role of input in determining susceptibility to discharge
Techniques used to study EPSPs elicited by impulses in single Some Principles Underlying Organization of Motoneuron Pools
afferent fibers How sensitivity in gradation of tension is achieved
EPSPs recorded from single motoneurons Basis for relation between motoneuron size and the force its
EPSPs recorded from populations of motoneurons motor unit develops
Amplitudes of EPSPs elicited by impulses in single fibers Actual sensitivity in grading muscular tension
Variability of EPSP amplitudes Mathematical derivation of a "principle of maximum grading
Factors responsible for variability in EPSP amplitudes sensitivity"
Boutons of la-fibers on motoneurons Recruitment order and minimum energy principle
Physiological analysis of location of boutons Collective action of motoneuron pool: role of input
Efficacy of synapses on different parts of motoneuron The size principle in la and group II sensory fibers
Physiology of la-terminals How does the central nervous system use the motoneuron pool?

423
424 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

ALL OF THE activity in the nervous system that influ- impart understanding rather than a mere accumula-
ences movement converges ultimately on motoneu- tion of facts.
rons. For this reason, Sherrington called motoneurons
the "final common path" to muscle. The cells and
axons that compose this common path are by no MORPHOLOGICAL CONSIDERATIONS
means uniform in their properties. They differ system-
atically from each other in size and other properties The purpose of this section is to provide a morpho-
and form a functional ensemble whose collective prop- logical basis for the physiological findings that are
erties are far more extensive than those of any single described in later parts of this chapter. Particular
motoneuron. The organization of this population of emphasis is placed on the dimensions of motoneurons,
cells derives from the intrinsic properties of the mo- the uniform scaling of the different parts of the neuron,
toneurons and from the inputs they receive. the distribution of cell sizes, and the functional impli-
In order to discuss this organization in a logical cations of these factors.
manner, this review is divided into six sections.
1. An anatomical description of the motoneuron
Columnar Arrangement of Motoneuron Pool
pool and certain aspects of the morphology of individ-
ual motoneurons is given. This account is limited to The population of motoneurons innervating a par-
information that seems relevant to a discussion of ticular muscle is called a "pool" (78). The members of
functional organization. a pool are not readily identifiable with ordinary stain-
2. The role of a motoneuron pool is to translate a ing techniques, because they appear to be intermingled
large, heterogeneous inflow of signals from certain with other motoneurons in the gray matter of the
peripheral structures and from many parts of the ventral horn. By sectioning muscle nerves Romanes
nervous system into a much smaller and simpler out- (302) showed that the chromatolyzed motoneurons
put that will produce precisely controlled tensions in supplying hindlimb muscles in the cat were disposed
a particular muscle. The characteristics of this output, in longitudinal columns within lamina IX. A recently
as judged by the firing patterns of individual motoneu- developed technique (230) has made it possible to
rons and motor units, are described. From the body of label the motoneurons supplying a particular muscle
information available from studies on animals and more clearly than with chromatolysis, so that they
humans a set of simple rules is derived that governs stand out distinctly from other cells. When the enzyme
the behavior of the motoneuron pool as a whole and horseradish peroxidase (HRP) is injected into a mus-
the muscle it supplies. cle, it is taken up by the motor axons in the muscle
3. Physiological studies on afferent systems that and carried rapidly by retrograde axoplasmic flow
send impulses directly or indirectly to the motoneuron back to their cell bodies, where it appears, after an
pool are described and an attempt is made to infer appropriate chemical reaction, as brown, intracyto-
some general principles that govern the organization plasmic granules. If the intramuscular injection of
of such inputs. HRP is well dispersed, so that all the motor terminals
4. The fourth section reviews evidence regarding the are exposed to it, a high percentage of the motoneu-
nonuniformity of motoneurons. The chief purpose is rons is found to contain the HRP reaction product.
to determine whether the differences between moto- Intracellular transport ofthe enzyme makes it possible
neurons in a single pool are sufficiently basic to classify to define all the dendritic and axonal projections of
them into different types or species. Other information motoneurons in some cases.
essential to a proper appreciation of pool organization Figure 1 illustrates the locations of labeled cells in
is also included. an experiment (47) in which HRP was injected into
5. Some of the factors underlying the differences in the left medial gastrocnemius (MG) and the right
the susceptibility of motoneurons to discharge are soleus (SOL) muscles of a cat. After reconstruction of
discussed. A new hypothesis regarding the role of the histological material, it was apparent that the
terminal arborizations of afferent fibers on motoneu- motoneurons innervating each of these muscles
rons is advanced to explain the size principle of re- formed a column of cells parallel with the long axis of
cruitment. Recent experimental evidence supporting the spinal cord in the lateral part of the ventral horn.
this hypothesis is presented. Throughout most of their courses the two pools ov~r­
6. Lastly, some theoretical and mathematical treat- lap to some extent. This arrangement permits close
ments of principles underlying the functional organi- apposition of pools supplying related muscles, which
zation of the motoneuron pool are called to the may facilitate the organization of input common to
reader's attention, although space does not permit them and of interconnections between them. Both a-
exten.sive discussion of them. motoneurons, which innervate the ordinary, large ex-
The field with which this chapter deals is changing trafusal muscle fibers, and y-motoneurons, which sup-
rapidly. We wish to emphasize that this account does ply the small intrafusal muscle fibers, were labeled
not provide a comprehensive review of the world lit- with HRP. The latter cells (originally called "small"
erature in this area. It is selective and attempts to motoneurons) were distributed randomly throughout
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 425

both cell columns. Since they provide the motor in- Scaling of Motoneurons
nervation for spindles located in the muscle supplied
Early anatomists were intrigued by the wide range
by the a-motoneurons, they should be regarded as
special components of the same pool. A third type of of sizes and by the geometry of the neurons they found
cell, the ,8-motoneuron, which innervates both extra- in the nervous system. Deiters (86) was apparently the
and intrafusal fibers, has recently been described by first to recognize that the diameter of an axon is
Emonet-Denand et al. (108). proportional to the size of its cell body. Cajal (291)
also showed that the size of a neuron is related to the
Dimensions of a-Motoneurons and diameter of its axis cylinder and, further, to the num-
Distribution of Cell Size ber and thickness of its collaterals and terminals. As
examples of large cells with thick axons he cited spinal
Physiological studies reveal that the functions of
motoneurons, giant corpuscles of the torpedo, Golgi
motoneurons are highly correlated with their size. In
particular, the relations between the size of a moto- cells of the cerebellum, horizontal cells of the retina,
neuron and 1) its susceptibility to discharge by excit- and, as a case of special interest, the Mauthner cell. In
atory inputs, 2) the suppression of its firing by inhibi- contrast, he pointed out that small neurons, such as
tory inputs, and 3) the properties of the motor units it the granular cells of the cerebellum and fascia dentata,
supplies prompted Henneman et al. (159) to conclude and the bipolar cells of the retina, all have thin axons,
that there was a "size principle" underlying these which give off few collaterals. In his chapter dealing
correlations. with dorsal root ganglia, Cajal points out that ganglion
The a-motoneurons are large cells with soma di- cells of large volume have thick axons and those of
ameters ranging from about 30 to 70 /Lm (339) and a small volume have thin axons (Figs. 157-159 in ref.
mean diameter of 48 /Lm, assuming they are spheres 291). Many other examples could be noted. The giant
(121). The distribution of their cell volumes (soma and cells that have been described in various species in-
dendrites) in the rhesus monkey is shown by the variably have large axons; small cells never do. In the
histogram in Figure 2, a rescaled version of one con- mammalian nervous system, there seem to be no strik-
structed from measurements of all of the motoneurons ing exceptions to this general rule.
in the seventh cervical segment (161). This highly The dimensions and proportions of spinal motoneu-
skewed distribution, with a preponderance of cells at rons have recently been scrutinized in more detail and
the smaller end of the scale, is characteristic of moto- with a greater variety of techniques than those of
neuron pools supplying large and small muscles in all other types of neurons because of the body of infor-
species. mation that clearly relates their functional character-
Using electrophysiological techniques, Clamann and istics to their size. After studying motoneurons physi-
Kukulka (62) found no evidence that motoneurons are ologically with intracellular electrodes that were filled
distributed spatially according to size within the pools with a dye to mark them, Barrett and Crill (13) then
supplying the medial gastrocnemius or plantaris mus- examined the labeled cells histologically. Using the
cles as Wyman (355) had suggested. Their study indi- conduction velocity of the axons as a measure of their
cates that any sufficiently large sample of motoneu- diameter (174), they found that axonal diameter cor-
rons obtained in a restricted region of the motoneuron related closely with the size of the cell body and the
pool or in the ventral root fIlaments issuing from it extent of its dendritic tree. By injecting horseradish
contains a wide range of sizes. peroxidase intracellularly (324) it is now possible to
Differences in the sizes of motoneurons allow for perform dimensional analyses of physiologically char-
considerable variation in the number of boutons on acterized neurons more precisely than ever before.
their surfaces. A summary of some early estimates of With this technique, Cullheim (80) has confirmed and
synaptic density was compiled by Illis (179). Conradi extended Barrett and Crill's findings by purely mor-
(71) made an extensive electron microscope examina- phological methods. The mean diameters of the cell
tion of cat motoneurons and Barrett and Crill (14) bodies of sciatic motoneurons injected with HRP
carried out a detailed study of the soma and dendritic ranged from 44 to 71 /Lm, which is in good agreement
trees of several motoneurons that had been injected with the results of electron microscope studies (71). In
with Procion dye. Combining the results of these two Figure 3 the diameters of a-motor axons at the initial
studies, Barrett (12) estimated the total number of segment and in the white matter are plotted against
synaptic endings at 20,000-50,000 per cell, depending the sizes of the parent cell bodies. In spite of the
on its size. scatter, due largely to representing cell size by the
The great differences in the surface area of moto- mean diameters in just one plane, significant (P <
neurons, and possibly in their geometry, may be the 0.001) positive correlations are present in both cases.
major cause for the wide range of input resistances In Figure 4 the axon conduction velocities are plotted
they exhibit (0.3-6.0 MO). The functional significance against cell body sizes, initial segment diameters, and
of input resistance in determining excitability and mean axon diameters in the white matter, with sig-
inhibitability is discussed in later parts of this chapter. nificant positive correlations in all three cases.
426 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

MG injection SOL injection

I
I

I
/
I
I
I . '.
(

.·..,,.
\
\
••'. AI - - - -
- --- ..
....::
.. .~.:...
~

~
:.'1 I ,
.. I,
I
.'.
,"...
..'~.,..:. ... B . .;. f' .
, ..... I
• ')' 0' 1
..~:
,.. :.. .
I
.A!

.......
I
, I
\

.........
I -' ':".
.,
..at.:
j
I '~"':' ~

!..,g,
c 1 _ .~.
..

.:~~ -'
......
• ••_ 1

...,.
. .,. I
·'. .
'101.

:,
.oil ..
I :'

o I ".,

"t I
If. E
,.
1

~
lSI

I
I
I
I

)
I mm

Cat MNP 5
-------------/ ---_/
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 427

Initial Segment of a-Motor Axons z

o
Physiological studies indicate that impulses are nor-
mally generated in the initial segment (IS) of a moto- ~
neuron's axon (75, 118). If other factors are equal, a :bs
Q.
given depolarization of the soma should cause a larger oQ.
flow of transmembrane current in a thin IS than in a
thick one. Thin initial segments may, therefore, have Z
lower thresholds than thick ones (183). Cullheim and ~10
Kellerth (82) have used intracellular injections of ;:)
horseradish peroxidase to delineate the intramedul- w
lary parts of the axons of sciatic a-motoneurons in the ~
to-
adult cat. The IS was invariably the narrowest part of
the motor axon. Its diameters ranged from 2.3 to 4.9 ~s
/lm (mean = 3.5 /lm) and its lengths varied from 20 to U.
35 /lm (mean = 26 /lm). No correlation was found o
between its diameter and length. When the diameters ~
of the IS were plotted against those of the same axon
in the white matter, a significant (P < 0.001) positive 2 4 6 8 10 ---- >12
correlation was found, as illustrated in Figure 5. In 4
view of these considerations, the dimensions and mem- MOTONEURON VOLUME x 10 ... 3
brane properties of the initial segment may playa role FIG. 2. Distribution of motoneuron sizes in the 7th cervical seg-
in determining the relative excitability of motoneu- ment of 6 normal rhesus monkeys. [Adapted from Hodes et al.
(161),]
rons. Their importance, relative to that of other fac-
tors, is, however, difficult to assess without further
information.
eral varied from 1 to 39 and the total number of such
branches from a single motor axon ranged from 0 to
Axon Collaterals of a-Motoneurons 69. A significant positive correlation exists between
The number and total extent of the axon collaterals the diameter of the first-order collateral and the num-
of a motoneuron are related to the size of the cell and ber of its end branches. This is another example of the
its functional properties. Cajal (291) found that a scaling principle in motoneurons.
correlation existed between the diameter of central Cullheim and Kellerth (81) have also used horserad-
axons and the size of their collateral trees. Cullheim ish peroxidase to examine the axon collaterals of dif-
and Kellerth's study (82) with horseradish peroxidase ferent types of motor units as classified by Burke et
shows that this finding applies to a-motoneurons. This al. (40,41). In this study, they were concerned chiefly
suggests that stronger physiological effects may be with "axon collateral swellings," discrete out-bulgings
exerted by recurrent collateral systems from large that usually are identifiable as synaptic boutons under
motor axons than from small ones. The number of the electron microscope. All motor units with fewer
first-order collaterals given off by a single axon ranged than 30 axon collateral swellings were found to be of
from 0 to 6. Most first-order collaterals give off 2-5 the slowly contracting, slowly fatiguing (S) type or the
branches of about equal size. The number of collateral soleus-S (SOL-S) type (see the chapter by Burke in
end branches originating from one first-order collat- this Handbook), while those with more than 80 swell-

FIG. 1. Reconstructions of medial gastrocnemius (MG) and soleus (SOL) nuclei from serial sagittal
sections in cat. Right, dorsal view of spinal cord outline (white matter-pia boundary in heavy lines)
on which are superimposed positions (dots) of MG (left hemicord) and SOL motoneuron cell bodies
(right hemicord). Boundaries between L 6 , L" and S, segments (identified by dorsal root entry zones),
are indicated by horizontal heavy lines; midline denoted by vertical dashed line. Dashed lines across
the cord denote levels (labeled A-E) at which reconstructions of cross sections were made. Left,
reconstructions of cross sections at levels A-E showing white matter-pia boundary in light lines and
gray-white matter boundary in heavy lines. Dashed lines at most lateral parts of cross sections
indicate estimated outline; the most lateral parts were lost in sectioning. All diagrams drawn on same
scale. Neurons indicated on each cross-section diagram are cells located within 300 JIm rostral and
caudal to that level. [From Burke et al. (47),]
428 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

A B
10

E 4
=l-
• . ,I, . . :::.
..' ...." ,.'
Ii: 3
w I ,
~ "
w
~ 2
o
':!:: 1

40 50 60 70 40 50 60 70
MEAN DIAMETER MEAN DIAMETER
OF CELL BODY ()Jm) OF CELL BODY Cum)

FIG. 3. A: diameters of sciatic a-motor axons of cat at initial segment (IS) are plotted against mean
diameters of their parent cell bodies (n = 68). Linear correlation coefficient (r) = +0.522 (P < 0.(01).
B: mean diameters of sciatic a-motor axons in white matter are plotted against mean diameters of
their parent cell bodies (n = 42). Linear correlation coefficient (r) = +0.516 (P < 0.(01). [From
Cullheim (80).]

A B
120 >- 120
~

u u
c
C; '00 w
g 100
:> :>

.5_ 80 " ~u;BO

:::- '" .:: E
':5E
::J - 50
u~
,5 60
z z
:) o
u u
40 ....
' --'-----"--'---:"":',....-.......---,,'::'- z 40 • !

Z
:) ~ 50 60 m 0 2 3 "
><
< MEAN DIAMETER ~ IS DIAME TER (/'''')
OF CELL BODY ()Jm)

>-
120
..... " ,"
u
..
'.
, , "
9 100
, ,
lJ.J
>
9",00
..... ,
.... ,

':5-S60
o
z
o
u 40 ! ! !

z 468 10
o><
<l
AXON OIAME TER IN
THE WHITE MATTER()Jm)

FIG. 4. Axon conduction velocities in cat are plotted in A against size of parent cell bodies (n =
24), in B against axon diameters at initial segment IS (n = 24), and in C against axon diameters in
white matter (n = 33). In A the linear correlation coefficient (r) = +0.673 (P < 0.(01); in B, r = +0.869
(P < 0.(01); and in C, r = +0.804 (P < 0.(01). [From Cullheirn (80).]

ings were exclusively of the fast, fatiguing type. The fication of these two types. The authors believe that
mean number of swellings found were as follows: the large variations in the axon collateral trees are at
FF = 98.1 ± 29.6; FR = 53.8 ± 16.5; S = 44.4 ± 14.1; least partly correlated with differences in the types of
SOL-S = 27.8 ± 13.1 (FF is fast, fatiguing; FR is fast, motor units they supply. Since different types of motor
fatigue resistant). These differences were statistically units vary in the size of their motoneurons, it is diffi-
significant. The axon collateral systems of the FF and cult to judge whether motoneuron size or motor-unit
FR motor units showed considerable differences, add- type correlates more closely with the morphology of
ing morphological support to the physiological classi- axon collateral trees.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 429

Recurrent Inhibitory Feedback 10

from Motoneurons
Control of a muscle demands various types of coor-
dinated activity of its motoneurons. Much of this
~ 2-8
a: a:
E

• •••• se
.....
o 0

..
W W
.. .. .-
coordination is produced by inputs that are shared by f-f-
o

a large percentage of the motoneuron pool, as de- wf- 6 o

L<l::
scribed in section ORGANIZATION OF INPUT TO MOTO- <l::L o

NEURON POOLS, p. 460. Another type of coordination,

00

o~4
still poorly understood, results from short-latency ZI
feedback mechanisms. Renshaw (297, 298) demon- 03
strated that each time a motoneuron discharged an ~w2
impulse a recurrent collateral of that axon transmitted I
f-
a corresponding signal that excited a special group of
interneurons located mainly in the ventral portion of 1 2 3 4 5
lamina VII near the motoneuron pool. These interneu-
rons, called Renshaw cells, send their axons back to IS DIAMETER Cum)
motoneurons in the pool of origin, where they exert FIG. 5. Relation between axonal diameters in spinal cord white
inhibitory effects. This inhibitory feedback is modu- matter and diameters of initial segments (IS) for 39 sciatic a-motor
lated by activity in dorsal root fibers and descending axons in cat. Linear correlation coefficient (r) = +0.698 (P < 0.0001).
systems. Its function, however, is not at all clear. [From Cullheim and Kellerth (82).1
Consistent with other findings that large motoneurons
exert greater effects than small ones, Ryall et al. (307) and evidence that this is transmitted electrically be-
reported that Renshaw cells are more strongly excited tween the somadendritic regions of the motoneuron
by collaterals of large motoneurons than of small ones. membranes was recently advanced by Magherini et al.
(248). Excitatory interactions between cat motoneu-
Direct Synaptic Interconnections rons that could be mediated by recurrent collaterals
Between Spinal Motoneurons have been described by Gogan et al. (124). However,
an unequivocal demonstration of the physiological
The possibility of even more direct interactions effect of activity in these collaterals has not yet been
between motoneurons via recurrent collaterals was made.
recognized many years ago. Cajal (291) reported that
some collaterals arborize extensively in the region of Species of Motoneurons
their motoneurons and suggested that they made syn-
aptic contacts with them. Later, collaterals terminat- For some time it was assumed that the motoneurons
ing not only in the Renshaw cell area but also among supplying a muscle were a homogenous population of
motoneurons were described by the ScheibeIs (309, cells differing only in size and in certain size-dependent
310). No wholly convincing demonstration of direct properties. Recent physiological studies (145, 146),
synaptic transmission from one motoneuron to an- however, indicate that there may be two or more
other has been made to date in mammals. Cullheim et species of a-motoneurons in a pool, differing in char-
al. (83), however, recently provided the best morpho- acteristics that are independent of size. Just as there
logical evidence that direct interconnections exist. By are distinct types of muscle fibers in a single mam-
injecting horseradish peroxidase into motoneurons of malian muscle, there may be different species of mo-
the eat's triceps surae he and his colleagues were able toneurons that innervate them and are responsible for
to trace some motor axons and recurrent collaterals to their contractile properties. However, no morphologi-
their terminations with both the light and electron cal basis for distinguishing different types of cells
microscope. In addition to the expected projection to within a pool, except by their size, has been discovered
the Renshaw cell area, axon collaterals were fre- to date. Studies on the distribution of enzyme activity
quently found terminating within the motor nuclei, in anterior horn cells have been carried out by Penny
where at least some of them made direct contact with and colleagues (279) in the hope of identifying different
triceps motoneurons. The presence of a synaptic com- species of motoneurons histochemically. Cytophoto-
plex at these junctions "satisfied the morphological metric measurements of the activities of five oxidative
criteria needed for postulating a genuine synaptic con- enzymes (succinate, malate, lactate, NAD+-linked iso-
nection." (83). citrate, and NADH dehydrogenases) have been made
Several reports of physiological interactions be- in the spinal cord of rabbits. The findings indicate a
tween motoneurons have appeared. Short-latency fa- unimodal distribution of the activities of all of these
cilitation of motoneurons by antidromic excitation of enzymes. Although the results do not offer a means of
adjacent motoneurons was described by Nelson (267). identifying different species of motoneurons, they are
Recurrent excitation of frog motoneurons by anti- consistent with the "constant proportion" hypothesis
dromic volleys was also reported by Grinnell (140), of Pette and co-workers (281) with regard to oxidative
430 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

enzyme activity. This was based on the finding that in maximal tetanic tensions varied from about 3 to 40 g
a number of tissues, including the brain, the activities (253); in the medial gastrocnemius they ranged from
of mitochondrial enzymes were always in constant about 0.5 to 120 g, a more than 200-fold difference. A
ratio to each other, although the absolute activities definite relation between axonal conduction velocity
varied from tissue to tissue. It should be emphasized (proportial to diameter) and maximal tetanic tension
that although there was a unimodal distribution of the was found in the deep and superficial lumbrical muscle
activities of all the enzymes studied, there was a very of the cat (6, 20) and in the soleus muscle of the cat
wide range of activities among anterior horn cells. (253). As illustrated by the three examples in Figure
For the present, therefore, the only evidence indi- 6 a linear relationship between conduction velocity
cating that there are different species of motoneurons and maximum tetanic tension was found in all 10
in a pool comes from physiological studies that are experiments on the soleus muscle. For other large limb
described in section FIRING PATTERNS OF INDIVIDUAL, muscles the evidence has been controversial (11, 262,
MOTONEURONS AND MOTOR UNITS, p. 435. Whether the 273, 283, 353). A recent paper emphasized the impor-
motoneurons innervating different types of muscles, tance of sample size in such studies (10). Jami and
such as the extraocular muscles, constitute different Petit (185), therefore, investigated four leg muscles
species is a question requiring information that is not (peroneus brevis, peroneus longus, soleus, and tibialis
yet available. anterior), examining 25%-50% of the motor unit pop-
ulation in each muscle. As Figure 7 illustrates, these
Terminals of Motoneurons in Muscle large samples showed unequivocally that there is a
demonstrable relation between the logarithm of the
Cajal (291) recognized that the number of terminals tetanic tensions developed by motor units and the
a motor axon gave off was proportional to the diameter conduction velocities of their motor axons. A total of
of its axon and, thus, to the size of its cell body. Eccles 453 units was studied to construct Figure 7. Axonal
and Sherrington (102) devoted considerable attention conduction velocities were grouped in classes of
to the branching of motor fibers and proposed that
"the larger fiber prepares early to form an extensive
motor unit," since division of large fibers begins far
proximal to the muscle. Thus the size of a motoneuron
is an important element in determining the capacity
of its motor unit to produce tension. However, other
factors such as the type of muscle fibers it supplies
also playa role.
The average "size" of a motor unit is determined by 50 60 70 80
dividing the number of fibers in a muscle by the
E 30 B
number of a-motor axons in the nerve supplying it. ......
bO
o
The extrinsic eye muscles contain as few as 3-6 fibers c:
o
per motor unit. In the soleus of the cat, Clark (64) 'iii 20

-
found an innervation ratio of 120: 1. The mean inner- c:
Q)

vation ratios vary widely in different muscles and may u

--
reach 1,000:1 or more in some large limb muscles. 'c 10
IV
Innervation ratio is usually correlated with delicacy of Q)

movement.
What is important for this discussion is the fact that E
:l 50 60 70 80
innervation ratios vary widely within a single muscle E
as a direct function of the conduction velocity and '~30 C
diameter of the motor axon. This became apparent in ~
the 1960s when a technique was developed that made
it possible to activate a single motor unit in an other- 20
wise quiescent muscle and to study its properties in
isolation (20,352). The axon of a single motor unit was 10
separated from others supplying the same muscle by
repeatedly subdividing ventral root fIlaments until a
thin strand was found that contained only one axon
supplying the muscle under investigation. The con- 50 60 70
duction velocity of this axon was obtained from mea- Conduction velocity (m /sec )
surements of conduction time and distance. When
6. Relation between maximum tetanic tension of individual
FIG.
single motor units were stimulated at various rates motor units in cat soleus muscle and conduction velocity of their
there were striking differences in the maximal tetanic axons in 3 different experiments. [From McPhedran, Wuerker, and
tensions they produced. In the soleus muscle, the Henneman (253).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 431

PERONEUS BREVIS PERONEUS LONGUS

100 100

I f t

:!
10
I t 10

j I f
c
~
j
..... f
VI
c
f
j
•1 .1
65 70 75 80 85 90 95 m/s 80 85 90 95 100 105 110 mlS

TIBIAliS ANTERIOR SOLEUS

100 100

I f
i tt
f

I It
10 10

I
:!
c
0
j
.....
VI
C 0

.1 .1
75 80 85 90 95 100 106 m/S 60 65 70 75 80 65 90 mlS

FIG. 7. Relation between logarithm of tetanic tensions developed by motor units and the conduc-
tion velocities of their motor axons in 4 leg muscles of the cat. Heavy dots on vertical bars indicate
mean value of tension ± SD. For peroneus longus and tibialis anterior, intervals corresponding to
conduction velocities under 85 mls contained too few motor units to allow valid calculation of mean
tension. Range of tensions developed by these motor units are represented by a vertical bar. Open
circle represents tension of a single unit. For soleus, 4 motor units innervated by axons, either faster
than 90 mls or slower than 60 m/s. are not represented. [From Jami and Petit (185).]

5 mls intervals. This study provides a firm basis for flexor digitorum longus, Olson and Swett (273)
the conclusion that axonal conduction velocity and grouped the 83 units developing less than 10 g each on
tetanic tension are correlated in large limb muscles. It an expanded scale, with intervals of 1 g. Even within
does not, however, explain whether the number or this restricted range, there was a clear tendency for
type of muscle fibers, or both, are responsible for the the number of units in each I-g interval to decrease in
correlation. It seems likely that within each type of number as their tensions increased. These histograms
motor unit there are units with numbers of muscle are typical of all muscles that have been studied in
fibers that vary systematically with the conduction this way, including such specialized muscles as the
velocity of the axon. tensor tympani (336) and those that move the eye.
The distributions of maximal tetanic tensions for The functional significance of this distribution pattern
soleus and medial gastrocnemius muscles in the cat is discussed later in this section (Matching the Prop-
are plotted in Figure 8. In both histograms the largest erties of Motoneurons and the Muscle Fibers They
number of motor units occurs in the O-lO-g range. In Supply, p. 433). It is sufficient now to point out that
general there is a progressive decrease in the number the abundance of small units makes possible precise,
of motor units in each higher decade. In a study of the finely graded control of muscle tension in the lower
432 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

30- This type of analysis cannot be applied to the gastroc-

Soleus (97 units)
nemius muscle, because it has three types of fibers
Avg. = 14.8 gm differing in size and contractile properties.
Only one study has been carried out on the anatomy
and innervation ratios of identified types of motor
t! 20- units (49). Direct morphological determination of in-

-
c:
:::J nervation ratios required complete reconstruction of
o the three-dimensional distribution of each motor
~
CIl unit's fibers within the whole medial gastrocnemius
.c
E muscle. Technical difficulties limited satisfactory re-
:::J constructions to two type FF units. The results indi-
Z
cated that the muscle unit territories extend through
a large fraction of the whole muscle volume, and
within this territory the density of fibers belonging to
a single unit is relatively low. At least 40-50 motor
10 20 30 40 units probably share anyone region within the medial
gastrocnemius. Burke and Tsairis (49) conclude that
Maximum tetanic tension (gm ) "an average muscle unit in the cat medial gastrocne-
mius contains between 400 and 800 muscle fibers ir-
A respective of physiological type." Unfortunately the
number of motor units satisfactorily reconstructed was
Medial gastrocnemius
too small to provide a direct morphological test of the
(103 units)
30- physiological findings described above; i.e., muscle
units that produce small tensions contain fewer fibers
VI
:!: than larger tension units. This conclusion was partic-
c: ularly compelling in the case of the homogenous soleus
.: 20- muscle of the cat. In harmony with this conclusion are
o
~
CIl
the results of Knott and colleagues (210), who showed
.c in studies of single motor units that there was a clear
E correlation between the surface area from which an
i 10·- EMG could be recorded, and the EMG amplitude and
the maximum tension the unit produced. This corre-
lation seems to indicate that strong motor units con-
tain more muscle fibers that are more widely dispersed
20 40 60 80 100 120 in the muscle than weaker units.
Maximum tetanic tension (gm )

B Morphology of Neuromuscular Junctions

FIG. 8. Distribution of maximal tetanic tensions of motor units
of cat's soleus (A) and medial gastrocnemius (B). [A, from Mc- This description of motoneurons would be incom-
Phedran, Wuerker, and Henneman (253); B from Wuerker, Mc- plete without including information on the junctions
Phedran, and Henneman (353).] they form with muscle fibers. Only certain aspects of
this topic that are germane to the subject of this
ranges, whereas the presence of very large units pro- chapter are described. Three types of neuromuscular
vides for the addition of large increments of tension junctions can be distinguished ultrastructurally (274).
during maximal efforts. The most obvious differences between them are the
Clark (64) found an innervation ratio of 120:1 for total sizes of the junctions and the amounts of surface
the soleus muscle. Correcting his estimate for the area they present for the discharge and reception of
small y-fibers to muscle spindles, which constitute transmitter. The axonal terminals on red fibers are
about 33% of the total motor outflow in muscle nerves, small and elliptical. The areas of contact they make
an average innervation ratio of 180:1 is obtained. This on a single fiber are relatively discrete and separate.
applies to the average motor unit which develops 14.8 The junctional folds, which the sarcoplasmic mem-
g of tetanic tension. By extrapolation the smallest brane of the muscle forms, are relatively shallow,
motor unit (3.2 g) and the largest unit (40.4 g) in this sparse, and irregular in their arrangement. Axoplasmic
sample consist of 39 and 491 muscle fibers, respec- vesicles are moderate in number and sarcoplasmic
tively. Since the cat's soleus is homogenous through- vesicles are rare. Axonal terminals on white fibers are
out, consisting entirely of p-fibers whose diameters longer and flatter and have a much greater area for
and histochemical appearance are extremely uniform release of acetylcholine. The postjunctional surface
(157), these estimates should be reasonably accurate. formed by the branching, closely spaced junctional
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 433

folds provides a vast receptive area for the transmitter. Matching the Properties of Motoneurons
At adjacent contacts on the same muscle fiber the and the Muscle Fibers They Supply
junctional folds may merge with one another, so that
synaptic contiguity is more continuous and wide- In the preceding sections it was pointed out that the
spread. Axoplasmic and sarcoplasmic vesicles are nu- dendrites, the axon, and the axonal terminals of mo-
merous. Intermediate fibers receive relatively large toneurons are all scaled in proportion to the size of
axonal terminals. Their junctional folds are the most the cell body. As a consequence of this scaling princi-
widely spaced and deepest of the three types. They ple, the postsynaptic effects a motoneuron produces
are relatively straight and unbranched. Axoplasmic at its various axonal terminals are also proportional to
vesicles are less numerous than in white fibers. Sar- the size of the cell. The most obvious of these size-
coplasmic vesicles are conspicuous near the deeper related effects is the capacity to produce tension. The
portions of the junctional folds. According to Padykula influence of a motoneuron on its muscle fibers, how-
and Gauthier (274), junctions on intermediate fibers ever, goes far beyond tension production. The pattern
are not just a gradation between the types on red and of contractile activity it induces in them causes them
white fibers but are a distinct third type. to develop a distinctive set of morphological, chemical,
On the presynaptic side of the myoneural junction and functional properties (308). Since the firing pat-
the dimensions and form of the axonal terminals are terns and mean daily activity of motoneurons vary
correlated with the caliber of the parent axon (271). systematically with cell size (and perhaps cell species)
On the postsynaptic side there is a positive correlation throughout a pool, the characteristics induced in their
between the size of the end plate and the muscle fiber muscle units vary correspondingly. The basic mecha-
diameter in mammals (5, 66, 69, 144, 271). Moreover, nisms by which contractile activity in a muscle induces
the distinctive features of the three types of junctions changes in its own properties are not understood, but
are correlated with the diameters of their muscle fibers they clearly enable the muscle to adapt to the amount
(274). The larger a muscle fiber, the smaller its input and pattern of usage it experiences. The result of this
resistance (198) and the greater the amount of trans- adaptation is that the contractile properties of motor
mitter required to depolarize it to its firing threshold. units are well matched to the activity patterns of their
Therefore, large white fibers require extensive synap- motoneurons.
tic surfaces. Kuno and colleagues (223) showed a pos- Figure 9 illustrates the distribution of some of the
itive correlation between the end-plate area and the properties of motor units that have been studied in
diameter of muscle fibers. They also demonstrated typical limb muscles. The histogram is a smoothed
that the number of quanta of transmitter released by version of the one shown in Figure 8 in which the
a presynaptic impulse was positively correlated with maximal tetanic tensions of gastrocnemius motor units
the size of the end plate. The greater number of are plotted against the number of units in each lO-g
synaptic vesicles in the axonal terminals on white tension range. As noted in subsection Dimensions of
fibers is presumably associated with release of more a-Motoneurons and Distribution of Cell Size, p. 425,
acetylcholine over a wider area. The resulting end- this skewed distribution, with a preponderance of units
plate potential is sufficient in amplitude and extent to producing small tensions, resembles the distribution
set up a propagated impulse in the large, low-resist- of cell volumes in Figure 2. The large number of units
ance muscle fiber. Smaller junctions with fewer vesi- at the left end of the histogram have axons of small
cles are adequate to ensure transmission to the more diameter and, thus, are innervated by a correspond-
easily excited red fibers. ingly large population of small motoneurons. The few
The pre- and postsynaptic differences in neuromus- motor units generating large tensions have thick axons
cular junctions that are associated with the three types and are supplied by a small number of large motoneu-
of muscle fibers are perhaps another indication that rons.
there may be three species of motoneurons supplying The largest motoneurons, which typically discharge
heterogeneous muscles. There is convincing evidence in brief, high-frequency bursts, supply motor units
that the activity pattern that motoneurons impress on with large, pale type A muscle fibers that contract
the muscle fibers they supply determines many of the rapidly and, therefore, require the fast rates of acti-
morphological, chemical, and functional properties of vation they receive from their motoneurons to produce
the muscle fibers (308). It is quite possible that these fused tetanic contractions. These motor units fatigue
activity patterns also determine the structural and quickly, sometimes in a few seconds, a characteristic
functional characteristics of neuromuscular junctions. that matches the brief duration of firing in their mo-
Although this has not been demonstrated, it seems toneurons. Their susceptibility to fatigue may be re-
likely because prolonged stimulation of a muscle nerve lated to lack of an abundant blood supply and to
can transform the muscle fibers it innervates from one insufficient mitochondria that contain the enzymes
type to another (308). It is probable that the neuro- necessary for aerobic metabolism. Instead, these pale
muscular junctions transmitting these impulse pat- fibers utilize glycogen, which is capable of supplying
terns are also transformed, since they must be adapted energy anaerobically at a rapid rate, but only for a
to the types of muscle fibers they supply. brief time. The glycogen is quickly depleted, and after
434 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

Distribution of

Motor Unit Properties

Number of units .. many. ........... few

Axon diameter. sma 11 .. . 1a rge
Number of termina1s .... few. ..many
Tetanic tension small... . 1arge
Speed of contraction .. slow or fast fast
Fatigue litt1e or none rapid
Metabolism .. aerobic anaerobic
Myoglobin . ...... p1entiful scanty
Glycogen . . little much
Mitochondria . high density very low density
Capillaries .. ...... rich supply few or none
Muscle fibers . small. red 1arge. pale
FIG. 9. Smoothed version of histogram in Figure 8B, illustrating distribution of some of the
properties of motor units in typical limb muscles. Height of columns indicates relative numbers of
motor units with the properties listed below. [From Henneman (154). Skeletal muscle: the servant of
the nervous system. In: Medical Physiology (14th ed.), edited by V. B. Mountcastle. St. Louis, MO:
The C. V. Mosby Co., 1980.]
a short period of maximal activity the tension output responding to the same influence, i.e.. the activity
drops to a low level and may take a long time to pattern of their motoneuron. It is perhaps not surpris-
recover. In contrast, the units that produce the small- ing, therefore, that the contractile properties in a
est tensions consist of thin, red fibers that are least motor unit are found in only certain combinations.
susceptible to fatigue. These units can maintain their The occurrence within one motor unit of a certain
maximal tensions almost indefinitely, as is often nec- capacity for producing tension, a particular contrac-
essary in postural activities. They are dependent on tion speed, a characteristic range of firing rates, a
oxidative metabolism for contractile energy. Their specific susceptibility to neuromuscular and contrac-
muscle fibers are richly supplied with capillaries and tile fatigue, a distinctive histochemical profIle, and
mitochondria. Their small diameters serve to reduce well-defined morphological properties probably rep-
the distance over which oxygen and other metabolites resents a set of correlated adaptations. The particular
must be transported from capillaries to mitochondria. combination of properties occurring in a unit undoubt-
All of these characteristics are well matched to the edly is the one with the greatest functional utility,
high level of daily activity in their motoneurons. As because all of the properties were determined by the
these examples illustrate, the distribution of firing
patterns in a motoneuron pool has its counterpart in same activity pattern. The advantages of combining
muscle where there is a matching distribution of con- certain properties will become clearer in section FIR-
tractile properties in the corresponding motor units ING PATTERNS OF INDIVIDUAL MOTONEURONS AND
(37, 45, 50). MOTOR UNITS, p. 435, when the actual discharge pat-
The adaptive mechanisms that determine what terns of individual motoneurons are described and the
properties develop in a motor unit are probably all operations of the pool as a whole are defined.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 435

Concluding Comments classify nerve fibers into three main types (A, B, C)
differing in size, and the first tentative attempts to
The morphological studies summarized in this sec- correlate fiber types with their functions were made.
tion indicate that motoneurons differ widely in size The voltages required to excite axons with external
and that all of their parts (dendrites, soma, initial electrodes increase as the diameters ofaxons decrease
segment, axon, collaterals, intramuscular terminals, (l09). It does not follow, however, that large neurons
and neuromuscular end plates) are scaled in propor- can be more easily discharged synaptically than small
tion to each other. Taken as a whole, the observations ones. The situation, in fact, appears to be just the
make an impressive array of evidence, providing an reverse, as Henneman (153) demonstrated in studies
anatomical basis for a size principle. on motoneurons. Reflexes were elicited in lumbar ven-
Although the morphological observations s~ggest tral roots by electrical stimulation of dorsal roots or
that in some respects large motoneurons are SImply large nerve trunks. When shocks of sufficient strength
expanded versions of small ones, the physiological were used, the response consisted of two phases (Fig.
findings of Harris and Henneman (145, 146) and others 10). The first phase was a brief, relatively synchronous
to be described later in section NONUNIFORMITY OF
MOTONEURONS, p. 485, indicate that motoneurons may
also differ in ways not related to their size. Cullheim
and Kellerth (81) report that the number of recurrent
collateral outbulgings (interpreted as synaptic termi-
nals) of a motoneuron correlate as well or better with
the type of muscle fibers innervated as with the axonal
diameter. In view of the anatomical evidence already
described and further physiological findings to be pre-
sented in the next section, it is probable that the
functional properties of motoneurons correlate not
only with their size but also with the neuromuscular
junctions they form and the types of muscle fibers
they supply.

FIRING PATTERNS OF INDIVIDUAL

MOTONEURONS AND MOTOR UNITS

Functional Significance of
Size of Motoneurons
Present concepts of the organization of the moto-
neuron pool can be traced to the interest of early
histologists in the significance of neuronal size and
shape. Ramon y Cajal (291) disposed of many of the
conjectures of his day by demonstrating that the di-
ameter of a nerve fiber is related directly to the size of
its cell body and by giving numerous examples of this
relation throughout the nervous system. His findings
have been repeatedly confirmed, in particular for mo-
toneurons, by electrical and anatomical studies. Func-
tional relations between the diameters ofaxons and
their electrical properties did not become apparent
until the introduction of the oscilloscope and the stud-
ies of Erlanger and Gasser in Electrical Signs of
Nervous Activity (109). In particular, Gasser (120)
showed that the amplitude of a nerve impulse recorded
externally from a peripheral nerve is related directly
to the diameter of the fiber in which it is conducted.
Thus, when impulses of different amplitudes are re-
corded from the same nerve, the largest signal signifies
FIG. 10. Reflex discharges recorded from a filament of 7th lum-
the firing of the largest fiber in the nerve and the
bar ventral root in response to stimulation of ipsilateral sciatic nerve
discharge of the biggest cell body. Smaller signals of cat. Numbers at right of each tracing indicate relative intensity
indicate firing in correspondingly smaller fibers and of stimulation. Initial deflection on left of each tracing is the early
cells. Electrical and anatomical criteria were used to discharge referred to in text. [From Henneman (153).]
436 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

discharge of short latency, followed by rhythmic fIring served discharges of motoneurons was a prolonged
sometimes lasting one second or more (the late phase fIring of internuncial cells, reaching an early peak of
of the well-known flexor reflex). To analyze the second intensity from which it declined smoothly to a resting
phase of the reflex in detail, monophasic recordings level. At its maximum the internuncial input was
were taken from fine filaments of ventral roots dis- apparently sufficient to cause repetitive fIring in a
sected under a low-power microscope. Individual im- wide range of cell sizes. As the excitatory activity
pulses of single motor fIbers that discharged repeti- subsided, it fell below the levels necessary to fIre the
tively could readily be distinguished in these records larger cells but remained adequate to discharge
by their amplitudes. As the stimulus applied to the smaller cells. Hence the patterns in Figure 10: short
ipsilateral sciatic nerve was increased, progressive trains of large spikes and longer trains of smaller
changes in the late response occurred. The top record impulses.
in Figure 10 shows that the stimuli reaching the mo- It should be emphasized that the relationship be-
toneurons were well above threshold for the early tween motoneuron size and susceptibility to discharge
reflex but were insufficient to evoke a late discharge. in these experiments was observed in recordings from
The three small spikes shown in this trace were prob- ventral root fIlaments where axons from several pools
ably impulses in y-motoneurons. Increase in the stim- of flexor motoneurons were intermingled. Undoubt-
ulus strength then caused a single a-motoneuron to edly, the observed fIring originated in different pools
discharge three times. Further increases in stimulus of flexor motoneurons (322). Nevertheless, the rela-
intensity caused progressively longer trains of im- tionships between input intensity and the size of out-
pulses in this fIber. At a higher stimulus intensity a put signals held with few exceptions. This suggests
second unit, with spikes about twice the amplitude of that several pools of flexor motoneurons, whose activ-
the fIrst, made its appearance, thereafter discharging ity is closely related during withdrawal reflexes, may
two, four, and fIve times in response to stronger shocks. function like one large, common pool in response to
A third and still larger unit appeared with stronger certain inputs. In 1926 Creed and Sherrington (79)
shocks, fIring twice at this intensity and four times at studied the concurrent contractions of flexor muscles
higher intensity. Recruitment of progressively larger in the "flexion" reflex and concluded:
motoneurons with increase in stimulus intensity was From the point of view of the afferent nerve the mus-
observed regularly. Discharges of slightly smaller mo- cular entity executing the [flexor] reflex movement
toneurons with stronger shocks were occasionally ob- is ... not this or that muscle but a composite aggregate of
served in single tracings, but not as a regular occur- motor units scattered through and forming parts of a
rence. In general, units that differed most in spike number of separate and even distant muscles. This com-
posite collection is given functional solidarity and homo-
amplitude also varied most in the strength of stimu- geneity by the likeness of threshold and latency obtaining
lation required to discharge them. Even small grada- among its units in regard to the afferent nerve bringing
tions in the spike amplitude of motoneurons, however, them into action. An increment in stimulus does not
were associated with measureable differences in their break this homogeneity but merely brings to it a certain
reflex threshold. addition again composed of units mutually similar in
From these observations it appeared that motoneu- threshold and latency although made up from muscles
rons may be graded according to their susceptibility anatomically separate.
to synaptic discharge in a flexor reflex. At the upper These studies suggested that orderly recruitment of
end of the scale are the largest cells, requiring the motoneurons according to their size might be a general
most intense stimulation. Smaller cells require corre- phenomenon, not limited to flexor reflexes. To inves-
spondingly less intense stimulation. At the lower end tigate this possibility a more comprehensive series of
of the scale are the y-motoneurons. In most recordings experiments was carried out using a variety of inputs.
from ventral root fIlaments these cells can be seen A convenient means of evoking a reflex discharge of
fIring steadily without externally applied stimulation. motoneurons is to stretch an extensor (antigravity)
The tendency to be continuously active is perhaps an muscle in a decerebrate cat. In such preparations the
indication that these small cells are so susceptible to normal response to stretch is exaggerated and the
excitation that the spontaneous activity of the spinal slightest elongation of the muscle results in the dis-
cord is suffIcient to keep them fIring. charge of motoneurons and the development of reflex
Although the discharge patterns recorded from dif- tension (233). The reflex is highly specifIc, the response
ferent ventral root fIlaments varied, certain features being limited to the motoneurons of the muscle that
were regularly observed. The most notable were 1) is stretched or its closest synergists.
recruitment of larger units and lengthening of dis- Figure 11 illustrates the results of stretching the
charge trains with stronger shocks; 2) regularity in the triceps surae muscle and recording from a fIlament of
rhythm of discharges of each motoneuron, with pro- the seventh lumbar ventral root. The tension applied
gressive decrease in their rate of fIring with increasing to the tendon of the muscle is measured by the sepa-
train length; and 3) an inverse relation between spike ration of the upper beams and the ventral root dis-
height and train length. These characteristics suggest charge appears in the lowest tracings. With the muscle
that the excitatory process responsible for the ob- completely relaxed the only activity usually recorded
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 437

STRETCH RELEASE electrically, whether the responses were elicited mono-

synaptically or polysynaptically, or whether the mo-
toneurons were flexor or extensor. In a further series
1 of experiments (325) trains of electrical stimuli were
. ttl''' I' JI'
~6 also applied to brain stem motor areas, cerebellum,
basal ganglia, and motor cortex of the cat. Regardless
of the site of stimulation, motoneurons whose axons
yielded the smallest impulses in ventral root filaments
were discharged at the lowest thresholds and other
units were recruited in order of increasing size. By
comparing each responsive unit with all of the others
in the same fIlament, a total of 396 pairs of motoneu-
rons was examined. In 86% of comparisons the smaller
3~~8 unit had the lower threshold, in 4% the thresholds
were indistinguishable, and in 10% the larger unit had
the lower threshold. The use of electrical stimulation
with widely spaced electrodes must have resulted in
9 discharges from several pools of motoneurons. The
,! , r', S ", "
low incidence of exceptions to orderly recruitment
suggests that most of the descending activity reaches
motoneuron pools through a few penultimate inter-
nuncial circuits (243) that transform this activity into
higWy ordered patterns of input. In conjunction with
previous findings, the results suggested that the sus-
FIG. 11. Stretch-evoked responses of 2 a-motoneurons recorded ceptibility of a motoneuron to discharge is strongly
from a fIlament of 7th lumbar ventral root of cat. Amount of tension correlated with its size regardless of the source of the
applied and developed reflexly is indicated by separation of 2 top excitation and the neural circuits that transmit it to
beams in each frame. Several seconds elapsed between successive
frames while muscle was stretched (1-5) and released (6-9). [From
the motoneurons, Input is not excluded as an impor-
Henneman et al. (159).] tant factor in this correlation, but it is difficult to
understand how each of the various inputs in these
was a continuous series of impulses of very low ampli- experiments could supply more excitatory fibers to
tude (not seen in this tracing), This activity was de- small cells than to large ones, as some authors (36,
rived from the axons of y-motoneurons, which are of 192) suggest.
smaller diameter than the a-fibers. Susceptibility to inhibition is also correlated with
In Figure 12 a series of single sweeps is reproduced cell size, as illustrated in Figure 14, The upper portion
to illustrate recruitment in a fIlament containing the of the figure shows orderly recruitment of three triceps
axons of five triceps motoneurons, With increasing motoneurons with increasing stretch. After the normal
stretch progressively larger impulses appeared, which pattern of recruitment had been established a 4-kg
are numbered from 1 to 5. The recruitment order was stretch, sufficient to elicit tonic firing of all three
1, 2, 3, 4, 5; the order of dropout was 5, 4, 3, 2, 1. motoneurons, was applied by elongating the triceps to
Figure 13 illustrates the distribution of stretch a fixed length. This stretch was maintained while the
thresholds for 208 motoneurons. The tensions required effects of inhibition, shown below, were recorded.
to elicit maintained rhythmic firing ranged from less Weak inhibition silenced the largest unit, Inhibition of
than 0.4 kg to more than 8 kg, moderate intensity suppressed firing of the unit of
These observations (159), together with those al- intermediate size as well, leaving the smallest unit
ready described, suggested that orderly recruitment discharging until a still stronger inhibition eliminated
must depend on differences in the excitabilities of the all responses to stretch. During each of the brief
motoneurons themselves, on some systematic differ- periods after inhibition there was a partial but incom-
ence in the input they received that was correlated plete recovery of the original pattern of response.
with their sizes, or on a combination of these factors. Results of this kind were obtained in a variety of
To distinguish between these possibilities, recruitment experiments, In general, the larger the unit, the more
was examined, using a number of different inputs. A readily it was silenced by inhibition. Regardless of the
variety of spinal reflexes was elicited to discharge existing level of excitatory drive, regardless of whether
flexor or extensor motoneurons (160). Responses of the motoneurons were responding rhythmically to
pairs of units to the same, simultaneously delivered stretch or monosynaptically to synchronous volleys,
stimuli were compared, In general, the smaller of any and regardless of whether the inhibition was direct,
two responding units was discharged at a lower inten- internuncially mediated, autogenetic, or recurrent, the
sity of stimulation regardless of whether the stimuli susceptibility of each cell to inhibition appeared to be
arose ipsilaterally or contralaterally, physiologically or closely correlated with its size. It was concluded that
438 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

FIG. 12. Stretch-evoked responses of 5 a-motoneurons recorded from a fIlament of the 1st sacral
ventral root of cat. Numbers above action potentials indicate rank of units according to size. [From
Henneman et al. (159).]

50
Exp.l. ~ 59 units
Exp.n. D 69 units
40
Exp.m. ~ 80 units

etI
±:
;30
....o
~
CD
.0

;20
z

10

.8 1.6 2.4 3.2 4.0 4.8 5.6 6.4 7.2 8.0

Threshold (k9 tension)
FIG. 13. Frequency distribution of thresholds of tonic responses to stretch in 3 experiments on
cats. Abscissa, threshold tension of deefferented triceps muscle. Ordinate, number of units whose
thresholds fell between values indicated on abscissa. [From Henneman et al. (159).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 439

FIG. 14. A: orderly recruitment in cat of 3 tri-

ceps motoneurons of different sizes in response to
increasing degrees of stretch. Tension produced by
stretch of triceps muscle indicated by separation of
top traces in each frame. B: orderly inhibition of
same 3 units during a constant stretch of 4 kg.
Records 1-4 on left show control responses of
stretch before, between and after 3 inhibitory stim-
ulations. Records on right were obtained during
100/s stimulation of ipsilateral deep peroneal nerve
at 3 intensities. Largest unit was silenced first (line
1), intermediate unit next (line 2), and smallest
unit last (line 3). Note lasting effects produced by
brief inhibitions, i.e., failure to recover to previous
levels in lines 2 and 4. [From Henneman et al.
(160).]

whereas the excitability of motoneurons (including the correlated with its size. This set of correlations was
role of input) is an inverse function of cell size, their referred to as the "size principle" (159, 160).
inhibitability is a direct function of cell size.
The net effects of cell size, excitatory input, and Measurement of Total Output
inhibition on the responses of three different moto- of Motoneuron Pools
neurons are represented quantitatively by means of
three-dimensional graphs in Figure 15. The responses By 1965 it had been shown that there was a striking
of the smallest and most excitable unit, which was degree of orderliness in both the neural and contractile
spontaneously active with the muscle relaxed, are properties of motor units. From the outset, however,
plotted in Figure 15A. The stretch threshold of the apparent exceptions in recruitment order had been
intermediate unit (Fig. 15B) was between 0 and 5 mm noted. It was difficult to ascertain whether these were
extension and that of the largest and least excitable due to 1) decreased amplitude of impulses in injured
unit was between 5 and 10 mm. At all levels of stretch fibers; 2) recording of impulses from separate but
the intensity of inhibitory stimulation required to si- synergistic pools, due to mixture of their axons in
lence a unit was always the greatest for the most ventral root filaments; or 3) true departures from the
excitable cell (Fig. 15A). In Figure 15D the responses general principle that cell size is highly correlated with
of all the motor units are combined in the contractile susceptibility to discharge. There was clearly a need
response of the whole muscle. for defining the operations of the motoneuron pool in
In summary, by 1966 it had been shown that the quantitative terms so that the activity of each cell
neural energy required to discharge a motoneuron, the could be related to that of the entire pool. A technique
energy it transmits and releases in its muscle, its was devised for measuring the simultaneous discharge
excitability and inhibitability, its mean rate of firing, of all the cells in a pool, and, from this, the percentage
and even its rate of protein synthesis (280) are all of output during any monosynaptic reflex (59). The
440 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

method is illustrated in Figure 16. The combined ously the monosynaptic reflexes of one triceps moto-
nerves to the medial gastrocnemius muscle (MG) and neuron were recorded from a thin fIlament of ventral
the lateral gastrocnemius and soleus muscles (LG-S), root. The vertical lines in the top tracing of Figure 18
which together make up the triceps surae muscle, were are the time integrals of successive population reflexes
stimulated with brief shocks (1 shock/s) that set up recorded at a rate of l/s. The height of each line varies
antidromic volleys in all of the a-motor fibers of these with the number of motoneurons discharged and their
nerves and dromic volleys in all of the group la-fibers. sizes. Just before this series was recorded the com-
The antidromic volley was recorded in the distal half bined MG and LG-S nerves were tetanized. The re-
of the seventh lumbar or first sacral ventral root (L 7VR sponses in Figure 18 were recorded during the decline
or SIVR) and displayed on one trace of a two-beam in the resulting PTP. As long as the responses of the
oscilloscope. The dromic volley elicited a monosyn- triceps population exceeded a certain level, the triceps
aptic reflex in the proximal half of the same ventral unit below discharged in every reflex. When the pop-
root, which was displayed on the other trace. The base ulation response declined below this level, the unit
lines of these two traces are superimposed in Figure ceased to respond. The first failure of the unit to
17. The left potential in each pair of responses is the discharge was followed by two successive responses
antidromic volley. It is a measure of the discharge of before the unit ceased to fire completely. Careful ex-
all of the MG and LG-S motor fibers in this ventral amination of the record reveals that the first failure
root. The right potential in each pair is the monosyn- was associated with a smaller population response
aptic reflex. Its size is determined by the number of than were the next two unit responses. The "critical
MG and LG-S motoneurons discharged during the firing level" (CFL) of this unit was, thus, between the
reflex and the sizes of their action potentials. Before levels associated with the first failure and the smaller
a 12-s conditioning tetanus to the muscle nerves at 500 of the next two population responses. The difference
shocks/s, the amplitude of the reflex was less than between these levels was less than 1% of the maximal
one-half that of the antidromic volley (top row of output of the triceps pool in this experiment.
responses in Fig. 17). After the tetanus the reflex The records in Figure 18 suggest that CFL is very
increased in amplitude due to posttetanic potentiation sharply defined. When the PTP was repeated several
(PTP) (237) and reached temporary equality with the times and sufficient observations were available within
antidromic volley (left side of second row). Thereafter, the critical range, however, it was usually apparent
the reflex response gradually declined back to its orig- that there was a slight degree of uncertainty in the
inal amplitude. At the peak of PTP all of the MG and firing of the motoneuron; i.e., there was a narrow range
LG-S motoneurons were evidently discharged reflexly, of population responses within which a unit might or
as nearly as could be determined by comparison of might not respond monosynaptically. This is evident
amplitudes. A more accurate measure of a compound in Figure 19, which reproduces the responses of three
action potential can be obtained by integrating it typical triceps units with relatively sharp "thresholds."
electronically. This was done in the experiments de- The plateau at the beginning of the top trace on the
scribed below. The time integrals had an absolute left indicates the 100% level of the population response
accuracy of 1%-2% and varied by less than 0.2% on at the peak of potentiation. During the decline in
repeated trials. Thus, any monosynaptic reflex could potentiation each of the three units abruptly stopped
be described quantitatively, using a scale of output responding at a different level and did not discharge
running from 0 to 100%. again in that series. Close inspection revealed, how-
ever, that the unit that ceased firing at approximately
Critical Firing Levels of Motoneurons the 60% level actually discharged once at 58.6% and
With the technique just described the monosynaptic failed to respond several times just above the 60%
reflexes of a large population of triceps motoneurons level. The range over which the unit's response was
were recorded in decerebrate cats (155). Simultane- unpredictable was called the uncertain range. The

FIG. 15. Graphic representation of discharge frequency of 3 motoneurons of different sizes (A-C)
in triceps surae of cat, and of contractile tension developed by the muscle itself (D) in response to
varying degrees of excitation and inhibition: x-axis, intensity of inhibitory stimulation applied to
ipsilateral deep peroneal nerve; y-axis, frequency of discharge in A-C and contractile tension in D; z-
axis, stretch of triceps surae, in millimeters. Data for all 4 graphs obtained simultaneously. Cells were
subjected simultaneously to various mixtures of excitatory (z-axis) and inhibitory (x-axis) stimuli.
Each plotted point represents mean of 2 successive determinations. Inhibitory stimuli were l00/s
shocks applied for 0.9-s veriods at 4-s intervals. Note that unit A (smallest) fIred spontaneously
without stretch, unit B began to discharge between 0 and 5 mm of stretch, and unit C (the largest)
between 5 and 10 mm. [From Henneman et al. (160).]
 41CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 441

A B

c o
442 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

Shmulah\l\g
~leetrodes
Muscle Nerve
~

a'V\hdromlc volley
iV\ \l\l\ot02'" fibe1"s

FIG. 16. Scheme of experiment used to measure maximal monosynaptic discharge of a motoneuron
pool. DR, dorsal root; VR, ventral root. [From Clamann, Henneman, et al. (59).]

critical firing level was defined as the arithmetic mean

of the uncertain range. In this series of experiments
203 triceps units were examined. Their CFLs ranged
from 0 to 98.7%. The mean uncertain range for the
entire group (2.1%) did not vary systematically with
the CFL.
No technique was available for measuring critical
firing levels directly during repetitive firing; instead, it
was possible to make use of the order of recruitment
in pairs of motoneurons as a measure of their suscep-
tibility to repetitive discharge and to compare this
order with the critical firing levels established in
monosynaptic tests. Two examples of this approach
are shown in Figure 20. Recordings were taken from
two different pairs of plantaris motoneurons in the
same experiment. As the intensity was first increased
and then decreased manually, the two units com-
menced and ceased~their firing at different intensities
of stimulation, thm;signaling their different thresholds
to repetitive discharge. In both pairs of units the
motoneuron with the lower CFL (shown at the ends
of traces b and c) began to discharge sooner during
the increase in stimulation intensity and ceased firing
later during the decrease. In the lower pair of units,
differing in CFL by 3.1%, the onsets of firing were
farther apart, and the last discharges were more sep- FIG. 17. Pairs of responses in cat recorded as in Figure 16 from
proximal and distal halves of 1st sacral ventral root in response to
arated in time than in the upper pair of units, which single-shock stimulation of combined nerves to medial and lateral
differed by only 1. 7%. gastrocnemius (MG and LG). Left and right deflections of each pair
Sixty-two pairs of motoneurons were compared as are antidromic and reflex responses, respectively, displayed on su-
illustrated in Figures 20 and 21. In 57 of these pairs perimposed traces of a 2-beam oscilloscope. First 10 pairs in top row
recorded before 12-s tetanus (500 shocks/s) to MG and LG nerves.
the unit with the lower CFL in monosynaptic tests Subsequent pairs recorded at 2-s intervals after tetanus. [From
was invariably the more susceptible to repetitive firing. Clamann, Henneman, et al. (59).]
In 5 pairs, which differed in CFLs by only 1. 7%, 1.1%,
2.0%,2.1%, and 5.0%, the unit with the lower CFL was
slightly less susceptible to repetitive firing. The minor activated several afferent systems in addition to the
disagreements between the two methods of ranking group la-fibers and presumably elicited descending
motoneurons were probably due to obtaining insuffi- activity from higher centers as well as recurrent inhi-
cient data in estimating the CFL monosynaptically. bition from the motoneuron discharges. The results,
They tended to disappear when a longer series of therefore, have considerable generality. They apply to
monosynaptic reflexes was recorded for this purpose. mixtures of inputs that might occur under normal
The repetitive stimulation used in these experiments conditions, not merely to group Ia inputs.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 443

FIG. 18. Simultaneous recordings in cat of a series of monosynaptic reflexes of triceps surae pool
(top) and a single triceps motoneuron (bottom) showing critical fIring level of the latter. Top trace,
time integrals of monosynaptic reflexes recorded at lis from proximal half of 1st sacral ventral root.
Height of each vertical line measures size of population response. Bottom trace, monosynaptic reflexes
of a single triceps motoneuron recorded from a small fIlament of 7th lumbar ventral root. Records
were made during decline in posttetanic potentiation following a brief train of conditioning shocks
that were applied to the combined medial gastrocnemius and lateral gastrocnemius and soleus nerves
at 500/s. [From Henneman et al. (155).]

FIG. 19. Critical fIring levels of 3 triceps motoneurons in cat. Single unit in bottom left had a sharp
threshold at 39%. The 2 units in bottom right tracing had thresholds of 60% and 33%. Plateau in top
left record indicates 1()()% discharge level of triceps pool. [From Henneman et al. (155).]

Relation of Critical Firing Level to this, which is fully described in reference 60. It in-
Axon Diameter and Motoneuron Size volved three separate procedures. 1) Monosynaptic
reflexes of single motor fibers were recorded mono-
Defining the CFL and relating it to rank order in phasically from ventral root filaments. A resistor was
repetitive firing made it important to correlate CFL placed in shunt across the recording electrodes and its
with axonal diameter and motoneuron size by direct value was varied until the action potentials were re-
quantitative methods. A technique was devised to do duced by one-half. The resistance of the nerve filament
444 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

FIG. 20. Susceptibilities of 2 pairs of cat plantaris motoneurons with known critical fIring levels
(CFLs) to repetitive discharge in cat. Electrical stimulation of plantaris nerve at 300 stimuli/s in both
cases. Stimulus intensity (i.e., voltage) indicated by level of trace a with reference to base line d.
Responses of plantaris units recorded on traces band c. CFLs indicated at right end of these traces.
Time marks at 1-s intervals for both top and bottom records. [From Henneman et al. (155).]

was then equal to that of the shunt. Dividing the Since both CFL and axon diameter were related to
voltage of the action potentials by the resistance of impulse size, they were related to each other, as illus-
the filament gave the axonal action current. 2) In trated in Figure 22.
experiments in which impulses were conducted antid- The linearity of the data points in Figure 22 suggests
romically over long distances to yield accurate mea- that CFL is a function of a single continuous variable,
surements of conduction velocity, it was shown that which is probably cell size. If CFL is precisely and
the axonal current of an impulse varied as the square linearly related to cell size, it must be concluded that
of its conduction velocity. 3) After the sizes of the the distribution of input to the pool and any other
impulses had been normalized in accordance with the factors that contribute to the relationship are also
resistances of their ventral root fIlaments, a direct correlated with cell size. It is appropriate, therefore,
correlation was found between impulse size and CFL. to ask whether the scatter in the data points of Figure
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 445

100

o
~
0

-SO
0::
W
o
~ 60 00
o
I
~
Z
<t 40 00
0::

FIG. 21. Differences in susceptibility to repetitive fIring of 2 cat .-

plantaris motoneurons whose critical fIring levels were 12.0% (top) z 20
and 15.5% (bottom). A: shocks applied to plantaris nerve were ::>
supramaximal for all group la-fIbers. Frequency was switched from
20 shocks/s initially to 30 shocks/s at 1st arrow, and to 40 shocks/
s at 2nd arrow. B: frequency of stimulation was 500 shocks/s
throughout and intensity was increased gradually as trace moved
from left to right. [From Henneman et al. (155).]
AXON SIZE (arbitrary units)
22 indicates a degree of variability or nonlinearity in 22. Relation between axon size and critical fIring level (rank
FIG.
this relationship. When the experimental observations order) of 21 cat plantaris motoneurons isolated in a single experi-
were repeated several times, variations were noted ment. [From Clamann and Henneman (60).]
that were sufficient to account for all of the scatter.
Technical limitations, therefore, account for the ob-
served scatter. Barrett and Crill (14) have demon-
II I
strated close relationships between input impedance
and conduction velocity and between soma size and 1'11111 H+t+t+++++++-:---,C F L
conduction velocity. Their observations are consistent
with the conclusion that CFL is a function of a single
continuous variable such as cell size, but, of course,
they do not provide direct proof of the relationship.

Effects of Inhibitory
Inputs on Critical Firing Level
and Rank Order During Repetitive Firing
Although the motoneuron pool in the preceding
FIG. 23. Critical fIring level (CFL) of a cat plantaris motoneuron
experiments was subjected to a variety of excitatory with and without inhibitory input from lateral popliteal nerve. Top,
and inhibitory influences, the specific inputs used to time integrals of monosynaptic reflexes of plantaris population
elicit monosynaptic reflexes and repetitive firing were elicited at 2-s intervals during declining phase of posttetanic poten-
predominantly excitatory to the pool. An important tiation. Bottom, simultaneous responses of a plantaris unit with a
critical fIring level of 61%. Alternate reflexes preceded by a 200-ms
question to be answered was whether the critical firing train of pulses to lateral popliteal nerve. Horizontal line indicating
levels and rank orders obtained in such experiments critical fIring level drawn as described in text. [From Clamann,
would be affected by the deliberate application of Gillies, and Henneman (58).]
various types of inhibition.
The oscillographic record reproduced in Figure 23 sponses of a single plantaris unit with a CFL of 61%.
illustrates the effects of a potent inhibitory input from Alternate reflexes of the plantaris pool were preceded
the lateral popliteal nerve on the monosynaptic re- by a 2oo-ms train of pulses applied to the lateral
flexes of a pool of plantaris motoneurons and a single popliteal nerve. The parameters of this pulse train
unit from the same population. The top trace records were adjusted to cause a 20% average inhibition of the
the time integrals of the monosynaptic reflexes of the maximum monosynaptic reflex of the plantaris pool.
plantaris population elicited at 2-s intervals during the As the monosynaptic reflexes declined from their peak,
declining phase of a posttetanic potentiation. The the absolute amount of inhibition resulting from the
lower trace, recorded simultaneously, shows the re- peroneal input remained approximately constant. The
446 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

single unit in the lower tracing discharged without Figure 24 illustrate this point. They show the effects
exception as long as the corresponding population of recurrent inhibition and lateral popliteal inhibition
responses reached the level indicated by the horizontal on the monosynaptic reflexes and the repetitive firing
line labeled CFL. This line was drawn at a level equal of a pair of plantaris motoneurons. In each of the four
to the size of the smallest uninhibited population comparisons shown in this figure the upper unit had
response that was accompanied by a unit discharge, a slightly higher CFL (18.6% in traces 1 and 3 and
and just above the largest inhibited response without 18.4% in traces 2 and 4). Each of the underlined
a unit discharge. The CFL of this unit was, thus, monosynaptic reflexes shown in Al and A2 was pre-
unchanged by the addition of a potent inhibitory in- ceded and accompanied by a 100-ms train of shocks
put. The effects of inhibition on 32 triceps surae units (100/s) applied to the proximal portion of the seventh
with CFLs ranging from 3% to 87% were examined as lumbar ventral root. The antidromic volleys set up by
above by stimulating antagonistic nerves in the same this stimulation resulted in recurrent inhibition me-
hindlimb or inhibitory sites in the medial reticular diated by Renshaw cells. Although 9 of the 18 reflexes
formation. The general conclusion was that inhibition in trace 1 were inhibited by this input, only 1 of those
had no significant effect on CFL in these experiments. in trace 2 was silenced. In the absence of recurrent
Experiments were also carried out to study the inhibition, both units responded every time. In A3 and
effects of inhibition on the repetitive firing of moto- A4 the same pair of units is seen responding rhyth-
neurons. As a rule, when any two motoneurons be- mically to repetitive stimulation of the plantaris nerve.
longing to the same pool were compared directly for When continuous recurrent inhibition was begun, as
susceptibility to inhibition, the results could be pre- indicated by the thicker base line in A4, the upper unit
dicted from prior measurements of their critical firing ceased firing and the lower unit slowed to about one-
levels. This was always the case when the CFLs dif- half its original rate of discharge. In Figure 24B, the
fered by more than 2.5% and was usually true when same pair of units was subjected to an inhibitory input
they were closer together. The records reproduced in elicited by stimulating the lateral popliteal nerve at

FIG. 24. Effects of recurrent inhibition (A) and lateral popliteal inhibition (B) in cat on monosyn-
aptic reflexes (traces 1, 2) and repetitive fIring (traces 3, 4) of a pair of plantaris motoneurons with
critical fIring levels of 18.6% (top) and 18.4% (bottom). A: in traces 1, 2 the underlined reflexes were
preceded and accompanied by a l00-ms train of shocks (100/s) to proximal portion of 7th lumbar
ventral root. Antidromic volleys set up by stimulation in AI, A2 resulted in recurrent inhibition
mediated by Renshaw cells. Of the 18 reflexes in trace 1, 9 were inhibited by this input; in trace 2 only
1 reflex was silenced. In A3, A4 the same pair of units is seen responding rhythmically to repetitive
stimulation of the plantaris nerve. When continuous recurrent inhibition was begun, indicated by
thick base line in )\4, top unit ceased fIring and bottom unit slowed to about one-half original rate of
discharge. B: same pair of units was subjected to inhibitory input elicited by stimulating lateral
popliteal nerve at l00/s. In B1, B2 the underlined monosynaptic reflexes were somewhat more
susceptible to this inhibition than to the recurrent inhibition. Of the 16 reflexes receiving this
inhibition, 15 in trace 1 and 4 in trace 2 were silenced. In B3, B4 the same pair of units fIring
rhythmically in response to 100/s stimulation of the plantaris nerve was subjected to an inhibitory
input whose intensity was gradually increased as the traces moved across the oscilloscope. The top
unit was fInally suppressed by this input whereas the bottom unit was only slowed in rate. [Clamann,
Gillies, and Henneman (58).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 447

100/s. The underlined monosynaptic reflexes in B1 of this input was gradually increased manually as the
and B2 were somewhat more susceptible to this inhi- beam moved across the screen. The intensity of the
bition than to the recurrent inhibition. Of the 16 inhibition required to silence each of the plantaris
reflexes receiving this inhibition, 15 in trace 1 and 4 in units was determined by measuring the level of the
trace 2 were silenced. In B3 and B4 the same pair of trace at the time of their last discharge. Similar mea-
units firing rhythmically in response to 100/s stimu- surements were made at six different levels of excit-
lation of the plantaris nerve was subjected to an inhib- atory input. The results are graphed in Figure 26. The
itory input whose intensity was gradually increased as data for the lower unit, with a critical firing level of
the traces moved across the oscilloscope. The upper 12.7%, are plotted with filled circles and a solid-line
unit was finally suppressed by this input, whereas the curve; those for the upper unit, with a critical firing
lower one was only slowed in rate. All four compari- level of 18.8%, are represented by open circles and a
sons in Figure 24 reveal that the unit with a slightly dashed curve. The plot shows that every increase in
lower CFL was the more resistant to inhibition regard- the excitatory stimuli required a proportionate in-
less of the mode of firing or the type of inhibition. crease in the inhibitory stimuli required to silence the
In a wide variety of experimental conditions the two plantaris units. For each of the six excitatory
activity of individual motoneurons is determined by inputs, the corresponding inhibitory input was greater
the relative magnitudes of the excitatory and inhibi- for the unit with the lower critical firing level. Fur-
tory inputs. The quantitative relations between the thermore, the difference between the two inhibitory
two types of input and their effects were examined inputs was approximately the same at all levels of
more closely by measuring the intensities of the stimuli excitation. It must be emphasized that the stimulus
used to produce the inputs. The record reproduced in intensities plotted in Figure 26 do not necessarily
Figure 25 is an example of the experimental data correspond to the magnitudes of the neural inputs that
obtained. Two plantaris units, whose CFLs had been result from the stimulations. Until the actual neural
measured previously, were set into rhythmic discharge inputs can be measured directly (no technique is cur-
by 100/s stimulation of the plantaris nerve. A sample rently available for this purpose) the quantitative va-
of the excitatory stimuli was led to the oscilloscope. lidity of these results remains to be verified. Still, the
The stimulus strength of this input was constant dur- findings provide evidence for the tentative conclusion
ing each test period but varied in different tests. The that inhibition often produced effects that are similar
lateral popliteal nerve was stimulated at 100/s to to a simple reduction in excitation. They also suggest
produce a competing inhibition. The stimulus strength that the competition between excitation and inhibition
in these experiments is a simple algebraic process.

FIG. 25. Records illustrating quantitative relationships between

intensity of excitatory stimulus causing 2 cat motoneurons with
critical fIring levels of 18.8% and 12.7% to discharge repetitively and
intensity of inhibitory stimulus required to silence them. Stimulus Excitatory Stimuli
strength of excitatory input (100/s stimulation of a plantaris nerve),
indicated by level of line E with reference to base line B, was FIG. 26. Graph showing linear relation between strength of ex-
constant during each test but varied in different tests. Stimulus citatory stimuli causing each of 2 cat motoneurons to discharge
strength of inhibitory input OOO/s stimulation of lateral popliteal repetitively and strength of inhibitory stimuli required to silence
nerve), indicated by line I, was increased as trace moved across them. Data obtained as illustrated in Figure 25. Data for lower unit,
screen. Intensity of inhibition required to silence each unit was with critical fIring level of 12.7%, are plotted with filled circles and
measured at time of their last discharges, indicated by dashed lines. solid line; data for upper unit, with a critical fIring level of 18.8%,
Time signals at top of record represent I-s intervals. [From Cla- are represented by open circles and dashed line. [From Clamann,
mann, Gillies, and Henneman (58).] Gillies, and Henneman (58).]
448 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

The effects of inhibition on 133 pairs of plantaris impossible to study in animals can often be carried out
motoneurons are compiled in Table 1. In 123 pairs quite simply by a cooperative subject. Furthermore,
(92%) the rank order for repetitive discharge was the the results are not open to the criticisms that they are
same as for monosynaptic reflexes. The differences in due to limited or unphysiological inputs, to the effects
the CFLs for this group of units ranged from 0.2% to of anesthesia, or to abnormal preparations such as
72.1% and the mean difference was 15.9%. In the 10 decerebrates. In humans, recordings are usually taken
remaining pairs with differences in CFLs from 0.4% to from muscles instead of nerves. Hence artifacts due to
9.9% and a mean difference of 3.2%, the rank order was shifting of the position of the electrodes relative to the
reversed for repetitive fIring. In three monosynaptic active muscle fIbers must be minimized. This move-
comparisons, i.e., 1.8%, inhibition reversed the original ment has been a serious problem in some studies,
rank order. In two of them the differences in CFLs leading to apparent reversals in recruitment order, and
were 1.0% and 0.4%. In four comparisons, i.e., 2%, it requires constant vigilance from experimenters. In
inhibition reversed the rank order established for re- a study by Olson, Carpenter, and Henneman (272),
petitive discharge. The mean difference in CFLs in orderly recruitment of muscle action potentials of
these four pairs of units was 0.8%. increasing amplitude was demonstrated on several cat
The six inputs used in these experiments were se- muscles. Some of the factors influencing the recorded
lected to provide contrasting types of inhibition for size of motor unit potentials and the apparent order of
testing: postsynaptic and presynaptic, spinal and su- recruitment were discussed. They included the dis-
praspinal, unilateral and contralateral, recurrent and tance of the active fIbers from the electrodes, differ-
nonrecurrent, cutaneous and muscular, and mixed and ences in the diameters of muscle fIbers, and possible
pure. Although these inhibitory systems and the effect variations in the density of the fIbers of a single motor
they exert differ from one another in many respects, unit. It is noteworthy that orderly recruitment was
their effects on the rank order of motoneurons in both readily observed in the cat soleus, a homogeneous
monosynaptic and repetitive tests were indistinguish- muscle in which differences in histochemical fIber
able. As Table 1 indicates, none of the inputs caused types and diameters were not signifIcant factors.
a signifIcant number of reversals. The small percent- Milner-Brown et al. (258) used the action potentials
ages of reversals in Table 1 all occurred in pairs that of single motor units in the fIrst dorsal interosseous
were nearly equal in CFL. muscle of human subjects to trigger an averager that
recorded the twitch contractions of motor units in that
Recruitment of Motor Units in Humans muscle. Those contractions that were time locked with
Despite the human and technical limitations in- the trigger signals from a particular unit were ex-
volved, studies on motor units in humans have added tracted by an averaging process from the other con-
greatly to the information gained from animal experi- tractile activity in the muscle. The twitch tension of
ments and facilitated its interpretation. Human sub- each unit was related to the total force produced by
jects can voluntarily make delicately graded contrac- the whole muscle at the instant of recruitment. Figure
tions, hold a desired tension, produce rapid or slow 27 shows the number of motor units plotted on a linear
contractions on demand, and, by monitoring their scale and on a logarithmic scale, with the tensions
performance on an oscilloscope, attempt to alter the indicated. As in experiments on motor units in cats
order of recruitment. Actions that are difficult or (253, 273, 353), there was an approximately exponen-
tial relation between the numbers of motor units iso-
lated in each size range and the tensions they produced
TABLE 1. Effects of Inhibition (many units developing small tensions, few units gen-
on Pairs of Motoneurons erating large forces). The three-dimensional fIgure of
Burke et al. (41) illustrating a tripartite classifIcation
Monosynaptic Repetitive of motor units unintentionally gives the impression
Reflexes Discharge
that large motor units are much more numerous than
Source of Inhibitory Input No. of No. of small ones. Milner-Brown et al. (258) concluded that
No. of reversals No. of reversals
tests of rank tests of rank
motor units in human muscles are recruited in orderly
order order progression according to the amount of tension they
develop. The larger motor units, which were recruited
-------

Lateral peroneal nerve 36 0 38 I

Posterior biceps semi- 36 I 37 0 at higher threshold forces, tended to have shorter
tendinosus nerve contraction times than the smaller units, as is com-
Sural nerve 7 0 8 0 monly found in animal studies (35, 253, 353). Milner-
L 7 ventral root 29 I 49 0
Contralateral Sl dorsal 32 I 34 2
Brown et al. (258) added in their discussion that
root these findings strongly confirmed the "size principle"
Brain stem 30 0 30 expounded by Henneman.... Indeed, the degree of or-
dering ... is remarkably high (linear correlation coeffi-
Totals 170 3 196 4 cients > 0.8 for all three subjects) when one considers
Adapted from Clamann, Gillies, and Henneman (58). that the recordings were made in experiments over a
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 449

so
A

-40 ....
. .' CJ A.M .
.....
.-:-, [;Iill R. Y•
:.:: . rzzJ R.B.S.
30 ~:" .
....
.
':'.
' .
20

..
...
c:

..
:I
L
0
0
... SO
E
10

0
0 2·5
.
.m.~···~,~!bzz:::::J----FJZ=ZZL..,I1'-F
5·0 7·5 12·5

0
L
GI
.0
E
:I
Z
10

o 25 SO 7 5 10
Twitch tenSIon (g)

FIG. 27. Number of human motor units having the twitch tensions indicated. A: linear scale. B:
logarithmic scale. Distributions are similar for all 3 subjects. The computed best-fitting line on the
semilog plot (B) indicates an approximately exponential relation between number of motor units and
twitch tension. [From Milner-Brown et al. (258).]

period of several months, and that there are inevitably fler (125) demonstrated strong positive correlations
many uncontrolled variables in human experiments. between threshold force of recruitment, EMG spike
Since the size of a unit recruited (IlF) is proportional to amplitude, and twitch tension in the masseter over
the mean force level (F) over such a wide range, the longer and greater ranges of force (1O-s plateaus vary-
fractional increments in force (IlF IF) produced by re- ing from 1 to 47 kg).
cruiting each unit will be constant.... This result is
reminiscent of the constant Weber functions (IlS/S)
The relation between the tension a motor unit de-
sometimes obtained when just-noticeable differences (IlS) velops and the conduction velocity (i.e., diameter) of
are measured at various stimulus intensities (S) in sensory its axon has also been well documented in humans.
experiments.. . Freund et al. (115) reported that the total muscle
See Henneman and Olson (157) for a fuller discussion forces at which motor units were recruited were a
of this analogy. Using the same averaging technique, direct function of their axonal conduction velocities
Yemm (359) also found a nearly linear relationship (CVs). Recently Grimby and Hannerz (139) and Borg
between recruitment force and twitch tension in the et al. (23) published studies on the discharge properties
more powerful masseter and temporal muscles for the of short toe extensor motor units activated voluntarily.
0-6-kg force range. Subsequently, Goldberg and Der- Certain motor units could be driven continuously by
450 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

voluntary effort at frequencies of lO/s or less. These cle contraction and the onset of fIring in three of its
units gradually accelerated their fIring with increases motor units. Both scales are logarithmic. With rapid
in contraction strength, and had maximum rates below contraction (300-3,000 g/s) recruitment was greatly
30/s during a sustained contraction but above 60/s in compressed in time. The delay in the onset of fIring in
twitch contractions. This group of units was found to low- and high-threshold units decreased from about
have CVs between 30 and 45 m/s. In contrast, larger 2,500 ms in the slowest contractions to less than 100
units with CVs between 40 and 54 m/s could not be ms in the fastest. The parallel curves indicate again
driven continuously, did not fIre repeatedly at rates that although the high-threshold units started fIring
below 20/s, accelerated rapidly with increasing effort, much sooner in rapid contractions than in slow ones,
and attained maximal fIring rates above 100/s. Units they were not preferentially activated.
that were intermediate between these two groups had The relation of the fIrst twitch contraction of a unit
intermediate CVs. In a prolonged contraction of con- to the total muscle force is the mechanical counterpart
stant strength only continuously fIring units were ac- of electrical recruitment. By measuring the contrac-
tive. In brief rapid contractions both types of units tion time of a unit, the change in total muscle force
participated. Thus, a clear relationship was estab- that occurs within that time during a rising contraction
lished between the voluntary discharge properties of can be calculated. In Figure 28D the increase in total
a unit and its axonal CV, as one would expect from muscle force that occurs within the mean contraction
animal studies. The authors (23, 139) point out that time of the muscle is plotted on the ordinate against
the axonal CV of a motor unit is directly related to the rate of rise of tension. The average decrease of the
the histochemistry of its muscle fIbers, to their twitch threshold force of recruitment of 12 units of this mus-
time, and to twitch amplitude, and inversely related cle with increasing rates of rise of tension is plotted
to the input resistance and afterhyperpolarization of for comparison. The two slopes are almost equal but
its motoneuron. Since the voluntary discharge prop- of opposite direction. Hence, the sum of these two
erties are related to CV, they should also be related to variables is approximately constant over the range
the other characteristics noted above. Accordingly, examined in these experiments. The decrease in the
the voluntary discharge properties may be a valuable threshold force of recruitment with increasing rate of
index to the other parameters of a motor unit that are rise of tension almost exactly cancels the increase of
diffIcult to study in humans. It is clear from this force in the whole muscle during the contraction time.
summary that the voluntary discharge properties of As a result, the authors (30) conclude that the me-
human motor units do not differ signifIcantly from the chanical recruitment of a motor unit occurs at approx-
characteristics of motor units in animals responding imately the same force level regardless of the rate of
involuntarily to various experimental inputs. rise of tension.
One function of the motoneuron pool that has re- This analysis provides new insights into the orga-
ceived little attention so far is its control of the precise nization of the motoneuron pool. Since the order of
timing of the mechanical events in single motor units recruitment does not change with the rate of rise of
relative to each other and to the development of tension, Bi.idingen and Freund argue that "... the
tension in the entire muscle. In an important study proportion between the force generated by recruit-
Bi.idingen and Freund (30) analyzed the mechanical ment and by fIring rate modulation of the units acti-
aspects of recruitment as they relate to the motoneu- vated prior to a particular unit remains unchanged.
ronal events. Single motor units from the human ex- This demonstrates how precise the tuning is between
tensor indicis muscle were investigated during volun- the recruitment of units and the activity of those
tary isometric contractions of equal amplitude carried already recruited." The fIndings demonstrate that the
out at different speeds, as illustrated in Figure 28A. activation of the motoneurons in a pool is organized
As Tanji and Kato (335) had shown, the threshold to achieve precise mechanical effects, i.e., stable rela-
force of recruitment for all motor units apparently tions between the force output of the individual motor
decreased as the speed of contraction was increased. units at the time of their mechanical recruitment and
Plots of the threshold forces versus the rates of in- the corresponding force outputs of the whole muscle.
crease in tension had approximately the same slopes These observations help to explain why the size and
in seven different units of the same muscle (Fig. 28C). excitability of a motor unit are necessarily related to
This monotonic decrease in threshold was consistently its contraction time, rate of fIring, and recruitment
observed in all units tested. No evidence was found to order.
indicate that high-threshold units were preferentially As described above, recruitment is orderly in con-
activated in fast, phasic (2,000 g/s) contractions. tractions of up to 2,500 g/s. Desmedt and Godaux (90)
Due to the general decrease in all force thresholds have developed a technique for investigating even
with faster contractions, the total time between the faster "ballistic" contractions that are completed in
recruitment of low- and high-threshold units dimin- too short a time to be controlled by sensory feedback
ished considerably. This is apparent in Figure 28B, signals from the activated muscles. The commands in
which plots the rate of rise of tension against the time ballistic movements are presumably preprogrammed
elapsing between the beginning of the voluntary mus- and must be dispatched to the segmental circuits and
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 451

A
400
c

c_-+-++11IltHIIll l l tttttll l l ----+1I ++JfI111HlI ! 111111tt+1

11 - -

D-----IH1IIHIf--------
300~~~
500 1000 1500 2000
R.t. 01 rls. 01 t.nslon (1I/s.cl
B o
"'\:
.g
.3000

.\~
lit CIt
.....,
~ .
u
41
• 150
i
!
• E

"'o~ ..
~
0

.\~
u
~ ~
300
iu
'0
•..
.~~.
c:
0
c:
41

•
41

;
~
41 30
E 30 300 3000 500 1000 1500 2000
i= Rete of rise of tension (9/sec: ) Ret. of rls. of tension (g/sec)
FIG. 28. A: fIring pattern of a high-threshold unit recorded from human extensor indicis during
isometric contractions with different rates of rise of tension (A-D). Each registration shows (top to
bottom) the spike record, isometric force, and tracking signal which was visually displayed to the
subject. Time scale in bottom 2 records is one-half that of top 2 records. During increasingly faster
contractions, unit starts fIring at successively lower force levels. B: dependence of time between
beginning of muscle contraction and firing onset of 3 motor units on rate of rise of tension. Double
logarithmic scale. Faster contraction causes earlier recruitment. C: change of threshold force of
recruitment (ordinate) associated with variation of rate of rise of isometric tension. Recordings from
7 motor units of extensor indicis of 3 subjects (. A, .). D: regression line of change of threshold force
of recruitment on rate of rise of isometric tension (y, = -0.067x + 198.66) calculated from 12 motor
units of extensor indicis. The increase of muscle force that occurs within the mean contraction time
of the muscle is shown for comparison (Y2 = 0.060x). [From Biidingen and Freund (30).]

motoneurons before the contraction begins. For study or more contractions were then ranked in order of
of these movements, subjects were asked to make 50 increasing peak force and the corresponding activity
or more ballistic contractions of tibialis anterior mus- of each unit was compiled as in Table 2. Unit 1 fired
cle at intervals of several seconds. Peak forces ranging in all 52 trials; units 2-5 failed to respond below certain
from 0.05 to 12.0 kg were produced in less than 0.15 s peak forces, but discharged with increasing probability
each. During each contraction the action potentials of in the progressively stronger contractions listed in the
the discharging unit at the recording site were re- table. The lower part of Table 2 gives the thresholds
corded along with the total contractile force. The 50 of the same 5 units for ballistic and slow ramp con-
452 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

TABLE 2. Motor Unit Threshold Data

Unit Unit
Trial Force, kg Trial Force, kg
2 3 4 5 2 3 4 5

1 0.07 + 27 2.27 + + +
2 0.08 + 28 2.42 + + +
3 0.15 + 29 2.72 + + + +
4 0.30 + 30 2.75 + + +
5 0.30 + 31 3.18 + + + +
6 0.31 + 32 3.20 + + +
7 0.31 + 33 3.33 + + + +
8 0.37 + 34 3.48 + + + +
9 0.38 + 35 3.51 + + + +
10 0.45 + + 36 3.78 + + + +
11 0.46 + + 37 4.54 + + + + +
12 0.53 + 38 4.84 + + + + +
13 0.75 + + 39 4.99 + + + + +
14 0.90 + + 40 5.15 + + + + +
15 0.90 + + 41 5.30 + + + + +
16 0.91 + + 42 5.45 + + + + +
17 1.06 + + 43 5.75 + + + + +
18 1.06 + + 44 6.96 + + + + +
19 1.06 + + 45 8.03 + + + + +
20 1.14 + + + 46 8.10 + + + + +
21 1.21 + + + 47 8.18 + + + + +
22 1.36 + + 48 8.63 + + + + +
23 1.38 + + + 49 9.54 + + + + +
24 1.51 + + 50 9.69 + + + + +
25 1.81 + + + 51 10.90 + + + + +
26 1.82 + + + 52 12.70 + + + + +
Ballistic force threshold, kg Ramp force threshold, kg
Unit
Range Mean Range Mean ± SD
1 0-0.07 0.035 0.14-0.48 0.34 ± 0.15
2 0.38-0.75 0.56 3.50-4.10 3.70 ± 0.28
3 1.06-1.81 1.43 3.60-4.40 4.14 ± 0.36
4 2.72-3.33 3.02 11.98-13.15 12.66 ± 0.57
5 3.78-4.54 4.16 13.73-16.41 15.46 ± 0.93
+, Motor unit discharg~; -, no discharge. Ballistic threshold was estimated by taking mean between maximum peak force for which a
unit never discharged and minimum force for which it always fIred. Ramp threshold estimated from data of 10 separate ramp contractions
producing 12 kg in 8 s. [Adapted from Desmedt and Godaux (90).]

tractions. The two sets of thresholds were compared Finally, evidence was presented (93, 95) that vibra-
by the rank correlation method, which yielded a very tion-induced inhibition of motor units silences them
high correlation coefficient (0.95). Four other muscles in exactly the reverse order from that in which they
(masseter, first dorsal interosseous, tibialis, and soleus) were recruited, as originally described in animal ex-
were later studied by the same investigators (91) with periments (58, 159).
similar results. Since the pyramidal tract is strongly Thus, many of the findings described in the early
involved in ballistic movements, it was suggested that part of this chapter have been confirmed in experi-
pyramidal signals impinge on the motoneuron pool in ments on human subjects, so that anesthesia, decere-
a pattern similar to that of the group Ia and II spindle bration, and other procedures, such as extensive de-
afferent fibers from the muscle itself. nervation and prolonged immobilization, could not
Further evidence was provided by Desmedt and influence the results. In addition, studies on humans
Godaux (92), who studied motor units in the first have added important new findings that provide fur-
dorsal interosseous muscle with the spike-triggered ther insight into the organization of the motoneuron
averaging technique, which allows the investigator to pool.
extract the tension myogram of individual twitch con-
tractions. It was clear with this approach that the Evidence Regarding Alternative
more rapidly contracting units were not preferentially Patterns of Recruitment
recruited in ballistic contractions. In fact, the fast
motor units were recruited after the slow motor units As the preceding sections have indicated, there is
even in fast contractions (94). This part of the study abundant evidence from animal and human studies
also showed that the motor units recruited at the that motoneurons in a pool are usually recruited in a
initial stage of a small ballistic movement produced stereotyped and repeatable order according to their
smaller forces than the higher threshold units. size. However, exceptions to this usual order have
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 453

been noted since the first investigations were reported TABLE 3. Recruitment Order and Amplitude of
and interest in their possible functional significance Motor Unit Spikes in Contractions Elicited
has increased. The early studies (159, 160) in which by Cortical Stimulation and by
discharges of motoneurons were recorded from ventral Pinching the Foot Pad
root filaments showed apparent reversals of recruit-
Cortex =
ment order in 11%-13% of comparisons. The extent to Motor Unit Cortex Pad
Pad
which these reversals were due to minor injuries to
fibers in dissected ventral root filaments or to the All pairs
Small always before large 27 24 20
possible mixture ofaxons from closely related but Tie 1 1 o
different pools was not known. With later, more quan- Large may precede small 6 (3) 9 (5) 2
titative techniques (155) reversals were seldom ob- Pairs of nonsimilar units
served in comparisons of units that differed by more Small always before large 23 20 17
Tie 1 1 0
than 2.5% in CFL (but see ref. 61). This led to the Large may precede small 4 (1) 7 (3) 1
belief that recruitment order was relatively, but not
Data for all 23 motor units of 6 small muscles (1st deep lumbricals
absolutely, fixed. Departures from the standard order of cat's foot). Recruitment threshold of each motor unit was com-
of recruitment have been reported by several authors pared to that of each other unit in same muscle. Total of 34 pairwise
(136, 137, 138, 335, 356, 206). Some of these studies comparisons were made. Peak-to-peak amplitudes of motor unit
have suffered from lack of information about the rel- spikes were measured. Cortex, recruitment patterns in response to
ative thresholds of different units, difficulty in repeat- cortical stimulation; Pad, recruitment patterns in response to pad
pinching; Cortex = Pad, same recruitment pattern for both kinds of
ing observations, and various technical limitations. stimuli. Numbers in parentheses indicate number of pairs where
Any deviation from the standard order of recruitment large unit consistently preceded small. Nonsimilar unit refers to
deserves special attention, however, because it indi- pairs where spike size and twitch amplitude in one unit were at least
cates that there are deficiencies in our knowledge of 30% larger than in the other. [From Kernell and Sjoholm (206).]
the mechanisms underlying recruitment and may pro-
vide useful clues regarding these mechanisms. This an ascending size order of recruitment in response to
section reviews evidence concerning variations in the pad pinching and vice versa. Hence, the exceptions
order of recruitment in animals and human subjects. were not the result of some difference in excitability
The best examples of variations in the order of between the various motoneurons, but were probably
recruitment are those illustrated in a study of Kernell due to differences in the distribution of input. In six
and Sjoholm (206), who produced maintained contrac- cats cortical stimulation was compared before and
tions of the first deep lumbrical muscle of the cat's after section of dorsal roots Ls-S3 . No obvious effect
foot either by electrical stimulation of the contralat- was produced by this procedure. In view of observa-
eral motor cortex or by pinching the foot pad, which tions of Thomas et al. (337), it would be valuable to
elicits a toe extensor reflex. The small size of this know whether the first deep lumbrical muscle in the
lumbrical muscle permitted the investigators to record cat is uni- or multifunctional, since only the latter type
the discharges of all the motor units of the muscle showpd reversals in their experiments.
electromyographically as well as the tensions they In a study of a larger pool of motoneurons, Kanda
produced. Cortical stimulation as well as pad pinching et al. (192) used bipolar hook electrodes on small
customarily recruited small units before large ones. natural filaments of the MG muscle nerve or fine
The order of recruitment caused by pad pinching often bipolar EMG electrodes to record electromyographi-
differed, however, from that caused by cortical stim- cally from motor units. Stretch or vibration of the MG
ulation. The most complete analyses of recruitment or SOL muscles was used to elicit reflex discharges of
were made in lumbricals containing five motor units motoneurons. Prompted by early reports (78) that
or less. Data from only six such muscles were compiled certain peripheral inputs could facilitate muscle ten-
in Table 3, but the results were confirmed in muscles sion in the fast gastrocnemius while simultaneously
with large numbers of units. A difference in recruit- inhibiting tension in the closely related, slow soleus,
ment order between two units in the table indicates Kanda and colleagues (192) used electrical or natural
that one of them could be silent while the other one stimulation of sural nerve afferents to alter the dis-
was discharging repetitively for at least 0.5 s. Table 3 charge patterns of MG motoneurons. The results were
shows pooled data from 23 motor units, analyzed in a variable in different preparations. Some low-threshold
manner similar to that used by Henneman et al. (159). units responding to stretch or vibration of the MG
The recruitment threshold for each unit was compared muscle exhibited brief slowing or cessation of firing at
to that for every other unit in the muscle. Thirty-four about the same time that higher threshold units, not
pairwise comparisons were possible. In a clear major- responsive to stretch or vibration, were powerfully
ity of pairs the smaller unit had the lower threshold. recruited by stimulation of sural afferents. As Figure
Exceptions were not rare, however. Unfortunately, 29 illustrates, the durations of the reversals produced
they varied in an unpredictable manner from cat to by sural stimulation were measured in tenths of sec-
cat and sometimes in the same preparation. Motor- onds. Activity in the lowest threshold unit (unit A)
unit pairs exhibiting some kind of sequence exception slowed and then ceased during the brief period when
in response to cortical stimulation generally showed the higher threshold units (B-E) were recruited. Units
454 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

Pinch

MG Filament

Unit
E

C
8
A

Sol EMG 2 mV

]
Sol Force
~ ~ !lOogm

. ._..·_.·_L·_.....
.
Time (0.1 .. 1.0 sec ) ~-
..l. ......I.l-..II'I.-.llft_.....-Jl,,_A.._......l.'-l,.. .......
=: J
~L.
FIG. 29. Records of EMG activity and force outputs of medial gastrocnemius (top 2 traces) and
soleus (bottom 2 traces) muscles of cat, together with simultaneous record of activity of a small
number of medial gastrocnemius (MG) motor axons in a natural fIlament of MG muscle nerve (3rd
trace). During MG vibration (160 Hz, 90 lUll) activity in whole MG increased and that in soleus (SOL)
showed a smaller increase. Pinching the skin over the lateral ankle with toothed forceps produced 2
bursts of MG activity, with recruitment of large amplitude motor axons in the MG fIlament, but
caused quite complete suppression of SOL activity. Isolated activity of 5 identifiable MG motoneuron
axons is shown in A-E. Note slowing of unit A discharge during each burst of activity in the higher
threshold units B-E. [From Kanda et al. (192).]

D and E began to fIre only during a 0.2-s pause in the orderly within different subgroups of a population.
activity of unit A. While the fIring behavior of unit A They add that "differential control is unusual and
was similar to the activity pattern in the whole soleus hard to demonstrate." As previously noted, Clamann,
muscle in Figure 29, the spike patterns in the higher Gillies, and Henneman (58) found no evidence that
threshold units (B-E) were unlike those of SOL mo- inhibitory inputs in general or sural inputs in particu-
toneurons or unit A. The authors interpret these rec- lar caused such differential effects. This study leaves
ords "as a clear and unambiguous example of differ- the impression that differential control plays minor
ential control of alpha motoneurons within a single roles of brief duration. At present, the fIndings have
motor unit pool." It should be noted that units B-E the greatest value "as clues for further investigation
were recruited in the order of their conduction veloc- of synaptic input" (192).
ities as Table 4 indicates. The authors point out that The CFL of motoneurons may also be altered dif-
the fIndings in no way suggest that recruitment order ferentially by certain inputs. The CFL was defIned
is at times chaotic, but rather may be intrinsically (subsection Critical Firing Levels of Motoneurons,
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 455

TABLE 4. Axonal Conduction guished by their firing patterns. About one-half of all
Velocities of Motor Units the motor units could be driven continuously by vol-
untary efforts, increased slowly in firing rate with
Conduction Distances, nun
Unit
greater contraction strength, and discharged for long
22 24 37.2 periods of time. Some of these "continuously firing
long-interval units" (c.l.m.u.s.) discharged regularly at
E 88
D 88 80 97 8-1O/s, while others did not attain regularity until
C 85 86 frequencies at 1O-15/s. During sustained maximum
B 79 80 89 efforts c.l.m.u.s. fired at 30-50/s and even more rapidly
A 73 70 81 in brief twitches of maximum rapidity and strength.
Gamma 28
About 10% of the motor units could not be driven
Velocities, in mis, measured at 3 different conduction distances continuously, but only for brief intervals. These "in-
in experiment shown in Figure 29. Distal electrodes fixed on muscle
nerve near entry to muscle; position of proximal pair varied.
termittently firing short interval motor units"
[Adapted from Kanda et a1. (192).] (i.s.m.u.s.) could not be driven voluntarily by weak or
moderate sustained contractions of constant strength
and did not fire at rates below 20/s. With strong efforts
p. 440) for monosynaptic responses to homonymous they discharged intermittently during steady contrac-
inputs. Since it correlates closely with motoneuron tions or in brief high-frequency bursts reaching 100/s
size (60), with recruitment order in mechanically or during rapid increases in contraction strength. Nearly
electrically elicited stretch reflexes, and with response one-half of the motor units had discharge properties
to various inhibitory inputs, it is regarded as a good somewhere between those of c.l.m.u.s. and i.s.m.u.s.
index of the rank order of a motoneuron with respect These "intermediate" units had various degrees of so-
to all other cells in the pool. Clamann and Kukulka called endurance, according to Grimby and Hannerz.
(61) have shown that more than two-thirds of MG Certain motor units could be driven continuously in
motoneurons showed a different CFL in response to the initial stage of an experiment but gradually lost
L 7 + SIDR stimulation than to stimulation of one this capacity after repeated contractions. The lack of
dorsal root (DR) alone. The same authors (63) have endurance noted by Grimby and Hannerz resembles
demonstrated that CFLs obtained by stimulating L 7 the rapid cessation of firing in large motoneurons of
+ SIDR are altered by electrical stimulation of Dei- cats whose muscle nerves are being stimulated repet-
ter's nucleus. This stimulation lowered the CFLs of itively at high frequencies, as noted earlier in this
motoneurons with axonal conduction velocities greater section. This phenomenon is not understood, but it
than 100 m/s and raised CFLs of those with conduc- may be due to a decrease in the effectiveness of the
tion velocities below 85 m/s. The effects may be due high-frequency input. It could hardly be due to failure
to inhibition of small motoneurons or facilitation of of contractile mechanisms, since the EMG does not
large cells. There is some evidence that both effects measure these.
occur. The significance of results achieved with arti- One c.l.m.u. and one i.s.m.u. were studied simulta-
ficial inputs that are far removed from natural ones is neously under many conditions. In sustained isometric
uncertain. The inability of human subjects to control contractions the c.l.m.u. was recruited before the
rank order voluntarily (subsection Evidence Regard- i.s.m.u. In a prolonged isometric contraction of con-
ing Voluntary Selective Control of Motor Units, p. stant strength or even during prolonged maximum
457) suggests that the inputs used in these experiments effort only the c.l.m.u. fired. On acceleration the
are not under voluntary differential control. i.s.m.u. was not recruited before the c.l.m.u. had at-
The most substantial group of studies on recruit- tained firing intervals as short as 30-40 ms, i.e., not
ment in human subjects are those of Grimby and before the c.l.m.u. had discharged near its maximum
Hannerz and their collaborators. They used EMG sustained frequency. On rapid acceleration the i.s.m.u.
recording techniques in a variety of muscles, reflexes, was recruited earlier. In twitch contractions selective
and voluntary movements. Various examples of rever- activation of the i.s.m.u. occurred if the muscle was
sals in recruitment order were described in normal relaxed before the twitch and minimum duration of
subjects and neurological patients. One of their most the twitch was intended. Some training was necessary
recent papers (139) is summarized here, because it before the subject was capable of adjusting his con-
seems to represent the culmination of skills, experi- traction so that only the i.s.m.u. was activated. Even
ence, and conclusions that 10 years of experimentation after intense training, however, the subject could not
have developed. References to their earlier studies are activate the i.s.m.u. selectively on each trial. The
found in this paper and in a more recent one by Borg authors (139) concluded that "under the experimental
et al. (23). In these two publications, the firing rates conditions used, the subjects were not capable of sys-
and recruitment orders of motor units in the short toe tematic activation of only i.s.m.u.s." In a subsequent
extensors of normal human subjects were investigated paper (23) it was shown that c.l.m.u.s. had conduction
in great detail. Two distinct types of motor units and velocities between 30 and 45 m/s; i.s.m.u.s. conducted
a large group of intermediate units could be distin- at 40-54 mls and intermediate units had intermediate
456 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

conduction velocities. Thus, a relationship was estab- corded before the Ls.m.u. (A). On prolonged maximum
lished between voluntary discharge properties and effort, only the c.l.m.u. responded (B). When sudden,
axonal conduction velocity (cell size). This finding rapid, and short contractions were intended, selective
indicates how closely these studies on human motor or predominant action of the i.s.m.u. occurred (C and
units resemble those on animals previously described. D).
As this summary indicates, there appear to be some Comparison of the latter condition with the 100-ms
differences in the results of Grimby and Hannerz (139) time mark indicates that these reversals were of very
and those of Desmedt and Godaux (90-92) previously brief duration, and the authors point out that their
described. These may only be due to differences in occurrence was "unpredictable." Whether the mech-
experimental conditions. Grimby and Hannerz believe anisms for reversal of recruitment order can be used
that in the rapid ballistic contractions of Desmedt and systematically, and have real functional utility, re-
Godaux isometric contractions were studied rather mains to be demonstrated.
than isotonic, as their paper states, and that this The recent experiments of Stephens and colleagues
accounts for the difference in their results. (330) have added a fascinating new observation to the
As mentioned earlier, both c.l.m.u.s. and i.s.m.u.s. study of recruitment in humans. Motor units in the
participate in ballistic isotonic contractions if the con- first dorsal interosseous muscle (1DI) were recorded
tractions are not made in any special way. If the along with the force produced by abduction of the
muscle is relaxed before a ballistic contraction and index finger. During normal voluntary abduction, mo-
both maximum speed and minimum duration of con- tor units were recruited into activity in a fixed order,
traction are intended, i.s.m.u.s. can be recruited briefly according to their size, as contraction was increased.
in isolation, i.e., according to the authors, "the order Subjects were unable to contract their 1DI in such a
of recruitment can be reversed." way that the unit with the normally higher threshold
If, however, the muscle is slightly contracted or of a pair discharged first or alone. Continuous stimu-
stretched before a ballistic contraction and lower force lation of the digital nerves of the index finger at 50/s
and a longer duration of contraction is intended, through ring electrodes around the finger, however,
c.l.m.u.s. can be recruited without i.s.m.u.s. According changed the order of recruitment in many cases. As
to Grimby and Hannerz, "a pre-existing contraction or stimulation was continued for 2-4 min the unit that
stretch of the muscle is necessary for studies of iso- normally had the lowest threshold (1) gradually de-
metric twitch contraction and this explains the ab- creased in firing rate until unit 2 was discharging
sence of recruitment order flexibility... " in the studies alone. When unit 1 stopped firing, the subject was able
of Milner-Brown et al. (258) and Desmedt and Godaux to keep unit 2 active alone at a contraction strength
(90). Figure 30 (139) illustrates the basis for their lower than the recruitment threshold for this unit in
conclusions. One c.l.m.u. and one i.s.m.u. were re- control conditions. Unit 1 could be made to fire, but
cruited. In sustained contraction, the c.l.m.u. was re- only if the force of contraction was greater than that

A I
..- ..-
It. II.
I'
~
I.....
Lt l

1'1
L

c D

FIG. 30. Simultaneous recording of one continuously fIring long-interval motor unit and one
intermittently fIring short-interval motor unit in human. A: increasing contraction. B: prolonged
maximum effort. C, D: rapid twitches. Time bar, 100 ms. [From Grimby and Hannerz (139).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 457

at which unit 2 fIred alone. The recruitment order was After an initial report of selective activation of single
thus reversed. According to the authors, the subject motor units in humans by Harrison and Mortenson
was quite unable to contract the muscle so that the (147), Basmajian (17-19) has repeatedly claimed that
previously low-threshold unit (1) fired alone. After when human volunteers are provided with visual and
stimulation was discontinued, several minutes were auditory feedback from their active motor units, they
required for the control pattern to return. As the effect easily learn within 10-15 min to control the activity of
of the stimulus wore off, unit 1 increased its fIring rate, any single unit within a given muscle. Since these
although the force exerted by the subject was decreas- reports disagreed with the majority of observations on
ing. Unit 2 could no longer be fIred alone at this time. animals and humans showing a relatively fIxed order
In further studies on the 1DI by Garnett and Ste- of recruitment, an attempt was made to determine
vens (119) each subject was asked to track a moving whether human subjects could indeed learn to exercise
target on an oscilloscope by moving the index fInger. selective control over their motor units. Subjects were
During each contraction, the recruitment threshold asked to discharge motor units in the extensor indicis
was defIned as that level of force at which the unit proprius muscle by dorsiflexing the index fInger. After
fIrst fIred continuously. During subsequent trackings, establishing the usual order of recruitment for two
the index fInger was stimulated via the ring electrode. motor units that were clearly distinguishable on the
In general, units recruited at less than 150 g (6% max oscilloscope, subjects were instructed either to reverse
force) had their thresholds raised by the skin stimu- the usual order or to silence the activity of unit 1
lation. Units with thresholds above 150 g had their without silencing unit 2. These were considered to be
thresholds lowered. As in the previous experiments, 2- simple but crucial tests of the claims for selective
4 min of stimulation was required to produce these control (158). All nine subjects were able, without
changes and 1-2 min passed before they had com- difficulty, to cause discharges of a single motor unit
pletely subsided. on the oscilloscope and to control its rate of fIring
The long delay in the onset of these effects, their from most recording sites. Slight voluntary increases
duration, and their slow subsidence are difficult to in muscle contraction usually resulted in an increase
explain. The results indicate that prolonged stimula- in the fIring rate of the fIrst motor unit and the
tion of cutaneous nerves may result in nonhomoge- recruitment of a second somewhat larger action poten-
neous distribution of inputs to a motoneuron pool. At tial that followed the same pattern as the fIrst. When
present, it is hard to put these fIndings in appropriate the subject was asked to reduce tension in the muscle,
perspective; it is even difficult to judge whether they the second unit was silenced before the fIrst and there
are physiologically normal. The fIrst dorsal interos- was a progressive decrease in the fIring rate of the
seous muscle is used in manipulative tasks, giving rise fIrst; the latter fInally stopped when the muscle was at
to cutaneous input of various types. However, brief complete rest. This pattern of recruitment could easily
stimulation of the digital nerves, such as might occur be reproduced in all subjects without special training.
normally, had little effect on recruitment in these In six of the nine subjects no changes in recruitment
experiments. A number of intriguing questions are, order were observed despite two hours of training and
therefore, raised by these fIndings. the help of audiovisual feedback. In each experiment
The foregoing review of evidence relating to alter- recordings were made from many sites, and the subject
native orders of recruitment indicates that this topic was encouraged to explore maneuvers that might lead
is not ready for any defInitive conclusions. Differences to alteration in recruitment. At each new site at least
in recruitment order have been demonstrated under 20-30 minutes was spent attempting to alter the nor-
circumstances that usually leave their signifIcance un- mal order. In not a single instance, out of hundreds of
certain. In general, they have been of short duration, trials, was anyone of these six subjects able to recruit
unpredictable in occurrence, and not under voluntary two units in their usual small-to-Iarge order and then
control. The unpredictability of departures from the turn off unit 1 without silencing unit 2. It should be
usual order contrasts with the predictability of the emphasized that even under almost isotonic conditions
standard order, as Grimby and Hannerz' studies em- the most minimal movements of the index fInger,
phasize. Further experimentation rather than specu- scarcely visible to observers, were associated with
lation is needed to clarify these problems. activity in both units. The CFLs of these pairs of
motoneurons probably differed only slightly, yet no
Evidence Regarding Voluntary reversals in recruitment order occurred.
Selective Control of Motor Units The results at almost all recording sites in the three
remaining subjects were similar to those just de-
To understand the signifIcance of deviations from scribed. In each of these subjects, however, there was
the usual recruitment order or of variability within it, one site at which some variability in recruitment order
it was important to establish the degree of selective was observed. Although one unit was recruited fIrst
control that the central nervous system could exercise and dropped out last in the majority of tests, the unit
voluntarily or even unconsciously over individual mo- that was usually recruited second was occasionally the
tor units. fIrst to respond and could then be activated repeti-
458 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

tively for some seconds without any activity in the and extracellular spike amplitude in motor neurons
first unit. These changes in recruitment order seemed which could be identified from animal to animal. He
to occur randomly. None of the subjects could, on found that these indices of motor neuron size were
demand, activate unit 2 at will or alternate the activity strongly correlated with each other and also with
of the two units in sequence. numerous functional properties of the motor neurons,
These studies agree with numerous reports that including adaptation to intracellular depolarization
human subjects can easily isolate single active units and neuromuscular effect, as well as recruitment se-
and control their frequency by simply changing the quence and discharge pattern during natural motor
contractile force voluntarily. This isolation is apparent output.
rather than real, however, and selective control could Lobster motor neurons are driven in rhythmic
only be demonstrated for the lowest threshold unit bursts of impulses by an "oscillator" in the correspond-
within recording range. There was no evidence of true ing abdominal ganglion. Coordinated movements of
voluntary or conscious control over the order of re- the four pairs of swimmeret appendages are brought
cruitment. These results demonstrate that the pattern about by interactions between these oscillators. The
of recruitment in human subjects is relatively but not oscillators, in turn, are controlled by identified excit-
absolutely fixed in the sense that it is not subject to atory and inhibitory "command" interneurons. Stim-
voluntary control. ulation of an excitatory command fiber in an isolated
In contrast to the results just described, a number nervous system elicits the rhythmic pattern of motor
of studies have been published (9, 18, 194, 341) sug- output that normally causes swimmeret beating. The
gesting that human subjects can activate motor units underlying neuronal mechanisms are, therefore, un-
selectively, provided they are given visual or auditory usually amenable to analysis.
cues signaling the occurrence of firing. To investigate The diameters of the somata were measured with
these claims Thomas et al. (337) recently studied three an ocular micrometer. They ranged from 25 to 125/Lm.
different muscles. In two, the abductor pollicis brevis Large cells invariably had rapid axonal conduction
and the extensor digitorum brevis, they were able to velocities and large extracellular impulses as illus-
demonstrate a significant number of reversals of re- trated in Figure 31. In neurons supplying a given
cruitment order. These are both muscles with more muscle, the relation between conduction velocity and
than one action. The reversals observed occurred only
between units with thresholds that differed by less 85}Jm
than 50 g and were accomplished by using the muscle
for movements in different directions. Reversals be- (2) 221m/sec 342m/sec 410m/sec
tween units with perceptibly different thresholds,
when activated along the same axis of contraction,
~~~
2mvl
were spontaneously reported to be "impossible" by all (3)
subjects. In the first dorsal interosseous muscle, which 2msec
has only one action, no reversals could be demon-
strated. "Thus, it appears," the authors conclude,
"that recruitment order reversals are possible to a v::--~ _~2mV(a.b)
10m~
limited extent in multi-functional muscles, but not in (4) ~=~~=== 1 20msec
single function muscles" (337).

Size Principle in Other Species

Recruitment of motoneurons in order of increasing
(5l~~iIr-
size by excitatory inputs of progressive intensity is not
limited to mammals. It has been observed in a wide ·:!tllW!II!lIIIIIIII!I!I!I! ,,""!!!!I!!!IIIIII'1!I11111111 -tWUl'llWIIIW'I»l'
variety of animals as phylogenetically distinct as cats (6)
I . !jj"I'l' 'ifi'i!'" 'II'" iill'l'lil'llll'lIli lilill
(160) and lobsters (85). Since its first description in
the cat (153), it has been reported in the natural motor
output to crab leg muscles (52) and eyestalk muscles
A B c
(165), as well as in cockroaches (277, 278), locusts FIG. 31. Properties of a small (A), medium (B), and large (e)
(168), dragonflies (257), mantids (301), water bugs motor neuron of lobster innervating the main power stroke muscle.
1, Soma diameters; 2, axonal conduction velocities; 3, amplitudes of
(342), newts (333), fish (169), chickens (113), skunks action potentials recorded with 2 extracellular electrodes at different
(340), and humans (115). positions on the motor nerve; 4, amplitudes of excitatory junctional
Davis (85) has made a comprehensive study of the potentials (EJPs) recorded from a single muscle fiber (top trace in
functional significance of motor neuron size in the each record); 5, adaptation to a maintained intracellular depolarizing
current; and 6, facilitation properties of extracellular EJPs (top
swimmeret system of the lobster. The advantages of trace in each record) during 50-Hz stimulation. Note antifacilitation
the invertebrate preparation enabled him to measure produced by the largest motoneuron. Time marks in row 5 (lowest
soma diameter directly, as well as conduction velocity trace), 0.1 ms. [From Davis (85).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 459

soma diameter was approximately linear. Since con- suggests that some rules governing the output of neu-
duction velocity is positively correlated with axon ronal networks are common to both simple and com-
diameter, soma and axon diameter were positively plex nervous systems (85). There is, moreover, sugges-
correlated as well. These indices of motor neuron size tive evidence that the size principle may apply not
were measured in dozens of experiments and found to only to motor neurons but to other neural cells as well.
correlate well with several functional properties of the Orderly recruitment of interneurons is apparent in
motor neuron. acridid auditory interneurons (304); the sensitivity of
A long-duration pulse of depolarizing current in- pacemaker cells is said to be inversely related to their
jected into a soma caused a series of action potentials size (33), and the size of sensory neurons in general
in the motor axon. Small motor neurons typically may playa role in determining their excitability (32,
responded with a train of evenly spaced impulses; in 357).
large motor neurons the interspike intervals became
progressively longer during the train. Accommodation Modulation of Firing Rate
at the site of spike initiation was presumably the cause
of this adaptation. Smaller motor neurons could also In 1929, Adrian and Bronk (3) noted that the force
be driven at higher maximum frequencies, an indica- of a voluntary contraction could be increased by two
tion that their refractory periods were probably mechanisms. The number of actively contracting mo-
shorter. It was regularly observed that the relative tor units could be increased ("recruitment") or the
increase in stimulus current needed to cause a given rates of the units already discharging could be accel-
increase in discharge frequency was related to motor erated ("frequency coding," "rate coding," "rate mod-
neuron size. That is, pacemaker sensitivity was appar- ulation"). Adrian and Bronk gave a major role to rate
ently inversely related to size. modulation. Recently recruitment has been empha-
The size of a motor neuron was also correlated with sized more because several investigators found little
its effect on the muscle fibers it supplied. The smallest increase in firing rate except near threshold (24, 128,
cells evoked the smallest excitatory junctional poten- 135). Moreover, the firing rate appeared to be stable
tials (EJPs), which on repetition displayed the most over a wide range of voluntary forces in humans (21,
neuromuscular facilitation (Fig. 31, line 4). The EJPs 57, 191).
of the largest motor neurons, in fact, nearly always Using the first dorsal interosseous muscle (lDI) of
showed antifacilitation (Fig. 31, line 6). These results humans, Milner-Brown et al. (259) averaged the
indicate that small and large motor neurons supplying steady discharge rate of single motor units at just
swimmerets have functional properties usually attrib- above the threshold required to recruit them. During
uted to slow and fast crustacean motor neurons (33). voluntary contractions over a wide range of forces, the
Increases in the frequency of stimulation of single relative importance of recruitment and rate coding
command interneurons caused a greater frequency of were compared. Subjects were asked to generate lin-
rhythmic bursts and also resulted in recruitment of early increasing or decreasing forces by moving the
additional motor neurons. This recruitment proceeded index finger laterally and tracking a triangular wave
in the order of increasing motor neuron size. The form on an oscilloscope. Single motor units began
thresholds of these motor neurons to natural synaptic firing at 8.4 ± 1.3 impulses/s and increased their rates
inputs, therefore, increased with their size as in ver- by 1.4 ± 0.6 impulses/s for each 100-g increment of
tebrates. When electrodes were placed so that the force. The mean firing rate did not depend on the
activity of all of the motor neurons could be monitored threshold for recruitment, although a relationship has
at the same time, it was apparent that with regard to been noted in other muscles (21,57). At intermediate
the recruitment sequence, at least, the size principle rates of increasing and decreasing force, firing varied
governs the entire population of swimmeret motor linearly with force over the entire range examined.
neurons, rather than simply the smaller groups of cells During slow increases in force, fIring rate tended to
innervating single muscles. Recruitment of motor neu- reach a plateau; during rapid increases an initial train
rons in order of increasing size occurred before and of impulses at a nearly constant rate occurred.
after abolishing all sensory feedback from propriocep- At each level the percentage of extra force due to
tors and other sense organs. Thus, the size principle in recruitment and that due to increased firing rate could
the swimmeret system does not depend on intact be computed for the population of units recorded.
reflexes, but results instead from central properties. Recruitment accounts for the majority of the increase
Davis (85) also showed that the temporal structure of at low levels. Although larger motor units were re-
bursts of motor neuron impulses depends strongly on cruited at higher levels, they were relatively few com-
the size of the motor neuron. He had pointed out also pared to the number of units already active, so that
that a number of other properties are correlated with the higher firing rate in the previously active units
the dimensions of cells, adding further support to the produced most of the extra force. Over the range
use of the more inclusive term size principle to de8cribe considered, increased rate produced about two-thirds
the many implications of motor neuron size. and recruitment only about one-third ofthe total force
The widespread distribution of the size principle accounted for.
460 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

Wani and Guha (343) have proposed a mathematical ship between recruitment and firing rates seen in all
model based on these findings of Milner-Brown et al. recordings, however, makes it unlikely that the soleus
(259) and on the contractile characteristics of motor pool has these alternatives. For questions of motor
units in animals (253, 353) and in humans (29,57). The control, it is important to understand which mecha-
model finds the twitch tension, firing frequency, and nisms may be responsible for the amount of rate
tension of any motor unit in the muscle at any instant modulation that occurs in parallel with recruitment.
during the gradation process. Expressions are also Kernell (204) has been analyzing this problem with
derived for finding total tension, tension due to re- models containing only five motoneurons. The factors
cruitment, and tension due to rate coding after the that influence the relation between recruitment and
recruitment of any motor unit in the muscle. The rate modulation are treated theoretically. They in-
model quantifies the interplay between recruitment clude the slope of the frequency-current relation of
and rate coding. The characteristics of the motor units the motoneurons, the distribution of recruitment
of the 1DI, evaluated with the model, closely resemble thresholds, the effects of recurrent inhibition, and
those obtained experimentally by Milner-Brown et al. other factors. If appropriately distributed, for example,
(259). a strong recurrent inhibition may have a very marked
In another very small muscle (the first deep lum- effect on the rate-recruitment relation.
brical of the cat) similar observations have been made More recently Hatze (149) has made a theoretical
(206). The tension produced by lumbrical motor units study of the contributions of recruitment and rate
may easily be more than doubled by synaptically coding under static conditions. An "activation sensi-
induced changes in firing rate. Kernell and Sjolholm tivity index" is derived from an "excitation variable,"
(206) go through a brief formal argument showing that and it is stated that this index is a "precise indicator
if the motor units of a muscle behave like lumbrical of the relative importance of the two control mecha-
motor units, the importance of rate modulation would nisms for any mode of contraction." The index is
remain valid regardless of the number of motor units applied to both experimental and stimulation results.
in the muscle. The author reports that the index "confirms earlier
It may be assumed that these findings apply to large and more intuitively based conclusions that recruit-
muscles, though with some modifications. In the lum- ment is the major mechanism responsible for the
brical muscle of the cat rate modulation is important grading of the force output at low levels (up to about
before all motor units are recruited. During a stretch 30% of the maximum tetanic force), while at interme-
reflex in the cat soleus muscle, however, a gradual diate and high force levels rate coding is the dominant
recruitment may apparently occur without much mod- factor." Thus, experiments, models, and mathematical
ulation of firing rates (89, 128, 135). Even a modest studies seem to be in general agreement.
amount of rate modulation, in the range of 5-10 im-
pulses/s, would have a significant influence on the ORGANIZATION OF INPUT TO
strength of a tonic reflex in the muscle, as Kernell MOTONEURON POOLS
(204) points out.
The widespread distribution of la-terminals to prac- The firing patterns of motoneurons and motor units
tically all homonymous motoneurons (255) is an im- described in the previous section are determined by
portant factor in rate modulation. Mendell and col- the intrinsic properties of the motoneurons and by the
leagues (269, 270) have shown that this is common in distribution of input to them. Each motoneuron re-
limb muscles (see section ORGANIZATION OF INPUT TO ceives thousands of terminal boutons from primary
MOTONEURON POOLS, this page). Any increase in syn- sensory fibers, spinal interneurons, and supraspinal
aptic excitation to the pool would, thus, occur simul- fibers. In spite of the great diversity of input, both
taneously in all motoneurons, including those already excitatory and inhibitory, that they receive, motoneu-
firing. In a discharging motoneuron this would tend to rons innervating the same muscle function as collec-
cause an increase in firing rate. Therefore, an increase tive entities. The activity of each cell in a pool is
in the number of recruited motoneurons would com- closely correlated with the activity of all the other
monly be expected to be accompanied by some in- cells in the pool.
crease in firing rate among all the discharging mem- The purpose of this section is to review and evaluate
bers of the pool. This is what has been observed. recent studies detailing how inputs are distributed to
In a contracting soleus muscle, as Kernell (204) motoneurons. There is little emphasis on synaptic
points out, a given length change tends to produce a mechanisms, which are well covered in recent reviews
greater relative change in tension at low stimulus rates by Burke and Rudomin (44) and Redman (295).
than at higher ones (132,249,284). Thus, according to Many of our present ideas concerning convergence
Grillner (132) the stiffness t:.F/ t:.L (where F = force and divergence in the nervous system, particularly in
and L = length) of a soleus contraction would, in the spinal cord, derive from the work of Cajal, Sher-
theory, be greater if it were caused by many units rington, and Lloyd. Cajal (291) provided evidence that
firing at slower rates than if it were produced by fewer single sensory fibers branch and send terminals to
units discharging at higher rates. The close relation- several motoneurons. From physiological studies,
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 461

Sherrington and colleagues (78, 321) inferred that important to the problem of distribution of la-fiber
many inputs converge on the motoneuron pool supply- connections to motoneurons. The discussion of this
ing a given muscle. Simultaneous activation of differ- subsection is qualitative; a more quantitative treat-
ent converging inputs could evoke responses that were ment is given in subsection Correlations Between
larger than the sum of the separately evoked reflexes. Morphology and Function, p. 476, where physiological
This summation was interpreted as the result of sub- and anatomical results are correlated.
liminal activation of some motoneurons by each input
alone. Additive effects in motoneurons, due to "over- THE MOTONEURON. Motoneurons are large cells in the
lap" of "subliminal fringes," were believed to bring ventral horn whose axons exit the spinal cord via the
ventral roots and run in peripheral nerves to supply
additional motoneurons to threshold.
A convenient way to begin this discussion of input muscle fibers. They include both a-motoneurons
to motoneurons is to describe recent findings on group which innervate extrafusal muscle fibers responsibl~
~or muscular force, and y-motoneurons, which supply
la-fibers that transmit impulses directly to motoneu-
rons. The inference of monosynaptic contact first be- mtrafusal muscle fibers of the muscle spindle. The la-
came possible after Lorente de No (239) had shown fibers project monosynaptically to a-motoneurons but
that the synaptic delay at the motoneurons in the not to y-motoneurons (98). This account is therefore
restri~ted to a-motoneurons and synapses upon them.
third cranial nucleus varied from 0.5-0.6 to 0.8-0.9 ms.
Renshaw (296) demonstrated that there were spinal An Important feature of the motoneuron is its very
reflex arcs consisting of only two neurons (one synaptic large surface area, which has been estimated in dye-
delay). Lloyd (235) went on to show that la-fibers filled neuro~s as ranging from 79,000 to 250,000 J.Lm2
from spindles were the only primary sensory fibers to (14). Accordmg to these authors 80%-90% of this sur-
make monosynaptic contacts with motoneurons. face area is dendritic.
These early conclusions have been amended since Motoneurons have 5-20 dendrites, each of which
recent studies revealed the existence of monosynaptic divides freq.uen~ly, with daughter branches tending to
connections from group II fibers supplying secondary be smaller m dIameter than the parents. The details
receptors in spindles, as described in subsection Group of this branching have become important in the de-
II Input From Secondary Endings in Muscle Spin- velopment of mathematical models of dendritic func-
dles, p. 479. tion (see ref. 288 for review). Dendrites are oriented
The introduction of intracellular recording by Eccles predominantly in the rostrocaudal direction. These
and collaborators (25) permitted direct access to the branches often extend more than 1 mm in either
excitatory and inhibitory events occurring in motoneu- direction from the cell body (87, 311, 332). Dendrites
rons. Intracellular recording revealed that the excit- from different cells are often arranged in bundles (251,
atory event was a depolarizing potential, an excitatory 311, 312). The functional significance of this bundling
is unknown.
postsynaptic potential (EPSP). When a muscle nerve
was stimulated, the amplitude of the EPSP was deter- All dendritic regions as well as the soma receive
~ined in part by the number of la-fibers conducting
synaptic connections; there are no regions devoid of
Impulses to the motoneuron (99). This subject and synapses (122, 354). Some order has been found in the
corresponding findings for inhibitory postsynaptic po- termination of various presynaptic systems on moto-
tentials (lPSPs) are discussed extensively from an neurons (218); however,· motoneurons do not receive
electrophysiological point of view in the first volume a highly topographic, layered series of connections
of this Handbook section on the nervous system (44) from various sources as do cerebellar Purkinje cells,
for example.
and are not treated here.
. The early intracellular investigations of synaptic ANATOMY OF la-BRANCHES TO MOTONEURONS. The la-
mput used electrical stimulation of peripheral nerves. fiber has a very extensive series of central connections
They allowed estimates of the number and potency of including cells of Clarke's column, spinal cord inter~
la inputs from various sources to different motoneuron neurons, and a-motoneurons (44). Only those branches
pools (99, 103). The extent to which la action is focused terminating on a-motoneurons are of concern here.
upon particular motoneurons began to become appar- Early studies at the level of the light microscope (180,
ent. To understand how individual motoneurons func- 327,331,334) gave confusing results, perhaps because
tion as parts of an ensemble, however, it must be t~e transverse sections used failed to provide a suffi-
known how the terminals of afferent fibers are distrib- CIently comprehensive view of the dominant rostro-
uted to the members of the pool. It was necessary, caudal polarity of the motoneuron. Similarly the la-
therefore, to study the effects of impulses in single la- fiber, which also has an important rostrocaudal com-
fibers on individual motoneurons and to determine the ponent, could not be viewed in its proper perspective
distribution of these effects. in transverse sections.
Schei?el and Scheibel (310), using Golgi techniques
Anatomical Studies
and sagIttal sections in kitten spinal cords, however,
In this subsection recent anatomical findings are were able to reconstruct the entire intraspinal course
surveyed with special emphasis on features that are of la-fibers. They described how afferent fibers bifur-
462 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

cate upon entering the spinal cord from the dorsal root la-SYNAPSES ON MOTONEURONS. This system contrib-
and run in both the rostral and caudal directions in utes only a small proportion (fewer than 5%) of the
the dorsal columns. Several branches are given off total number of boutons on these cells (71, 72, 252; see
from the dorsal column and drop ventrally through subsection Correlations Between Morphology and
the medial part of the dorsal hom and into the ventral Function, p. 476). This has complicated the analysis,
hom (Fig. 32), where their collaterals are organized particularly from the point of view of statistical relia-
primarily in a mediolateral plane. Each collateral pro- bility (for review see ref. 73 and also ref. 212). Conradi
vides terminal branches to the deeper layers of the (71), using ultrastructural methods, has described the
dorsal hom and also to motoneurons in the ventral bouton population on motoneurons and estimated
hom. It has since been learned that group II fibers which types are contributed by group la-fibers. Only
also project into the ventral hom (117) and make type M and larger type S boutons, which are 3-7 JLm
monosynaptic contact with motoneurons (subsection in diameter and contain clear spherical vesicles, be-
Group II Input From Secondary Endings in Muscle come fewer in number following dorsal rhizotomy;
Spindles, p. 479); la-fibers, therefore, cannot be iden- degenerating boutons were not found sufficiently often
tified simply by the presence of a projection into the in this study to warrant quantitative analysis (but see
ventral hom. ref. 222). Most boutons (e.g., types T, P, C, and smaller
The anatomy of single la-fibers has recently been type S) survive this procedure, indicating that their
examined in much more detail by staining single iden- axons have not been injured and, therefore, are not
tified cells with HRP (26, 51, 163). Reconstructions derived from sensory fibers. Conradi (71) found that
from either cross sections (26, 163) or sagittal sections type M and S boutons of sensory origin were localized
(51) reveal a picture (Fig. 32) that is qualitatively primarily on the proximal dendrites, where they were
similar to the Scheibels' description. Several collat- estimated to make up about 0.5% of the total (see also
erals from each la-fiber leave the dorsal columns and ref. 212). In a similar study, McLaughlin (252) found
drop down into the ventral hom to make contact with only type M boutons, which also were most concen-
motoneurons. Rostrocaudal spread of individual col- trated on proximal dendrites of motoneurons. An ad-
laterals within the ventral hom was limited (Fig. 32). ditional factor that may further decrease the total
The terminal fields of individual collaterals did not number of boutons on distal dendrites is the dimin-
overlap. A more quantitative discussion of these find- ished surface area per unit of length in distal regions
ings is deferred until subsection Correlations Between ofthe motoneuron (14). These regional ultrastructural
Morphology and Function, p. 476, where both physi- studies are extremely difficult, due to the widespread
ological and anatomical results can be evaluated to- dendritic tree and the heterogeneous inputs, only a
gether. However, a reasonable conclusion at this point small portion of which arise from the afferent system
is that the morphology of la-fibers allows each of them of interest.
access to large numbers of motoneurons. Studies on the identity of group la-boutons and the

FIG. 32. Morphology of la-collaterals to motoneurons and arrangements of boutons on dendrites.

A: sagittal section through lumbosacral cord of 40-day-old kitten (section shown at right). Primary
fibers in dorsal columns (a) send ventrally directed branches to ventral horn (b) to terminate in region
of motoneurons (c and d) that have sagitally running dendrites. Primary afferent branches also give
off branches to dorsal horn and intermediate zone; these are described in more detail in the original
illustration of Scheibel and Scheibel (310) from which this figure is modified. B: scale drawing
illustrating morphology of collaterals of a single triceps surae la-fiber as revealed by horseradish
peroxidase (HRP) filling. This 3-dimensional representation provides both a cross-sectional (trans-
verse) and sagittal perspective. The la-fiber in the dorsal columns is drawn as a thick black line along
with the dorsal root ganglion cell and dorsal root portion of the cell. Four collaterals emanating from
the dorsal column region are illustrated, each one coursing through the dorsal horn (in a ventral-
lateral and rostral direction) to terminate in the cylindrical motoneuron region (lamina IX). Other
terminals are given off by the same fiber in laminae VI and VII. C: morphology of a single soleus la-
fiber in the dorsal columns with 9 branches given off to ventral horn; HRP technique. Two
motoneurons-soleus (SOL), which innervates type S motor unit, and lateral gastrocnemius (LG),
which innervates a type F, fatigue-resistant motor unit (R)-were also filled with HRP. Circles and
stippling denote somata and maximum dendritic extents of these neurons. Note that at most only 2
of 9 la-fiber branches have access to each of these motoneurons. Also ventrally directed collaterals of
this fiber do not exhibit rostral component shown by Brown and Fyffe (Fig. 32B). Top and bottom
solid lines denote dorsal and ventral border~ of cord. Dotted lines represent dorsal and ventral limits
of ventral horn. D: origins of primary collaterals of MG, LG-S, and posterior tibial (PT) la-fibers. E:
schematic drawing of disposition of la-fiber boutons on a-motoneurons. In each case, both la-fiber
and motoneuron filled with HRP. Note in SOL motoneuron that bouton A is on proximal dendrites
in contrast to B-F, which are on distal dendrites. Terminals on LG-type FR motoneuron are of en
passant variety (triplet) and of terminal variety. [A, adapted from Scheibel and Scheibel (310); Band
D, from Brown and Fyffe (26); C and E, from Burke et al. (51).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 463

distribution of this population on the motoneuron fibers terminate on single motoneurons. For example,
surface are fundamental to our understanding of la- different types of motoneurons might have systemat-
motoneuron synaptic transmission. To further analyze ically different locations for la inputs, which would
the mechanisms by which input can lead to highly contribute to the order in which they are recruited. In
organized output behavior, however, we must also be order to study the termination of single la-fibers on
able to describe where the boutons from single la- motoneurons it is necessary to label them with some

o CIudIl ROIlrol

MG I I I
I tI
I

i I

LG·S
I
I ~
i i I
!It
I

I
~'

PT I I
I~~,
I

11 '
I
I I
+ I

POIitton of
8 --L do<ul . _ onlry
~ PoUtm of

+ col.......
Injection IiltI

SOL 10

E
>--i
Imm

SOL 10

c SOLEUS
GROUP 10
AFFERENT ~,

. J
LG - FR

SOL 10

SOL-S LG-FR
)=¢f=====
IL7 LG-F(FF?)

mm
464 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

marker, so that their boutons can be identified. A of a muscle nerve dissected until it apparently con-
particularly thorny problem in such an analysis is the tained a single la-fiber. Burke and Nelson (42) inves-
relatively extensive dendritic tree of the motoneuron, tigated rhythmic EPSPs recorded intracellularly dur-
which makes it difficult to sample the entire postsyn- ing stretch of the homonymous muscle. These EPSPs
aptic surface available to the la-fiber. Iles (177) labeled exhibited characteristics indicating that they were
small groups of dorsal root fibers with cobalt and evoked directly or indirectly by impulses in single la-
studied the terminals given off by single C0 2 + -filled fibers (i.e., silence during muscle contraction, in-
fibers to single motoneurons. These studies were nec- creased firing rate as a function of stretch, regular
essarily limited to regions near the soma, since the discharge), but they could be detected in only 10% of
motoneurons were only lightly stained and dendrites the motoneurons that were penetrated. Letbetter et
could not easily be followed. The boutons tended to a1. (232) recorded EPSPs evoked by intracellular stim-
be clustered, although this impression may have been ulation of la-cell bodies in dorsal root ganglia. Several
biased by the relative difficulty in seeing boutons other investigators have stimulated peripheral nerves and
than juxtasomatic ones. have severed dorsal roots progressively to achieve
An important advance in this analysis has been single fiber inputs. Others have worked with filaments
recently published by Burke et al. (51). Horseradish containing a small number of la-fibers and adjusted
peroxidase was injected into a physiologically charac- the peripheral shock to stimulate only a single fiber
terized la-fiber. Similar techniques were used to inject (176, 182). In order to resolve these small EPSPs
HRP into identified motoneurons (homonymous or clearly from other synaptic noise, averaging tech-
heteronymous in relation to the la-fiber) at a different niques have been used.
time during the same experiment. The sagitally s~c­ One of the most useful methods is to trigger an
tioned spinal cord was then processed to identify the averaging computer with single la impulses recorded
HRP reaction product, which is easily visible. Both la- from a dorsal root filament. Only EPSPs that are time
fiber terminals and motoneuron dendrites fill very locked with these impulses are resolved; all other
extensively, perhaps completely. Since synaptic con- intracellular activity is averaged out (Fig. 33). This
nections occur between la-fibers and a high proportion technique does not require a dorsal root filament with
of homonymous and heteronymous motoneurons (sub- only one active la-fiber. In fact, the presence of two
section PERCENTAGE OF MOTONEURONS RECEIVING active la-fibers in a filament allows the investigator to
TERMINALS FROM SINGLE la-FIBERS, p. 473), this tech-
trigger first off one afferent and later the other and,
nique has a high yield even though no physiological
thus, to study the EPSPs produced by two different
evidence is obtained for connectivity between the filled
la-fibers in the same motoneuron (256, 315). This
afferent and motoneuron. The results were quite dra-
technique, first used by Mendell and Henneman (254,
matic: a single la-fiber sends several collaterals into
255), is called spike-triggered averaging (STA). Watt
the ventral horn, only one making contact with the et al. (345) and Nelson and Mendell (269) have devised
single motoneuron whose dendrites are easily visible. techniques to ensure before beginning intracellular
Most boutons are localized close to one another but in recording that fibers have not been damaged central
one such case (of three reported), the la-collateral was
to the recording point during dissection.
observed to terminate with several boutons on differ-
ent closely related dendritic branches and also to send EPSPS RECORDED FROM POPULATIONS OF MOTONEU-
a process terminating in a single bouton close to the RONS. The limitations of studying EPSPs in single
soma (Fig. 32). These results are presently limited to motoneurons have recently been removed by a new
light microscope observations where synaptic contact technique introduced by Luscher, Henneman, et a1.
can only be inferred. Furthermore, no correla,tive in- (244), which permits recording of EPSPs from large
formation is yet available on the shape of the EPSPs
populations of cells. By infiltrating a ventral root with
produced by la-fibers whose bouton numbers and
locations are known (subsection PHYSIOLOGICAL ANAL- isotonic sucrose solution so that its extracellular re-
YSIS OF LOCATION OF BOUTONS, p. 468). sistance is extremely high, the interiors of its axons
Despite the exciting progress of these anatomical can be used as wick electrodes to record voltage
studies, much of the current knowledge concerning changes in all the motoneurons of that spinal segment.
the details of la-fiber connections to motoneurons With an averaging computer triggered by the action
comes from physiological studies. potentials from a functionally isolated, but intact af-
ferent, the synaptic effect of a la impulse upon a large
Techniques Used to Study EPSPs Elicited population of motoneurons can be extracted from the
by Impulses in Single Afferent Fibers physiological and electrical noise in the recording sys-
EPSPS RECORDED FROM SINGLE MOTONEURONS. Kuno tem. Some of the early results of studying input with
(215) was the fircit to study EPSPs that were elicited this technique are described in subsection FACTORS
in motoneurons by impulses in single afferent fibers. RESPONSIBLE FOR VARIABILITY IN EPSP AMPLITUDES,
This was done by electrical stimulation of a fine strand p.466.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 465

c
0-1
EJ
FJ'
r --
G

FIG. 33. A: arrangement for spike-triggered averaging. Muscle-stretch activated group la impulse
recorded from dorsal rootlet (a) is used to trigger an averaging computer, using a pulse-height analyzer
to select impulses of constant height produced by a single fiber. Each trigger pulse causes averager to
process the next several milliseconds of signal recorded from the microelectrode impaling the
motoneuron. Motoneuron is identified by antidromic invasion following muscle nerve stimulation (b).
If la-fiber acting as source of trigger signal sends terminals to motoneuron, then an EPSP is averaged.
B: examples of single 20-ms records (sweeps) from a medial gastrocnemius (MG) motoneuron triggered
by an MG group la impulse. The EPSP is seen at start of sweep superimposed on highly variable
synaptic activity. Note variability in amplitude of la-evoked EPSP. C-F: 4, 16,64, and 256 sweeps are
averaged, respectively. Note improvement in resolution of EPSP as number of sweeps averaged is
increased. Only potentials synchronized with the MG afferent are preserved. Others are asynchronous
with respect to the trigger signal and so are averaged out. G, H: dorsal root recording (G) and
simultaneous intracellular recording from motoneuron (H) at slower sweep speed. Note 2 la impulses
in G and synaptic noise in H. Calibration, 500 p.V for B-F, and H; 0.7 ms for B-F; 5 ms for H. [A,
adapted from Mendell and Henneman (255).]

Amplitudes of EPSPs Elicited by this variability is itself subject to some control (305).
Impulses in Single Fibers Whatever its cause, reliable comparisons can only be
made between averaged EPSPs.
VARIABILITY OF EPSP AMPLITUDES. Consideration of The EPSPs obtained by electrical stimulation meth-
the results in studies performed in different laborato- ods (176, 182, 219) are much larger on the average
ries with the various methods described above, indi- than those obtained using spike-triggered averaging
cates that the average size of EPSPs elicited by im- (254-256, 266, 269, 315, 345). Examination of typical
pulses in single fibers depends on the techniques used amplitude distributions for each technique (Fig. 34)
to measure them. It is important to recognize that the reveals important differences at both ends of the spec-
EPSP at any synapse does not have a fixed amplitude trum; spike-triggered averaging yields no EPSPs larger
but varies in size from stimulus to stimulus (42, 215, than 800 p.V, with more than one-half of them being
256). Whether this variability is due to differences in smaller than 100 p.V. In contrast, very few EPSPs
transmitter release from terminals (34, 215, 256), in- evoked by electrical stimulation are smaller than 100
termittent conduction in presynaptic terminals (106), p.V and they range in amplitude up to and beyond 2
or the superposition of synaptic noise on a constant mV.
EPSP (105) is not yet clear (see ref. 44 and subsection Differences between the EPSPs obtained by spike-
Physiology of la-Terminals, p. 472). The extent of triggered averaging and electrical stimulation might
466 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

100 A
80 N=342

60
40
20
z
'0 0 + ++
~
B
0

100
80 N=590
60
40
20
0 + ++
0 400 800 1200 1600 2000 >2500
Ampl itude (microvolts)
FIG. 34. Amplitude histograms of single-fiber EPSPs in homonymous and heteronymous moto-
neurons. A: histogram obtained using electrical stimulation method. B: histogram with spike-triggered
averaging method. Arrows represent values of less than 1% of N. Afferents and motoneurons are
mainly from triceps surae and semitendinosus muscle groups in A and exclusively so in B. [A, from
Jack et al. (82); B, from data of Mendell and Henneman (255), Nelson and Mendell (269, 270), Scott
and Mendell (315).]

be explained by a number of factors. For example, a which recordings were made. It is also possible that
criterion used in the studies with electrical stimula- these EPSPs were elicited by an interneuron that was
tion, namely that the EPSP exhibit a single time driven rhythmically by la impulses.
course with no discontinuities, might fail occasionally, The wide range of mean EPSP amplitudes described
so that the EPSP would be elicited by impulses in above may be due to the use of different techniques.
more than a single fiber. This would account for the It is worth reemphasizing that the spike-triggered
few relatively large EPSPs, but unless this were a averaging technique provides the most complete as-
common occurrence it would not explain why so few surance that only a single fiber is generating the EPSP
averaged EPSPs of less than 100 J..L V were recorded and that it is a la-fiber. This technique has consis-
under these conditions. However, this method selects tently yielded EPSPs smaller than 800 J..L V in animals
la-fibers with the lowest electrical threshold and im- with intact spinal cords. Larger EPSPs in such prep-
pulses in fast conducting fibers have been shown in arations have only been obtained using the other
other experiments (255) to evoke the largest EPSPs techniques. Despite these differences there is general
(see the next subsection). Furthermore, it seems pos- agreement that most EPSPs are 200-300 J..L V or less.
sible that motoneurons would receive less interneu- Their small size indicates that a considerable la con-
ronal and afferent input under these conditions than vergence is required to discharge motoneurons. How-
during spike-triggered averaging, when the muscle is ever, under normal conditions the motoneuron mem-
stretched and a large fraction of the dorsal rootlets are brane potential might hover close enough to threshold
still intact. The synaptic potentials produced by these so that input from a few la-fibers would be sufficient
other inputs might tend to act as a shunt, thereby to discharge it.
reducing the amplitude of EPSPs evoked by the la-
fiber under study (15, 38). FACTORS RESPONSIBLE FOR VARIABILITY IN EPSP AM-
The rhythmic EPSPs recorded by Burke and Nelson PLITUDES. Even when variability introduced by the
(42) had mean amplitudes ranging from 680 J..LV to use of different techniques is eliminated, a wide range
1,570 J..LV. These may have been the largest EPSPs of EPSP amplitudes is observed. Certain factors influ-
evoked by la-fibers, perhaps the only ones large encing EPSP amplitude have been identified. Mendell
enough to be detected in single sweeps. They were and Henneman (255) reported that the mean EPSP
detected in only about 10% of the motoneurons from amplitude evoked in experiments in which large affer-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 467

ent fibers (measured by conduction velocity) were 2.5

stimulated, was larger than in experiments when small
afferents were stimulated. This finding has been ex-
tended by Henneman and colleagues (244), who have
recorded the electrotonic potential from an entire
ventral root that had been perfused with isotonic
sucrose (subsection Techniques Used to Study EPSPs 2.0
Elicited by Impulses in Single Afferent Fibers, p. 464).
The root was held under constant conditions, during
which the projections of many la-afferent fibers were
studied using the spike-triggered averaging technique.
The size of the response, which reflects both the
-31.5
>
number of motoneurons activated as well as the am-
plitudes of their individual EPSPs, was found to be a..
systematically greater for large la-fibers than for small a..

-
en
ones within a given experiment (Fig. 35). These results a..
suggest that larger sensory fibers exert more potent o
synaptic actions, perhaps by giving off more terminal ~1.0
branches. A further example of the capacity of differ- ='
.=
ent afferent fibers to evoke EPSPs of different ampli- a.
tudes is provided by lateral gastrocnemius la-afferents, E
which evoke larger EPSPs on the average than medial
«
gastrocnemius, soleus (315), or semitendinosus la-af-
ferent fibers (269) despite similar conduction veloci-
ties.
05 1T
• •
Mendell and Henneman (255) also found that when
experiments were performed in high spinal, unanes- I~l ·
thetized preparations, the EPSPs were on the average '1 1
about twice as large as in intact, unanesthetized prep-
arations. This was attributed to the lack of anesthetic; 0.0 ............'-----l----L--L--'---'---'-.....J--'--'-->----'--....L.-...L..-~
however, recent experiments (268) suggest that acute 40 50 60 70 80 90 100 110
transection of the spinal cord in the anesthetized prep- Conduction velocity of afferent fibre (m/sec)
aration can also lead to considerably larger EPSPs via FIG. 35. Plot of mean amplitudes of postsynaptic population
mechanisms which are still obscure. These results potentials (PSPPs) elicited by afferent impulses with mean condi-
underscore the importance of the experimental con- tion velocities shown on abscissa. The PSPPs were evoked by
ditions in determining EPSP amplitude. impulses in single group la and II fibers of cat medial gastrocnemius
The existence of pre- and postsynaptic sites at which muscle. using spike-triggered averaging. Responses recorded from
1st sacral ventral root and lower part of 7th lumbar ventral root.
EPSPs evoked by la-fibers can be influenced is now The 153 data points were grouped in pentads of conduction velocity
well accepted. Stimulation of various segmental (101, and mean and standard deviations of PSPPs for each group are
216) and supraspinal areas (53, 54) results in depolar- shown. Mean data points fell into 2 groups, each of which was fitted
ization of la-terminals and depression of EPSPs with- with a straight line by method of least squares, using point at 72.5
out any change in their time course. This is probably mls for both calculations. [From Luscher, Henneman, et al. (244).]
due to presynaptic inhibition (314). Whether there is
diminished release of transmitter due to presynaptic (203). Several investigators (36, 220, 255) have ob-
depolarization (195) or prevention of impulse propa- served that EPSPs are larger on the average in small
gation at terminal branch points by cathodal block motoneurons that have larger input resistances than
(167), or both, is not known. The EPSP amplitude in large ones with small input resistances (Fig. 36).
might also be diminished by shunting if it were evoked Further evidence for the role of cell size is that
during an afferent barrage in other fibers converging individual EPSPs are larger in soleus than in medial
on the same motoneuron (15, 38). These mechanisms or lateral gastrocnemius motoneurons (315). This fac-
are discussed in more detail in Burke and Rudomin's tor undoubtedly contributes to the greater reflex ex-
review (44). For the present, it is sufficient to note citability of small motoneurons than large ones (see
that alterations on either side of the synapse can affect section HOW SIZE OF MOTONEURONS DETERMINES
the amplitude of the EPSP and that differences in THEIR SUSCEPTIBILITY TO DISCHARGE, p. 491).
descending, segmental, or propriospinal activity deter- The use of axonal conduction velocity to assess
mined by the type of experimental preparation might variations in EPSP properties in motoneurons of dif-
cause variations in EPSPs. ferent sizes is justified by the substantial evidence that
Another source of systematic variability in EPSP recruitment order is a function of motor axon diame-
amplitude is the input resistance of the motoneuron ter, which is proportional to axonal conduction veloc-
468 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

600
A
. B
Burke et al. (46) have demonstrated that motoneurons
supplying muscle units of various types generate ag-
:; gregate EPSPs whose amplitudes appear to depend
.
-=400
more on motor unit type (obtained from intracellular
to
stimulation of the motoneuron and measurement of
... .
"0
:J
.4;;:
Q. 200
the resultant muscle contraction) than on axonal con-

. . .. · . ...
E
<l
duction velocity. Although motoneurons in type FF
(fast, fatiguing) motor units have conduction velocities
similar to those supplying type FR units (fast, fatigue
o
C 0 resistant), the aggregate EPSPs are apparently smaller
in the former. This subject is discussed in greater
~400 detail in the chapter by Burke in this Handbook.
to •
. Boutons of la-Fibers on Motoneurons
"0
:J
:t

··· . ..
Q. 200
E PHYSIOLOGICAL ANALYSIS OF LOCATION OF BOUTONS.

. .
<l
The time course of individual EPSPs is a good indi-
cation of the location of the active boutons on moto-
o 90 95 100 105 110 90 95 100 105 110 115 neurons. Analysis of aggregate EPSPs (36) is compli-
Conduction velocity of motor axon (m/sec) cated by asynchrony in the arrival of the afferent
1·0 impulses as well as by differences in the location of
0·9
• the active synapses. In order to use EPSP time course
to infer synaptic location precisely, certain information
~ 0·8 concerning the electrical properties of the cell soma
g and its dendrites as well as the time course of the
~
4>
0-7 synaptic current must be known. These topics have
;:l
:':: been treated at length in reviews by Rall (288) and by
-a 0·6 Burke and Rudomin (44) in the first volume of this
E
III Handbook section on the nervous system and also in
il<
00 O·s. a recent review by Redman (294); they are discussed
il<
>il only briefly here.
E 0·4 Rall (286-288) has developed an electrical model of
d
III
:.a the motoneuron in which the dendrites are repre-
...
Ql
0·3
sented as cylinders divided into compartments of equal
"" 0·2 electrotonic length. The EPSPs produced by activa-
tion of single compartments and also by simultaneous
0·1 or sequential activation of several compartments have
been computed (286). The EPSPs generated by the
action of boutons located on distal dendrites would, in
Input 1"('~i~taJl('" (\\!l) general, be expected to have slower time courses than
FIG.36. Dependence of amplitude of single-fiber EPSPs on size those elicited by somatic boutons (Fig. 37A). An ad-
of motoneuron in cat. Top, relationship between EPSP amplitude ditional variable, which may affect the shape of an
(obtained by spike-triggered averaging) and motor axon conduction EPSP, is the time course of the synaptic current.
velocity determined for 4 separate group la afferent fibers. Axonal
conduction velocity determined by antidromic stimulation. Medial
Indirect evidence suggests that this may vary from
gastrocnemius (MG) la-fibers and motoneurons in each case. Note synapse to synapse (176) perhaps as a result of con-
inverse relationship. Bottom, each point represents median value of duction-time differences in the presynaptic terminal
amplitude distribution of a presumed single-fiber EPSP evoked in systems (141, 177,275). A relatively slow EPSP might
a motoneuron during stretch of triceps surae. The EPSPs are not result from a fast-current transient acting on distal
restricted to those generated by la-fibers. Many of these EPSPs are
probably produced by intemeurons. It is assumed that variation of
dendrites or from a slow-current transient acting on
these EPSPs with motoneuron input resistance as shown would also the proximal dendrites or soma. A very fast EPSP is
occur in the subpopulation of la-evoked EPSPs. Small cells with likely to result from a brief-current transient acting
large input resistance (203) generate larger EPSPs. 0, Gastroc- proximally. Factors that may also influence the time
nemius motoneurons supplying type F (rapidly contracting) motor
units; &, MG motoneurons supplying type S (slowly contracting)
course of an EPSP include the time constant of the
motor units; and e, soleus motoneurons. [Top from Mendell and postsynaptic cell (7; 3 < 7 < 12 ms) and the ratio of
Henneman (255); bottom from Burke (36).] dendritic to somatic conductance (p; 3 < P < 25).
The influence of the distance between synapse and
ity (see sections
FIRING PATTERNS OF INDIVIDUAL recording point on the time course of synaptic poten-
MOTONEURONS AND MOTOR UNITS, p. 435, and NON- tials was first demonstrated experimentally by Fatt
UNIFORMITY OF MOTONEURONS, p. 485). However, and Katz (112) at the neuromuscular junction (Fig.
these studies may provide an incomplete picture. 37B). As the recording microelectrode was moved
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 469

A B
10mVJ
S
0

i~
1r--
0·5; .

b c d 1·5/ -
2/

2-5

3·5

4
o I
10
!
20
I
30
I

FIG. 37. A: EPSP shapes computed for input restricted to specific regions of the dendritic tree of
cat. All EPSPs assumed to be recorded from compartment 1 (the soma) of a lO-compartment
motoneuron. Compartments 2-10 represent dendritic regions of equal electrotonic length but increas-
ing distance from the soma. Top. time course of the conductance transient generating EPSPs is given
by dotted line; a is EPSP expected for a transient delivered to all compartments simultaneously.
Bottom, b-d are EPSPs expected for transients delivered to compartments 1, 4, and 8, respectively.
Ordinate is in arbitrary units; to produce EPSPs with equal peak amplitudes as shown, intensity of
conductance changes for d > c > b > a. Note differences in time course of EPSPs, and for b-d, note
differences in time of their onset. B: end plate potentials at different locations in same muscle fiber
(frog sartorius in vitro). Microelectrode was inserted sequentially into curarized fiber and end plate
potential in response to motor nerve stimulation was measured. Numbers at left of each trace
represent distance in millimeters from end plate focus. S is the stimulus artifact. Time is in ms. [A,
from Rail (287); B, from Fatt and Katz (112).]

progressively farther from the synapse, the end plate components (Fig. 38C), but this is exceptional. These
potentials (EPPs) became slower and smaller (see next observations indicate either a compound time course
subsection). This is in qualitative agreement with the for synaptic current or more than one location for the
predictions of the Rall model (Fig. 37). Individual synaptic boutons from a single afferent fiber, or both.
EPSPs in motoneurons show similar variations (176, Rall et al. (289) concluded that some synapses were
182, 255, 263, 269, 289, 315) suggesting the basic cor- probably distributed to more than one compartment.
rectness of the Rall analysis. Moreover, half-width A similar conclusion was reached by Mendell and
(duration at half amplitude) and rise time (time to Weiner (256), who compared EPSPs produced by two
peak) are positively correlated (176, 182,255,263,269, la-fibers in the same motoneuron (see anatomical
289,315) as predicted by the theory (287). Impulses in evidence in support of this in subsection Anatomical
a single la-fiber produce EPSPs of very different time Studies, p. 461).
courses in different motoneurons (Fig. 38A). Con- Methods of calculating the distance from synapses
versely, impulses in two la-fibers can elicit slow and to soma from EPSP time course have been described
fast EPSPs in the same motoneuron (Fig. 38B). The by Jack et al. (182). Use of these techniques (176, 182)
time course of an individual EPSP is not an intrinsic has revealed that the majority of la-EPSPs are gen-
property of either the la-fiber or the motoneuron erated by terminals located on the proximal half of the
alone; other factors that vary from synapse to synapse dendritic tree. Synapses on the soma and distal den-
also influence the time course. drites occur relatively infrequently. This distribution
Some individual EPSPs have half-widths that are agrees with the anatomical findings of Conradi and
too long for the rise time (based on Rall's calculations McLaughlin (71, 72, 252) as described in subsection
for input localized to a single compartment and a Anatomical Studies, p. 461. Studies have been limited
simple synaptic current time course). Occasionally, to EPSPs with simple time courses. The electrical
such EPSPs appear to consist of two well-defined stimulation methods used by these investigators prob-
470 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

B #
A .:

------
.~.

-....
r ...,,_
.

.1. ._. . . .

.--.......,
----.--<-

FIG. 38. A: single-fiber excitatory postsynaptic potentials (EPSPs) obtained using spike-triggered
averaging produced by the same medial gastrocnemius (MG) la-afferent fiber in 6 different MG
motoneurons in cat. Note different shapes of the 6 EPSPs. Calibration pulse at the end of each sweep
is 200 /l V. B: single-fiber EPSPs (by spike-triggered averaging) produced in the same LG motoneuron
by an LG la-fiber (top) and a soleus la-fiber (bottom). Calibration pulse is 100 /lV, 1 ms. Note
differences in shape of these EPSPs. C: example of a single fiber EPSP (MG la-fiber and MG
motoneuron) with compound shape (spike-triggered averaging). Calibration is 1 ms, 50 /lV. [A, from
Mendell and Henneman (255); B, from Scott and Mendell (315); C, from Mendell and Henneman
(255).]

ably activated the largest la-afferent fibers. These Data on individual EPSPs suggest less attenuation
indirect estimates of the location of synaptic terminals of distally generated EPSPs than predicted by Rall's
may, therefore, be biased. theoretical studies. A negative correlation between
The rise times of individual EPSPs are longer in EPSP amplitude and rise time would be expected, but
small than in large motoneurons (315). A similar con- this is actually weak as shown in Figure 39A (see also
clusion was reached by Burke (36) from analysis of refs. 22, 34, 217, 255, 256). Rise time may not be an
composite EPSPs. These data alone cannot be used to accurate indication of synaptic location because of the
infer that synaptic terminals are located more distally variations in synaptic-current time course (0:) as well
in small motoneurons because the longer time con- as motoneuron time constant (7'). However, the find-
stants of small motoneurons (48) would tend to pro- ings of lansek and Redman (176) indicate that den-
long EPSPs irrespective of the location of the termi- dritic EPSPs recorded at the soma are about the same
nals. However, Jack et al. (181) have taken the time- amplitude as those generated at the soma where syn-
constant factor into account and have shown a tend- aptic distance is estimated taking p and 7' into account
ency for la-terminals on small motoneurons to be (Fig. 39B). The correlation between amplitude and
located more distally than those on large cells. In half-width is unexpectedly positive (see ref. 256 for
experiments on individual motoneurons, they found discussion) .
no evidence that la-fibers with different peripheral The small differences between amplitudes of EPSPs
conduction velocities terminate on different parts of produced by dendritic and somatic synapses do not
the somadendritic tree. Conduction velocity of la-af- necessarily argue against the Rall motoneuron model.
ferent fibers central to the dorsal root ganglion is The prediction that dendritic synapses should evoke
considerably less than in the periphery (238). smaller EPSPs than somatic ones requires assump-
tions concerning the uniformity of synaptic input,
EFFICACY OF SYNAPSES ON DIFFERENT PARTS OF MO- postsynaptic receptors, and motoneuron properties at
TONEURON. Rall's analysis has been instructive in those parts of the motoneuron. Differences in these
evaluating the effects exerted by boutons located on properties have been postulated to explain the simi-
different regions of the motoneuron surface. Rinzel larity in amplitudes of distally and proximally gener-
and Rall (300) have calculated that an input transient ated EPSPs. These include increased receptor sensi-
(current source) delivered to a single dendritic branch tivity at dendritic synapses (176, 256), enhanced resis-
generates a much smaller EPSP at the soma (by tivity of dendritic membrane compared to the soma
factors of up to 235) than the same input at the soma (111, 175, 196), a greater number or size of boutons at
itself. Several factors enter into the analysis, including dendritic synapses (294), and a regional membrane
nonlinear summation in the dendrites, the attenuation specialization with reversal potentials that are more
introduced by the dendritic cable, the number of positive for dendritic synapses, with a resultant de-
neighboring branches activated (288), and the prop- crease in nonlinear summation (294). At present, none
erties of the current, e.g., steady vs. transient (15, 290). of these explanations has decisive experimental sup-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 471

.·...
N
I>
......
I>
.. :..
... • •
.••. .
.............................................................................................................- - .

··.....
• !.

1.0 3.0

(e; 2,5, 193 iN)

•
• •• • e
•• •
•
• • •• •
• •
••
., • •
••
• •• • •• •
•

o 0·2 0·.. 0·6 0·8 1·0 1·2 1·.. 1·6 1·8 2·0
Distance from soma to synapse (x/A.)
FIG. 39. A: plot of ratio of excitatory postsynaptic potential (EPSP) amplitudes fVdV 2 ) against
ratio of their rise times (RT./RT.). Each point was obtained by consideration of 2 averaged single-
fiber EPSPs (spike-triggered averaging) evoked by different la-fibers in the same motoneuron. Medial
gastrocnemius (MG) la-fibers and MG motoneurons throughout. By defmition RT 1 < RT 2 so that
RT2 /RT 1 > 1. Furthermore VdV 2, the ratio of the amplitudes, is defined so that amplitude offaster
rising EPSP is in numerator. Majority of points have V,/V2 > 1 suggesting a tendency for faster rising
EPSPs to be larger than slower rising ones. Correlation between V/ ,/V 2 and RT2IRT 1 is very weak.
B: plot of single-fiber EPSP amplitude (obtained by electrical stimulation) against distance (in units
of space constant A) between the terminals and the electrode (assumed to be in soma). The EPSPs
are chiefly from hamstring and triceps surae motoneurons with both homonymous and heteronymous
combinations. [A, from Mendell and Weiner (256); B, from lansek and Redman (176).]

port; in fact, anatomical studies suggest that la-bou- among the unit EPSPs produced by impulses in a
tons on dendrites are not greater in size (222) or single la-fiber because the observed amplitude histo-
density (72) than those on more proximal regions. gram was narrower than expected from the quanta!
Nonlinear summation of synaptic potentials would analysis (i.e., fewer large potentials). However, Burke
be expected if the EPSPs approached the equilibrium (34) demonstrated that aggregate EPSPs evoked by
potential and/or if appreciable shunting of one by the volleys in different muscle nerves sum linearly in the
other took place. Kuno and Miyahara (219) concluded majority of cases. The maximum deviation from per-
that substantial nonlinear summation could occur fect linearity was 17% (i.e., 83% of the expected value
472 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

calculated from adding the separately evoked EPSPs). nomial statistics. Quantal content was generally less
Since the terminals of a single fiber are more likely to than 5 with occasional synapses exhibiting values up
be clustered than those from two muscle nerves, this to 20-25 (215, 219, 220). Similar findings were made
difference in summation is not surprising. Further- by Burke and Nelson (42) and Mendell and Weiner
more, it is likely (219) that this effect would be greatest (256). These results suggested that transmitter release
for dendritic terminals where the EPSPs would be mechanisms at the la-motoneuron synapse were qual-
expected to be largest due to locally high input resist- itatively similar to those at the neuromuscular junc-
ance. The occurrence of very large synaptic potentials tion (88). The number of units released by a single
at dendritic terminals has been confirmed by lansek impulse at the neuromuscular junction, however, is
and Redman (176). about 2-3 orders of magnitude greater than at the la-
The inference of nonlinear summation was also motoneuron synapse.
made (269, 315) to explain the finding that the ex- Edwards et al. (105) analyzed similar data but ap-
pected aggregate EPSP amplitude, calculated from plied different calculation procedures. Charge transfer
individual EPSP amplitudes, was usually larger than (time integral of the voltage) rather than peak ampli-
the experimentally determined value (99, 103). The tude of the EPSP was used for analysis. Furthermore,
la-terminals from a single muscle nerve may be in they took account of variability introduced by the
much closer proximity on the motoneuron than those synaptic noise. The distributions of EPSP plus noise
from different muscle nerves. This could account for and noise alone were calculated. They solved for the
the smaller amount of nonlinear summation in the signal distribution, which, when convolved with the
results of Burke (34). It would be important to learn observed noise distribution, gives the EPSP plus noise
whether certain combinations of inputs or types of distribution. In over one-third of the individual EPSPs
motoneurons exhibit more nonlinear summation than only a single peak was obtained for the EPSP charge
others. distribution, and in about another one-third, two peaks
Many authors have tended to equate the amplitude were found whose amplitudes did not have a simple
of the synaptic potential recorded at the soma with its integral relationship as in a quantized process. Peaks
effectiveness in causing action potentials. However, corresponding to failures, which were quite common
since the ultimate effect of EPSPs is to cause depolar- in the earlier analyses, were seen only rarely. These
ization of the initial segment, and since summation of results suggested neither a Poisson nor a binomial
EPSPs is necessary to reach threshold, the duration process.
of the EPSP is an important variable. Barrett and These two approaches lead to very different conclu-
Crill (15) have suggested that the time integral of an sions: individual EPSPs in single sweeps are very small
EPSP is perhaps a better measure of an EPSP's effec- and are accompanied by considerable synaptic noise
tiveness. RaIl (285) has made the interesting sugges- from other inputs, making direct analysis difficult.
tion that dendritically generated EPSPs, which are Fitting the observed distribution with a Poisson proc-
usually slower (measured at the soma) than somati- ess (256) or calculating a peak charge distribution
cally generated ones, might act to regulate excitability. (105) requires statistical verification (e.g., l analysis)
More proximally placed inputs would be apt to cause and this has been obtained in both cases. The major
specific impulse patterns. Luscher, Henneman, et al. difficulty is that there is no unique solution to this
(244) have provided experimental support for these problem and statistical methods cannot determine
suggestions. Theoretical work indicating the impor- that one solution is significantly better than another.
tance of the first derivative of EPSP time course in Edwards et al. (106) used their convolution tech-
raising the probability of firing (209, 211) is also con- nique to analyze two individual EPSPs whose com-
sistent with these ideas. pound shapes made it very unlikely that the respon-
sible terminals were located on a single compartment
Physiology of la- Terminals of the motoneuron. Calculation of the charge histo-
grams revealed that each EPSP consisted of two
It is not feasible at present to make satisfactory peaks. The EPSP associated with each peak was found
intracellular or extracellular recordings from identified to have a different shape and the probability of occur-
regions of la-terminals. The recording of individual rence of these peaks was different. These results were
EPSPs in a motoneuron, however, may provide some interpreted as indicating two sets of terminals from a
indirect evidence regarding la-terminals, permitting la-fiber located on different compartments of the mo-
inferences to be made concerning their function. toneuron with different probabilities of activation.
Kuno (215) suggested that fluctuations in EPSP This suggested intermittent conduction in the presyn-
amplitude at a particular synapse were the result of a aptic terminals. The alternate possibility, namely that
variable release of transmitter. He plotted the distri- these intermittent events reflected a stochastic release
bution of amplitudes resulting from stimulation of process, was rejected because the charge histograms
single la-fibers or small numbers of them. These am- did not satisfy binomial or Poisson statistics.
plitude histograms were found to have multiple peaks An important common thread in these two groups
that satisfied Poisson statistics or, in some cases, bi- of results is that stochastic or probabilistic processes
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 473

probably occur at some locus along the path of the TABLE 5. Percentage of Triceps Surae Motoneurons
presynaptic la-fiber. At present, it is not clear if this Receiving Homonymous or Heteronymous
is in the terminals and consists, for example, of a Projections From la-Fibers of Triceps Surae
quantized release process obeying Poisson or binomial Muscles as Indicated by Spike- Triggered Averaging
statistics, or if it is located at points of fiber branching, Motoneuron
which can be regions oflow safety factor (141, 275). At Afferent
issue is the nature of the unit or smallest EPSP MG LG SOL
produced by a single fiber. (See refs. 44, 51 for further MG 78 54 63
discussion.) In a more general context, attempts to (89) (69) (24)
deal fully with the relationship between the anatomy LG 49 71 68
(49) (84) (19)
of the la-fiber distribution to motoneurons and the
SOL 42 74 100
resultant functional action must await resolution of (46) (61) (18)
these problems. MG, medial gastrocnemius; LG, lateral gastrocnemius; SOL, so-
leus. Values represent percentage of motoneurons that develop
Distribution of Ia Excitation excitatory postsynaptic potentials. Numbers in parentheses repre-
to Motoneuron Pools sent total number of observations. [From Scott and Mendell
(315).]
PERCENTAGE OF MOTONEURONS RECEIVING TERMI-
NALS FROM SINGLE la-FIBERS. In studying EPSPs pro-
duced by impulses in single la-fibers, the opportu- was attributed to the finding of these more narrowly
nity arises to determine the percentage of motoneu- projecting afferents. Similar la-fibers were found to
rons in an identified pool to which a given la-fiber arise from the LG muscle. Soleus la-fibers projected
projects. Both la-fibers and motoneurons are identi- to all (100%) of the SOL motoneurons, but the sample
fied by their responses to electrical stimulation of size for this study was relatively small, and this result
muscle nerves. The spike-triggered averaging tech- may be less reliable than for MG and LG afferents.
nique is particularly suited to this type of study be- Fibers projecting to less than 65% of homonymous
cause impulses can be recorded from a single la-affer- motoneurons have been called type Y. They have
ent in an intact dorsal rootlet for many hours, and the conduction velocities similar to fibers projecting to
responses of many motoneurons can be studied. Men- more than 80% of motoneurons, called type X fibers.
dell and Henneman (254,255) demonstrated that ac- These are arbitrary classifications, and fibers with
tivation of single MG la-fibers evoked EPSPs in vir- intermediate projection percentages exist. Type Y fi-
tually all (94%) MG motoneurons. They inferred that bers tend to produce smaller EPSPs than type X
each of the approximately 60 la-fibers (56) from MG fibers. It is possible that the smaller projections of
projected to each of the 300 a-motoneurons (102) type Y fibers result from inability to detect the smaller
innervating this muscle. Kuno and Miyahara (220) EPSPs they produce. More recent studies in prepara-
performed a similar experiment but found a much tions with spinal cord transection indicate that con-
lower percentage (44%). They ascribed this difference nectivity has been underestimated in intact prepara-
to their inability to resolve very small EPSPs, since, tions and that the error is particularly severe for type
rather than averaging, they examined single sweeps. Y fibers (see next subsection). Similar connectivity
This illustrates an obvious but important technical has been reported by Watt et al. (345) for triceps surae
problem in such studies, namely that failure to detect afferents and by Nelson and Mendell (269) for semi-
small EPSPs can lead to an underestimate of the total tendinosus afferent fibers. In the latter case a mixture
la projection. The presence of a tonic inhibitory proc- of type Y and type X afferent fibers was seen in about
ess (pre- or postsynaptic) could cause the same prob- the same proportion as observed for triceps afferent
lem. Recent evidence indicates the existence of such fibers.
processes in preparations with intact spinal cords (see
next subsection) and suggests that they play an im- FACTORS INFLUENCING PERCENTAGE OF MOTONEU-
portant role in determining the percentage of func- RONS RECEIVING HOMONYMOUS PROJECTIONS. Esti-
tioning synapses. mates of the projections of single MG la-fibers to MG
Scott and Mendell (315) investigated the projection motoneurons in preparations anesthetized with pen-
of all three types of triceps surae efferents (MG, LG, tobarbital and with intact spinal cords have ranged
SOL) to their homonymous motoneurons in the intact from 78% to 94% (255, 315). This variation reflects at
preparation (Table 5). Although the majority of la least in part the proportion of type Y afferent fibers in
impulses from MG elicited EPSPs in virtually all the sample (269, 315) and the sizes (conduction veloc-
homonymous motoneurons, a significant minority ity) of the afferent fibers (Table 6). In anesthetized
evoked EPSPs in only 65% or less of the homonymous preparations with spinal cord transections at T 13 or Ls,
pool. The average percentage for all MG la-afferent single la impulses elicit EPSPs in 98% of the homon-
fibers in these experiments was 78%. This decrease ymous motoneuron pool and no type Y fibers are ever
from 94% reported by Mendell and Henneman (255) encountered (268). Since these changes occur soon
474 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

after transection (subject to time required to record weaker than that to homonymous motoneurons. The
from enough motoneurons and also to sampling error), studies carried out by Lloyd (235) using ventral root
growth of new terminals is an unlikely explanation. It discharges, and subsequently by Eccles and collabo-
seems more probable that these "new" connections rators (99, 103) using intracellular recording, con-
already existed in the intact cord but were inactive. firmed the existence of a weaker monosynaptic excit-
This increase in functioning connections can occur in atory input from la-fibers to heteronymous than ho-
preparations exhibiting normal EPSP amplitudes (268, monymous motoneurons.
270), so it seems unlikely that an improvement in In the context of the previous analysis of la-evoked
resolution of small EPSPs is the major factor respon- EPSPs, the smaller aggregate EPSPs in heteronymous
sible for the higher percentage of projections observed. motoneurons might have resulted from fewer projec-
Since the changes after transection occur in a mono- tions to heteronymous motoneurons, smaller EPSPs
synaptic pathway, their locus must be restricted to the in the motoneurons, or both. There is agreement
presynaptic terminal arborizations or to the synapse among several groups that single la-fibers project
itself. It is known that descending impulses can cause more widely to homonymous than to heteronymous
presynaptic depolarization of la-terminals (53, 54), motoneurons (220, 255, 269, 315; 345). This remains
which might diminish or eliminate conduction through true even after spinal cord transection (268). In early
terminal branches (167) and reduce release of trans- studies (255, 345) individual EPSPs in homonymous
mitter (198, 234). Interruption ofthis descending input motoneurons (MG) were found to be only slightly
following spinal cord transection might increase the larger on the average than in heteronymous motoneu-
number of motoneurons to which presynaptic im- rons (LG plus SOL). It was concluded that the larger
pulses have access by a process of presynaptic disin- aggregate EPSPs in homonymous motoneurons were
hibition. Under these conditions the distinction be- due to the projection of more la-fibers to homonymous
tween type X and type Y afferent fibers would disap- than to heteronymous motoneurons.
pear, since each afferent could evoke EPSPs in each This issue was subsequently reexamined by Scott
homonymous motoneuron. It follows that these differ- and Mendell (315), who made a special effort to distin-
ences observed in preparations with intact spinal cords guish the projections to LG and SOL motoneurons
may reflect functional (e.g., effects of presynaptic in- (whose axons run in the same peripheral nerve), since
hibition) rather than anatomical factors (i.e., number EPSPs would be expected to be larger in SOL cells
of motoneurons to which terminals are sent). due to their smaller size (subsection FACTORS RESPON-
In conclusion, it appears that projections to moto- SIBLE FOR VARIABILITY IN EPSP AMPLITUDES, p. 466).
neurons can be underestimated in experiments on the In these studies motoneurons were first classified by
intact spinal cord because of inhibitory effects exerted antidromic stimulation as either MG or LG-SOL. If
on terminals by systems desending from more rostral they were LG or SOL, the motoneuron was stimulated
regions. The findings as a whole suggest that each la- intracellularly to localize the muscle in which the
fiber sends terminals to all of its homonymous moto- contraction occurred. The projections from LG and
neurons. SOL la-afferent fibers to these three types of moto-
neurons were also examined. The results (Table 7)
Comparison of Projections to Homonymous were subjected to a two-way analysis of variance,
and Heteronymous Motoneurons which showed conclusively that, on the average, indi-
As early as Sherrington's observations (233) it was vidual EPSPs from any of the three afferents were
clear that input to heteronymous motoneurons was larger in homonymous than in heteronymous moto-

TABLE 6. Relationship of Axonal Conduction TABLE 7. Mean Amplitude of Excitatory

Velocity of la-Afferent Fibers and Postsynaptic Potentials Elicited by
a-Motoneurons to Percent of Homonymous Homonymous and Heteronymous Inputs
Projections Linking Them
Motoneuron
Conduction Velocity Projection Frequency Projection Frequency Afferent
Range, mls From Afferent Fibers to Motoneurons MG LG SOL
70-80 50% (16) 89% (45) MG 95 59 79
80-90 81% (80) 78% (131) (69) (37) (15)
90-100 77% (103) 76% (114) LG 76 14 84
100-110 87% (87) 83% (95) (24) (60) (13)
110-120 96% (25) 82% (17) SOL 50 65 143
(19) (45) (18)
Data derived from studies on homonymous projections, including
medial and lateral gastrocnemius, soleus and semitendinosus. Num- Excitatory postsynaptic potentials (EPSPs) in triceps surae mo-
bers in parentheses represent total number of observations. la- toneurons produced by impulses in triceps surae group la-afferent
afferent conduction velocity was often not available. Values in left fibers. EPSPs obtained by spike-triggered averaging. Values are in
column refer to conduction velocities of la-fibers (middle column) microvolts. Values in parentheses are number of observations mak-
and to conduction velocities of motoneurons (right column). ing up each mean value. Note that largest values are on diagonal
[From data of Mendell and Henneman (255); Nelson, Mendell, et representing homonymous combinations. [From Scott and
al. (268); and Scott and Mendell (315).] Mendell (315).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 475

neurons, excluding other factors such as the tendency only be speculative at this time. It is important to
for EPSPs to be larger in SOL motoneurons because emphasize that the higher percentage of homonymous
of their small size. The largest EPSPs (greater than projections might be due to the larger individual
150 p. V) occurred almost exclusively in homonymous EPSPs they elicit. The EPSPs too small to be resolved
motoneurons; small EPSPs occurred in both types. In would undoubtedly be more common in heteronymous
several cases, the EPSPs evoked in the same moto- motoneurons and would therefore reduce the projec-
neuron by impulses in one homonymous afferent and tion frequency to motoneurons in this pool more than
one heteronymous afferent were compared (Fig. 38B). to the homonymous pool. However, if resolution were
Under these more controlled experimental conditions improved, it seems likely that additional small EPSPs
where motoneuron properties were identical for each would be found more frequently in heteronymous
comparison, the homonymous EPSP was found to be motoneurons. Since the mean EPSP amplitude is cal-
significantly larger on the average than the heterony- culated taking into account only nonzero values, the
mous one. differences in mean EPSP amplitude between homon-
A possible explanation for larger EPSPs in homon- ymous and heteronymous motoneurons might be even
ymous motoneurons might be a tendency for la-fibers larger than those found with the present level of
to terminate in a preferred location (173, 236), e.g., resolution.
closer to the soma of these cells (but see subsection Since large EPSPs cannot be accounted for by a
EFFICACY OF SYNAPSES ON DIFFERENT PARTS OF MO- more proximal location of the active boutons on the
TONEURON, p. 470). Jack and Porter (184) did, in fact, soma-dendritic membrane, other factors must be re-
find a tendency for aggregate EPSPs in homonymous sponsible. Kuno (215) proposed that EPSP amplitude
motoneurons to have slightly briefer rise times than was related to the mean quantal content of the EPSP,
in heteronymous ones (but see ref. 220). Scott and which was verified for individual EPSPs by Mendell
Mendell (315) confirmed this finding (see also ref. 263) and Weiner (256). A more general conclusion, in view
although the differences in mean rise time, which were of the discussion in subsection Physiology of la-Ter-
very small, could hardly account for the differences in minals, p. 472, is that larger EPSPs are produced by
amplitude, given the very weak relationship between the action of a greater number of functional presyn-
amplitude and rise time (see subsection EFFICACY OF aptic "units."
SYNAPSES ON DIFFERENT PARTS OF MOTONEURON, p. A specific proposal (269) that could account for
470). Thus, it seems likely that factors other than these observations is that afferent fibers give off more
location of terminals on the somadendritic tree ac- boutons on the average to homonymous than to het-
count for larger EPSPs in homonymous motoneurons. eronymous motoneurons (Fig. 40). This would result
Similar studies have been carried out on the projec- in a larger mean EPSP in homonymous motoneurons.
tions of la-fibers from the semitendinosus muscle to If, in addition, the number of boutons exhibited vari-
homonymous and heteronymous (posterior biceps, ability about these means, one would expect 1) a
semimembranosus) motoneurons (269). The results smaller number of la-projections to heteronymous
confirmed the findings obtained from analysis of tri- motoneurons, since a greater portion of this distribu-
ceps surae connections, namely, that homonymous la- tion of bouton numbers would be truncated at 0 (no
fibers project to a higher percentage of motoneurons boutons; therefore no projection); and 2) restriction of
and elicit larger EPSPs than heteronymous fibers. the largest EPSPs to homonymous motoneurons
Attempts to account for differences in projections (315), since only homonymous projections would in-
to homonymous and heteronymous motoneurons can clude the largest number of boutons. It is possible that

Mean Boulon
10 AFFERENT Number ProjFreq "leon Amp
5.0 5/5 5units
1.8 3/5 3 unifs

5
o.o I 6
FIG. 40. Model of projection of single la-fiber to homonymous and heteronymous motoneurons
(circles). Systematic differences in mean number of boutons provided by single la-fibers to homony-
mous and heteronymous motoneurons would lead to differences in projection frequency and excitatory
postsynaptic potential (EPSP) amplitude in these 2 types of projections. Mean bouton number was
chosen using estimates of lies (177) as a guide. Numbers below each circle represent number of
boutons provided by single la-fiber to each motoneuron.
476 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

motoneurons have a fixed number of receptor sites for single la-fiber to make contact with virtually all mo-
each type of afferent fiber, with more reserved for toneurons in the homonymous pool despite the extent
homonymous than heteronymous la-fibers. This hy- of the pool (up to 8 mm) along the rostrocaudal axis
pothesis implies a variable number of boutons per la- (47, 302). Scheibel and Scheibel (310) suggested that
fiber (51, 177). In its simplest form, it requires that the a single la-fiber might project to 90-900 a-motoneu-
boutons be identical, so that their number is the rons. Physiological studies indicate that a la-fiber from
decisive variable. The evidence for this is by no means MG makes contact with virtually all of the 300 MG
clear. Some investigators (222, 252) have reported that motoneurons and at least 60% ofthe 600 heteronymous
all la-boutons are similar, whereas Conradi (72) has cells belonging to the LG and SOL pools.
described two types, M and S (subsection Anatomical An important difference between the results of Golgi
Studies, p. 461). It is conceivable that the boutons on studies on kittens and more recent studies with horse-
homonymous motoneurons may be qualitatively dif- radish peroxidase (26, 51, 163) is in the number and
ferent from those on heteronymous cells, e.g., the spacing of the ventrally directed la-collaterals. Schei-
larger type M boutons may be restricted to homony- bel and Scheibel (310) reported that 10-100 collaterals
mous motoneurons. Another variable of possible sig- were given off at intervals of 100-200 J1.m. Studies with
nificance is bouton size. At the neuromuscular junction HRP indicate that only 5-11 collaterals are given off
the number of quanta released is directly correlated at intervals of about 1 mm. Depending upon which set
with the size of the nerve terminals (223). Current of figures is used, each collateral could contact as
evidence suggests that there is a fivefold variation in many as 50-60 motoneurons (26) or as few as about 10
the length of bouton contacts on motoneurons (222), (177, 310).
which may account for differences in synaptic efficacy. These differences between the number and spacing
Variations in the areas of synaptic contact may explain of collaterals are not easily resolved at this time. It
the differences in the amplitude of EPSPs, which does not seem likely that some branches of la-fibers
occur upon stimulation of small and large la-fibers have failed to be filled with HRP, since recent physi-
(subsection VARIABILITY OF EPSP AMPLITUDES, p. 465). ological mapping of these collaterals using spike-trig-
The optimal technique for resolving these issues may gered averaging of the presynaptic spike (see next
involve filling single afferents and motoneurons with subsection) indicated that the number of collaterals
HRP, followed by ultrastructural analysis of the syn- has not been underestimated (265). A more intriguing
apse-a difficult task. possibility is that since the Golgi studies were per-
formed in kittens whereas adult cats were used for the
HRP and electrophysiological work, the collateral pat-
Correlations Between
Morphology and Function tern may undergo changes during maturation. How
such reorganization occurs is not known, and no de-
In a Golgi study on kittens Scheibel and Scheibel scriptions of degenerating branches are available.
(310) provided anatomical evidence that may explain However, a substantial loss of boutons from certain
some of the physiological findings that have been regions of the motoneuron occurs in the period im-
described. It is now generally agreed that the soma mediately after birth (74, 303), and this remodeling
and dendrites of motoneurons are disposed in a pre- process may be related in some manner to reorgani-
dominantly longitudinal column running for a consid- zation of the la-projection.
erable distance in the rostrocaudal axis of the spinal Many individual EPSPs have time courses indicat-
cord (309, 332). According to Scheibel and Scheibel, ing that the active la-terminals are located in close
each primary la-fiber in the dorsal column of the proximity to each other (e.g., on a single compartment
kitten gives off a series of primary afferent collaterals of the motoneuron). In commenting upon this fact,
at intervals of 100-200 J1.m (Fig. 32). These collaterals Mendell and Henneman (255) used the then available
leave the dorsal column "at right angles and drop evidence on spacing between collaterals, i.e., 100-200
ventrally with almost plumb-line precision" (310). Sev- J1.m (310), to conclude that although many collaterals
eral collaterals run together, forming a "microbundle", from the same la-fiber might have access to any mo-
which passes perpendicularly through the dendritic toneuron whose dendritic tree has a rostrocaudal ex-
field of the motoneurons. Each primary afferent col- tent of 1-2 mm (332), the time courses of the EPSPs
lateral gives off terminals that ramify almost entirely indicated that only one of these collaterals might
in the mediolateral plane with little rostrocaudal ex- actually synapse on the motoneuron. This suggested
tension. Thus, the terminal field of each collateral is an interesting possibility, namely, that after one set of
virtually two-dimensional and makes contact with a endings from a particular la-fiber is established on a
very limited extent of the longitudinally running mo- motoneuron, no others from different collaterals of the
toneurons. The thickness of these two-dimensional same fiber can develop, even on other dendrites of
fields varies from 10 to 50 J1.m, with a mean of about 25 that motoneuron. With the more recent information
J1.m. indicating a spacing of 1 mm between collaterals, it
The primary afferent collaterals projecting into the now seems unlikely that more than a single collateral
ventral horn at different rostrocaudallevels allow each of a la-fiber has access to a given motoneuron. How-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 477

ever, it remains to be seen whether individual EPSPs delay is probably the time required for electrotonic
in the kitten have a more complex time course than in conduction in motoneuron dendrites. This was ana-
the adult, suggesting early functional la-motoneuron lyzed by Rall (287) and Iles (178), who showed that
connections that are later eliminated, as has been EPSPs generated in the distal dendrites of a model
observed at the neuromuscular junction (28, 293). motoneuron should exhibit longer delays than those
produced in the soma or proximal dendrites. This
prediction was confirmed in experiments showing (Fig.
Latency of EPSPs
41) that the latencies of EPSPs increased with their
Early studies with aggregate EPSPs elicited by vol- rise times (255,263). Differences in the synaptic delays
leys in la-fibers fixed their synaptic delay at about 0.5 of homonymous and heteronymous EPSPs were also
ms (25). When individual EPSPs were investigated, observed that were consistent with the differences in
however, it became apparent that synaptic delay var- their rise times [Fig. 41; (263)].
ied considerably from synapse to synapse. Variability In addition to these differences from synapse to
was first seen when the interval between the arrival of synapse, variability in delay has also been observed in
the dorsal root impulse and the onset of the EPSP was single sweeps at a given synapse (34, 70, 215, 218, 266,
measured (255). Intracellular recording from the mo- 295). This phenomenon is similar to that described
toneuron with spike-triggered averaging often reveals previously at the neuromuscular junction (16, 197).
a small brief wave that precedes the onset of the The variability in delays at motoneurons has been
EPSP; this has been interpreted as the arrival of the localized to the synaptic region by the finding that the
afferent impulse in the terminal region (189, 345). latency of the presynaptic spike is constant, unlike the
Munson and Sypert (263) demonstrated that the la- fluctuations in latency observed in the subsequent
tency between this presynaptic spike and EPSP onset EPSPs (70). These delays can vary by as much as 300-
also fluctuates, thereby localizing the variability to the 400 fLs but most of them are within a 1oo-fLS range and
synapse itself. the distribution of latencies is unimodal. The EPSPs
A major factor underlying differences in synaptic with the shorter latencies have the same time courses

~ o LG/S

• 0
• •
0

•
• 0
•
•
• ••
r,.
0

••
Q
• 0

••
:. • •
• oj. • • • .0. •
••
•I ••i .-. 0 •

....,,.. ....
• ~ Q

: ~ 8 ••
• • '.
~.. : o
:
i
!
O--~---r--"""r---"""T"--oor---.....,--""T"--~-""'"T--""
o .1 .2 .3 .4 .e .6 .7 .8 .9 1.0
LATENCY (mSec)
FIG. 41. Scatter diagram of relationship between rise time and latency for single-fiber excitatory
postsynaptic potentials (EPSPs) in 78 medial gastrocnemius (MG) and 23 lateral gastrocnemius and
soleus (LG-SOL) motoneurons. MG la-afferent fibers throughout. Dashed vertical line at 0.27 ms
denotes minimum latency recorded in heteronymous LG-SOL motoneurons. Inset shows intracellular
(A), extracellular (C) records from a motoneuron with spike-triggered averaging. Note reversal of
EPSP, but not diphasic prespike following withdrawal of electrode. Calibration pulses 50 /lV, 1 ms. B
is same as A with both scales expanded by 4. Latency was measured from negative (downward going)
peak of prespike to onset of EPSP. [From Munson and Sypert (263).]
478 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

(rise time, half-width) as those with the longer laten- leagues (see review in ref. 44). They report that fluc-
cies, which suggests that they are generated by the tuations in the amplitudes of la-evoked EPSPs are
same terminals and that the latency fluctuations are correlated to a significant degree in different motoneu-
a property of the release process (70). rons in a pool from which intracellular recordings are
These findings are consistent with the idea that the made simultaneously (305). Primary-afferent depolar-
la-motoneuron synapse is chemical rather than elec- ization, which may affect many related la-terminals
trical (266). This issue is currently controversial and synchronously, may account for some of the correla-
further discussion is beyond the scope of this chapter tions in motoneuron discharge. Whether widespread
(see review in ref. 295). depolarization occurs via the action of interneurons,
One consequence of these fluctuations in latency is which branch to terminate on many la-terminals, or
that temporal parameters (e.g., rise time) of averaged because of correlated firing among large pools of in-
individual EPSPs are composite values. Any attempt terneurons having relatively narrow projections, is not
to estimate location of synaptic terminals based on known. Whatever the mechanism, the findings suggest
EPSP time course (subsection PHYSIOLOGICAL ANAL- that modifications in synaptic input may be exerted in
YSIS OF LOCATION OF BOUTONS, p. 468) is affected by such a way as not to disturb the relative sttengths of
this process, since measured values are somewhat inputs to motoneurons in a pool, and thus to preserve
larger than the actual ones. recruitment order.

Other Examples of Divergence la-Projections to Motoneurons Controlling

in Inputs to Motoneurons Other Parts of the Body
Less direct methods of studying the inputs to mo- The previous discussions have centered on reflex
toneurons also indicate considerable divergence of af- connections involving the hindlimb of the cat. In this
ferent terminals to a motoneuron pool, resulting in a subsection functional aspects of la-motoneuron con-
significant degree of correlated firing in the cells of the nections for muscles controlling other regions of the
pool. Working in the respiratory system, Sears and body are briefly discussed.
Stagg (317) have used spike-triggered averaging to Forelimb muscles in the baboon have a substantial
show, by means of pre- and poststimulus time histo- complement of spindles. Afferent fibers from them
grams, that the discharge of motor axons is highly project to both homonymous and heteronymous mo-
correlated with the firing of other motor axons. This toneurons (65). Projections to heteronymous moto-
result was perhaps predictable in view of previous neurons are generally, although not always, weaker
findings that impulses in single la-fibers excite all the than those to homonymous motoneurons as measured
members of a pool at approximately the same time. by their mean aggregate EPSP. The differences, how-
Use of the EMG rather than recordings from motor ever, do not appear to be as great as for motoneurons
nerves makes the Sears and Stagg technique suitable supplying cat hindlimb muscles.
for human studies, which is an important advantage. The spike-triggered averaging technique has been
The preceding study led to the hypothesis that the used to examine the monosynaptic connections made
joint occurrence of EPSPs elicited in motoneurons by by afferents from jaw elevator muscle spindles to
impulses in branches of common stem presynaptic masseter and temporalis motoneurons (7). Intracellu-
fibers causes a transient depolarization in one respi- lar recording revealed monosynaptic projections in
ratory motoneuron at the time of discharge in another only 14 of 91 instances and a mean amplitude of only
motoneuron of the same pool. The hypothesis was 18 /LV. These connections were studied further by
tested by averaging the synaptic noise of inspiratory using the spike-triggered averaging procedure to mea-
motoneurons (209). A depolarization centered around sure the extracellular fields produced by impulses in
the time of the trigger spikes was consistently observed single la-fibers. These fields were found to be highly
and was designated the average common excitation localized within the motor nucleus and were found to
(a.c.e.) potential. The a.c.e. potentials had a mean be somewhat larger in response to input from dynamic
amplitude of 32 /LV when the trigger signals came from spindles than from static ones, again confirming the
a fIlament in the same segment, but only 19 /LV when smaller monosynaptic action from the slower conduct-
the fIlament was two segments rostral to the motoneu- ing fibers of smaller diameter discussed in subsection
ron. The a.c.e. potentials were recorded when either FACTORS RESPONSIBLE FOR VARIABILITY IN EPSP AM-
a- or y-discharges served as triggers. This was regarded PLITUDES, p. 466. The authors explained this very
as evidence for a-y-coactivation by common presyn- limited projection as possibly resulting from the com-
aptic axons. The authors have developed a theory that pound structure of these muscles, which have multiple
accounts for the main features of the a.c.e. potential. insertion points, unlike limb muscles with single inser-
Another indication that impulses in branches of tions. Spindles from a given part of the muscle might
common stem presynaptic fibers are an important therefore project to only certain functionally related
mechanism for correlating the activity of members of motoneurons.
a motoneuron pool was noted by Rudomin and col- Considerable effort has been devoted to the study
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 479

of reflexes in respiratory muscles, where it is known synaptic delay, as measured from the prespike to the
that a powerful monosynaptic reflex exists from large- onset of the EPSP, was identical to that of EPSPs
diameter fibers (316) and a weaker one from group II elicited by la impulses (328), as illustrated in Figure
fibers (207, 208). The details of the distribution of la- 42.
fibers from external and internal intercostal muscles Stauffer et al. (328) found that impulses in group II
in various segments to the motoneurons innervating fibers from MG spindles evoked EPSPs in 52% of MG
these muscles are not known. Since prevention of motoneurons and in 26% of LG-S cells, compared to
spindle unloading by tracheal occlusion during inspi- 87% and 61%, respectively, for impulses in la-fibers
ration causes increased inspiratory motoneuron dis- under similar experimental conditions (345). The
charge (76), it would appear that la excitation is di- mean amplitudes of the group II EPSPs were 38 p.V
rected toward homonymous motoneurons, as in limb and 12 p.V compared with 67 p.V and 54 p.V for la-
muscle reflexes. EPSPs. These findings suggest that group II fibers
Neck muscles are well supplied with spindles that give off fewer terminals to motoneurons than la-fibers.
send fibers to make monosynaptic connections with This is consistent with the conclusion reached in ex-
motoneurons innervating these muscles. Recent ex- amining EPSPs elicited by impulses in la-fibers of
periments indicate that these connections are quite different diameters (Table 1). The smaller EPSPs
selective (4), with splenius group la-fibers sending few evoked by impulses in group II fibers probably repre-
terminals to biventer cervicis and complexus motoneu- sent an extension of the relationship between la-fiber
rons, and spindles of the latter muscles sending la- diameter and individual EPSP size. Luscher, Henne-
terminals to their own motoneurons rather than to man, et al. (244) found that the amplitudes of the
those supplying splenius. This organization is similar postsynaptic population potentials (PSPPs) recorded
to that seen for limb muscles with respect to projec- from an entire ventral root were directly related to the
tions to homonymous and heteronymous motoneu- conduction velocities of the la or group II impulses
rons. Functional studies indicate that splenius and the that evoked the PSPPs. Of course, factors other than
biventer cervicis-complexus group have different ac- afferent fiber diameter probably influence the size of
tions (299). Despite the existence of these well-defined PSPPs. For example, if the number of terminals given
monosynaptic connections as revealed by recording off by an afferent fiber is a function of the diameter of
aggregate EPSPs, monosynaptic reflexes to these mus- that fiber, the relationship between conduction veloc-
cles may be hard to demonstrate under some experi- ity of afferent fiber and amplitude of EPSP should be
mental conditions (1). the same for both group la and group II afferents.
These examples indicate that the widespread pro- However, Luscher, Henneman, et al. found different
jections of la-fibers from hindlimb muscles in the cat slopes relating these variables, which indicates that
are not necessarily present in all la-motoneuron sys- factors other than the number of terminals alone
tems. One must be careful in extrapolating findings influence the size of a PSPP.
from this well-studied system to all others. Luscher, Henneman, et al. (244) have also found
that the rise times of PSPPs evoked by group II
Group II Input From Secondary impulses are slower than those of PSPPs elicited by
Endings in Muscle Spindles la impulses. This indicates that the synaptic terminals
given off by group II fibers may be located, on the
In the past few years it has become apparent that average, more distally on the dendritic tree than ter-
group II fibers from muscle spindles must send some minals from la-fibers. These results suggest a possible
of their terminals directly to homonymous motoneu- distinction between dendritic and somatic activity.
rons (207,208, 244, 328). Early studies using electrical The more dendritically located synapses, generating
stimulation of group II fibers in muscle nerves with slower EPSPs (measured at the soma) might be chiefly
unavoidable activation of their la- and Ib-fibers had responsible for the maintenance and regulation of the
revealed central synaptic effects, similar to those elic- general excitatory state of the motoneuron pool. They
ited by impulses in cutaneous fibers, that were inhib- would be well suited to process input from secondary
itory to extensor motoneurons (104, 229). The different endings, which are most sensitive to the static length
synaptic effects of impulses in group II and group la of the muscle. In contrast, the more proximally located
fibers were always puzzling because of the similarity input would be apt to cause rapid triggering of reflex
in location and adequate stimulus of their receptors. discharge. This would be an appropriate response to
The use of the spike-triggered averaging technique to input from primary receptors, which are most sensitive
study the connections of group II spindle fibers has to rapid changes in muscle length. The sustained
demonstrated the existence of a monosynaptic excit- synaptic activity in the dendrites, produced by input
atory projection to homonymous extensor motoneu- from secondary endings may, therefore, modify the
rons (Fig. 42) and the probable existence of polysyn- effectiveness of the input from primary endings. This
aptic projections to motoneurons from the same fibers. functional distinction in the central processing of sig-
Although the conduction time in the central processes nals from primary and secondary endings may shed
of group II fibers was slower than that of la-fibers, some light on how inputs from different peripheral
480 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

0nIc» II

I
I) I)

8 8
8 8
15 4 4

I8lN I 2
0
02 OA Q8 08 02 0.4 0.6 0.8

0nIc» 1A
A

I
I)
B 8
C a: 8
15 4

5m8ec J 2
0
0.2 0.4 0.6 0.8 02 0.4 0.6 0.8
ConcU::tion Tine (maec) SynspIlc DeBy (maec)
FIG. 42. Comparison of synaptic delay associated with single-fiber excitatory postsynaptic poten-
tials (EPSPs) produced by group Ia and group II fibers in cat. Left, trace A represents potential
recorded inside a medial gastrocnemius (MG) motoneuron with spike-triggered averaging from a
group II afferent fiber. The large upward potential is EPSP, which is preceded by a brief diphasic
potential considered to arise from activity in presynaptic terminals of group II fiber. Trace B is record
obtained by spike-triggered averaging from a position just outside motoneuron. The early potential is
not inverted, indicating that it is produced outside motoneuron and unlike EPSP, which reverses,
showing that it is generated across the membrane. Trace C is averaged record recorded from group
II fiber in dorsal rootlet. Interval from dorsal root spike to presynaptic spike is conduction time (CT)
and remaining time to EPSP onset is synaptic delay (SD). Right, histograms of conduction time and
synaptic delay for group II fibers (top) and group la-fibers (bottom). Note similarities in synaptic
delay for both types of fiber and trend for shorter conduction time in (the larger) group-Ia fibers.
[From Stauffer et al. (328).]

receptors can act together at the level of the motoneu- must be involved in determining susceptibility to in-
rons. hibition, e.g., the amount of excess excitation (section
FIRING PATTERNS OF INDIVIDUAL MOTONEURONS AND
Inhibitory Inputs to Motoneurons MOTOR UNITS, p. 435). The mechanism by which an
orderly distribution of inhibitory input is achieved
Motoneurons also receive inhibitory inputs which
without interfering with recruitment order is not
play an important role in determining their discharge
known. It apparently does not involve widely diverging
properties. Recruitment and critical fIring level studies
projections at each synaptic level (see next subsec-
have shown that inhibitory inputs from many sources
tion).
affect motoneurons according to their size, with large
motoneurons being more easily silenced. This is op- Ia INHIBITORY INTERNEURONS. The only interneu-
posite to the situation with excitatory inputs where ronal input studied systematically at the level of in-
large motoneurons are the least susceptible to dis- dividual IPSPs is the projection from interneurons
charge (section FIRING PATTERNS OF INDIVIDUAL MO- mediating disynaptic Ia inhibition (188, 189). The ap-
TONEURONS AND MOTOR UNITS, p. 435). The organi- propriate interneurons were identifIed from their pat-
zation of inhibitory inputs to motoneurons is, unfor- tern of responses to peripheral muscle nerve stimula-
tunately, more difficult to characterize because the tion and by their projection into the motoneuron pool.
projections from the periphery are not monosynaptic. Their synapses on antagonistic motoneurons were
Burke et al. (46) have obtained evidence that the studied by a modifIed version of the spike-triggered
amplitude of aggregate disynaptic IPSPs in antagonist averaging described in subsection Techniques Used to
motoneurons is directly correlated with the amplitude Study EPSPs Elicited by Impulses in Single Afferent
of their monosynaptic EPSPs, the largest IPSPs being Fibers, p. 464 (Fig. 43). Instead of natural activation
found in small cells. Since the small cells in a pool are ofthe interneuron, it was made to fIre by iontophoresis
the last to be silenced by inhibition, additional factors of glutamate. The interneuronal spikes triggered the
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 481

- ..
A
a
01 .... .. -.,. ¥ 4 1

2mVJ
---.---- '1'--- """"'j'-t
I I
~
-. I
..I....
'" -~
I

20 msec
B

~~J
, iii i "
1 msec
FIG. 43. Inhibitory postsynaptic potential (IPSP) in a cat motoneuron analyzed by spike-triggered
averaging. A: top trace is intracellular record from a posterior biceps or semitendinosus motoneuron.
Bottom trace is extracellular spike activity from an interneuron satisfying the criteria required for a
la-interneuron. Interneuron is excited by iontophoretically applied glutamate and discharges rhyth-
mically. Each spike is associated with an IPSP (overlaid by row of dots). B: averages of 512 sweeps
triggered by interneuronal spike from motoneuron (above) and interneuron (below). Trigger level was
set so that averaging began at beginning of interneuronal spike. [From Jankowska and Roberts (189).]

computer, which averaged the intracellular postsyn- proximal dendrites or soma, since the rise times and
aptic potential in the motoneuron. The IPSPs were half-widths were always quite brief. Individual la-
also examined in single sweeps without averaging. EPSPs exhibit a much wider range of rise times and
The projection of 12 quadriceps-activated interneu- half-widths. which overlaps the values observed for
rons (Q) was examined in 43 posterior biceps-semiten- these IPSPs. Reversal of individual IPSPs by intra-
dinosus motoneurons (PBST) that were within 100 p.m cellular chloride injection was always seen, also sug-
of the interneuronal terminals. A projection from a Q gesting that the boutons from a single interneuron
interneuron to this subset of PBST motoneurons was were grouped on the soma of the motoneuron. Com-
found in 38 of 43 instances (88%). Since single Q plex nonuniform reversals were often seen in the case
interneurons were found to project to about 50% of the of aggregate IPSPs.
length of the PBST motor pool, giving off terminals to In many cases, averaged individual IPSPs were pre-
about 40% of this region and innervating about 90% of ceded by a small spike similar to that occurring prior
the neurons within this restricted region, it was esti- to individual EPSPs (subsection Latency of EPSPs,
mated that single Q interneurons terminate on about p. 477). The latency between presynaptic spike and
20% of the entire PBST motor pool. The PBST inter- IPSP onset ranged from 0.28 to 0.42 ms, which is
neurons were found to innervate a somewhat smaller consistent with a monosynaptic connection.
proportion of Q motoneurons. It was suggested that a The distribution of disynaptic inhibition from MG
divergence of single la-fibers to supply a large propor- la-fibers to motoneurons has been studied with the
tion of the interneurons, as is the case with the mono- spike-triggered averaging technique (345). Although
synaptic la-motoneuron projection (subsection PER- the amplitude of the recorded IPSP is a poor measure
CENTAGE OF MOTONEURONS RECEIVING TERMINALS of this disynaptic projection, because the interposed
FROM SINGLE la-FIBERS, p. 473), could result in a interneuron does not allow the motoneuron to respond
widespread effect from single la-fibers to motoneurons in one-to-one fashion, a useful estimate of the diver-
over this disynaptic inhibitory pathway. gence of single la-fibers to target cells connected via
The mean amplitude of these individual IPSPs was interneurons can be obtained. The IPSPs with laten-
52 p.V. The distribution of amplitudes in repeated cies consistent with a disynaptic pathway (1.2-2.4 ms)
trials indicates a quantal release mechanism (224). were found primarily in tibialis anterior and extensor
Comparison with the mean aggregate IPSP (3.6 mV) digitorum longus motoneurons, ankle flexors which
suggested that about 70 interneurons converge on each one would expect to show disynaptic IPSPs in re-
PBST motoneuron. sponse to MG la impulses. Longer latency IPSPs
Shape-index plots (half-width vs. rise time) indicate suggestive of trisynaptic action were also observed,
that these interneuronal terminals are located on the again mostly in ankle flexors.
482 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

RENSHAW CELLS. Another well-defined inhibitory in- to selectively inhibit small cells. However, the evi-
put to a-motoneurons is provided by Renshaw cells. dence discussed in section FIRING PATTERNS OF INDI-
These cells are situated in the ventral horn medial to VIDUAL MOTONEURONS AND MOTOR UNITS, p. 435, in-
a-motoneurons (186). They are activated by the re- dicates that recruitment occurs according to the size
current axon collaterals of motoneurons and respond principle for both fast ballistic movements as well as
to ventral root stimulation with a characteristic high- for slow ramp movements and that deliberate appli-
frequency repetitive discharge (97, 348). The output of cation of Renshaw cell inhibition does not silence
these cells is directed toward 0'- and y-motoneurons firing in small motoneurons before that in larger cells
(97), other Renshaw cells (306), and interneurons me- (see Fig. 24).
diating Ia inhibition to motoneurons (172). In each Analysis of the effects exerted by Ia-interneurons
case, their action is inhibitory. A more complete review and Renshaw cells on motoneurons requires consid-
of these topics is given by Burke and Rudomin (44). eration of the other inputs that drive these interneu-
Antidromic stimulation studies have revealed that rons as well as the synaptic relationship between them.
Renshaw cell axons can project to sites up to 12 mm Such a discussion is beyond the scope of this chapter,
apart along the spinal cord (190). These fibers are not but the reader is referred to recent articles and reviews
restricted to the gray matter, but run in the ventral on this subject (e.g., refs. 116, 171, and the chapter by
funiculi. Terminals to the motor nuclei appear to be Baldissera, Hultborn, and Illert in this Handbook).
limited to within 4 mm of the Renshaw cell. More
distant terminals apparently go to Ia inhibitory inter- Group Ib Input From Golgi Tendon Organs
neurons (190). There is no direct evidence concerning
the degree of convergence of Renshaw cells on moto- Tendon organs respond primarily to tension pro-
neurons; in particular, the amplitudes of their individ- duced by the active contraction of motor units directly
ual IPSPs are not known, so that comparison with in series with the receptor (166). Previous studies (100,
their aggregate IPSP is not feasible. Although the 229) indicated that activity in this projection to mo-
recurrent IPSP is finely graded upon ventral root toneurons was mediated by a circuit with one inter-
stimulation suggesting substantial convergence from neuron, i.e., a disynaptic afferent pathway, and was
these cells (97, 348), Renshaw cells discharge more inhibitory to homonymous motoneurons. The Ib-fi-
rapidly in response to increasing strength of stimula- bers do not send terminals directly into the ventral
tion. The increase in IPSP amplitude could, therefore, horn. Many of their terminals end in lamina VI (27,
represent temporal summation as well as spatial sum- 163, 347) where they synapse with a well-localized
mation. The finding that a single Renshaw cell gives group of interneurons (240). The complete projection
off terminals to motoneurons only within 4 mm of its of these interneurons is not known, but many of their
cell body suggests that divergence of single cells cannot axons go directly to motoneurons. In addition, such
include an entire motoneuron pool whose cells can interneurons receive extensive convergence from other
extend over as much as 10 mm (302). Although the segmental and descending pathways (242).
number of motoneurons contacted by a single Ren- The distribution of Ib effects to motoneurons of
shaw cell is not known (313), the distribution of recur- various species has been studied by means of spike-
rent inhibitory action seems to be restricted to the triggered averaging (345). In numerous homonymous
motor pool whose nerve is stimulated and to its sy- and heteronymous motoneurons IPSPs were seen
nergists (97). In addition, motoneurons receiving this after prolonged averaging (400-16,000 sweeps). The
inhibitory action tend to be located in close proximity responses were somewhat jagged and irregular despite
to the antidromically activated ones, which is not this extensive averaging, an indication that only oc-
surprising in view of the short-range projection of casional Ib impulses triggered the discharge of the
Renshaw cells to motoneurons (190). inhibitory interneuron that elicited the IPSP. Never-
There is evidence that activation of large motor theless, these potentials were more clearly defined
axons produces a much larger excitatory effect on than similarly evoked EPSPs found chiefly in the
Renshaw cell firing than activation of small motor motoneurons of direct and indirect antagonist muscles.
axons (193, 307). The notion that large motor axons The IPSP latencies ranged from 0.6 to 6.0 IDS, sug-
provide a more powerful input to Renshaw cells is gesting the existence of di- and trisynaptic linkages
further supported (307) by evidence that small reflex from Ib-fibers to motoneurons. The amplitudes of
ventral root volleys (in small fibers) elicit much these responses were small (5.2 ± 3.5 jlV), as would be
smaller Renshaw effects than similarly sized anti- expected in averaging over pathways involving more
dromic volleys (in large fibers). These observations than one synapse. These results, obtained with phys-
are further examples of the size principle; larger mo- iological excitation of receptors, confirm earlier, less
toneurons exert larger synaptic effects than small cells. direct methods in showing that impulses from Golgi
Since studies with electrical stimulation suggest that tendon organs exert disynaptic inhibitory effects on
Renshaw inhibition is most potent in small cells (97, both homonymous and heteronymous motoneurons.
131, 214), it has been suggested that during some They also confirm the existence of excitatory connec-
phasic movements large motoneurons might be able tions to motoneurons of antagonists. These studies
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 483

suggest that the projection from individual Ib-fibers is exists in the monkey. Comparison of the spontaneous
not sufficiently focused on the intemeurons to fire EPSPs, ranging from 0.2 to 0.6 mV in amplitude (319),
them reliably, judging from the relative infrequency of with the aggregate EPSPs (less than 1.2 mV) suggests
responses in motoneurons. Recent HRP studies (163) a restricted convergence of different fibers from the
on single Ib-fibers demonstrate that they distribute red nucleus on single a-motoneurons. Effects of mem-
collaterals at several rostrocaudallevels to intemeu- brane polarization indicate that terminals of rubro-
rons, much as la-fibers do. spinal fibers are located at similar distances from the
soma as la-terminals (320). The reticulospinal path-
Monosynaptic Input From way is uncrossed and projects both in cat and monkey
Descending Pathways to hindlimb flexors (134, 241), forelimb flexors and
extensors (349), and neck and back motoneurons (350).
A number of descending systems project monosyn- The small size of aggregate EPSPs again indicates
aptically to motoneurons. In particular, evidence exists relatively limited convergence from this source on a-
for direct projections from vestibulospinal, reticulo- motoneurons (318).
spinal, rubrospinal, and corticospinal systeIllS. How- The corticospinal pathway is largely crossed and
ever these projections differ substantially from one makes monosynaptic connections with flexor and ex-
another. In some cases, the nature of a given projection tensor motoneurons, supplying forelimbs and hind-
varies with the species of animal. This general subject limbs in the monkey, but not in the cat. In the cat,
has recently been reviewed by Burke and Rudomin individual fibers may send terminals to both cervical
(44). and lumbar cord (323). The aggregate EPSPs are in
The vestibulospinal tract is well organized in all general large compared to those produced by volleys
mammalian species. Two uncrossed descending sys- in reticulospinal and rubrospinal pathways. A much
tems exist: the lateral vestibulospinal tract originating stronger projection is directed to motoneurons con-
from the cells of the lateral vestibular (Deiters') nu- trolling distal than proximal muscles (65, 187, 282).
cleus and the medial vestibulospinal pathway, which Injection of intracellular currents (320) indicates that
originates from the medial vestibular nucleus. The the active terminals are more proximal than those
lateral system has a strong excitatory projection to from la-fibers. The smoothly graded nature of the
neck motoneurons and to hindlimb extensor motoneu- aggregate EPSP indicates that there is considerable
rons, particularly those of the knee and ankle (133, convergence from corticospinal axons onto a-moto-
349) but not to motoneurons supplying forelimb mus- neurons (187, 282).
cles (349). One descending fiber may influence neurons In primates some descending systeIllS are reported
in both cervical and lumbar segments (2). Monosyn- to terminate selectively on large a-motoneurons. Ru-
aptic vestibular effects on motoneurons are quite brospinal (8, 39) and large corticospinal (8) axons send
weak, with maximal composite EPSPs generally excitatory terminals to large motoneurons. Small cells
smaller than 2 mV (133, 241). Spike-triggered averag- in the same pool are apparently unaffected or, in some
ing studies have revealed a limited divergence from cases, are inhibited. Corticospinal axons of smaller
descending fibers to motoneurons with relatively small diameter distribute their terminals more uniformly to
individual EPSPs (less than 40 p.V). The time course motoneurons of all sizes. In cats these systeIllS do not
of these individual EPSPs as well as of aggregate project directly to motoneurons, but in humans there
EPSPs (292) suggests more proximal locations of ves- is apparently a larger direct projection. Detailed phys-
tibular terminals than of la-terminals generally. iological studies of these descending projections simi-
The medial vestibulospinal tract terminates exclu- lar to those on single la-fibers have not been carried
sively in rostral spinal cord segments. An unusual out. Other direct projections that are involved in
feature is the generation of monosynaptic IPSPs in movement, such as the vestibulospinal (8), exert excit-
neck motoneurons (349). This pathway has been stud- atory actions on both small and large motoneurons in
ied with the spike-triggered averaging technique and the same pool. At present there is little evidence that
it was found that the IPSPs were generally smaller higher centers and their projections are used in isola-
than 80 p.V. Their reversibility with somatic hyperpo- tion to achieve special patterns of recruitment within
larizing currents as well as their brief times to peak a pool.
(292) indicate that these IPSPs are generated at prox-
imal boutons. Single fibers in this tract project to less Topographic Factors Governing Development of
than 50% of the motoneurons in a pool. Connections of la-Fibers to Motoneurons
Two descending systeIllS with relatively little direct
input to motoneurons are the rubrospinal and reticu- One of the most important questions facing neuro-
lospinal tracts. In the cat only a small monosynaptic scientists is how the neurons in the brain establish
rubrospinal projection to a-motoneurons exists (164) their specific patterns of connections. The projections
although this crossed pathway provides terminals to of primary sensory neurons to motoneurons provide a
all levels of the spinal cord. A small monosynaptic specific example of this connectivity that has obvious
projection to motoneurons supplying distal muscles advantages for study. Since motoneurons to different
484 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

muscles are often adjacent to one another and slightly that the afferent fiber entered. The entry points of
intermingled in the spinal cord (44, 302), ingrowing these gastrocnemius fibers were extremely important
afferent fibers must have some way of establishing the in determining the size of the PSPPs they evoked.
necessary orderly connectivity. Various hypotheses These general findings indicate that a la-fiber gives off
have been advanced to explain the precision of such more terminals to motoneurons in the regions near its
connections. The chemoaffinity hypothesis fornlU- entry point. Connectivity is not determined entirely
lated by Sperry (326) proposes that neurons become by species specificity as some early findings may sug-
chemically differentiated during development depend- gest.
ing on the positions they occupy. This enables the To demonstrate the effect of the precise level of
axons of neurons to recognize matching or comple- entry on the amplitude of a PSPP, data from individ-
mentary labels on target neurons. Eccles (96) proposed ual experiments were utilized. Figure 45 illustrates
that the surface properties of motoneurons are the how the amplitudes of PSPPs depend on both the
basis for recognition by afferent fibers. The finding conduction velocities and entry points of the afferent
that the mean size of an aggregate EPSP is larger in impulses eliciting them. The conduction velocity and
homonymous than heteronymous motoneurons (99) the entry level of the afferent fibers are plotted on the
led Eccles to postulate that la-fibers establish connec- horizontal plane. The amplitude of the corresponding
tions with motoneurons by a process of "species rec- PSPPs are plotted on the vertical scale. Data from 11
ognition." The foregoing parts of this section provide out of 22 afferent units in this experiment were se-
considerable new evidence consistent with the hypoth- lected for display according to the following criteria.
esis of species specificity (see also refs. 225-228 for First, six units with the same entry level (2.0 ± 0.2
evidence concerning inherent species specificity of mo- mm) were chosen to illustrate the effects of their
toneurons). different conduction velocities. Second, seven units
Others have challenged the notion of a unique role with the same conduction velocities (84 ± 1.0 m/s)
for species specificity, suggesting "that there is a cer- were selected to show the effects of their different
tain stage in the development of most neural centers levels of entry. The illustration shows how the ampli-
during which they become topographically polarized tude of PSPPs increases with the conduction velocity
in such a way that the constituent neurons acquire of the impulses evoking them. A clear maximum of
some determining characteristic that establishes the PSPP amplitude can be seen for units entering the
spatial organization of the projection as a whole" (77). caudal part of SIDR (illustrating that PSPPs are larg-
Basing their arguments on genetic considerations and est when entry level and recording level coincide). The
analogies with sensory systems, Wyman et al. (355, PSPP size decreases for units entering the cord either
356) have suggested that topographic factors may more cranially or more caudally. The decrease in
account for connectivity in the spinal cord without amplitude with distance is, however, not symmetrical
specification of connections by neuronal species. A in both directions. These findings suggest that soma-
recent study with a new electrophysiological technique totopic factors are more important than previously
(156) provides considerable new evidence regarding believed in the development of these monosynaptic
the importance of such topographic factors. connections. The somatotopic distribution of termi-
Postsynaptic population potentials (PSPPs), evoked nals described here provides a possible explanation for
in large numbers of motoneurons by impulses in single the topographically localized stretch reflexes within a
la-fibers or spindle group II fibers, were recorded single muscle originally described by Cohen (67, 68)
simultaneously from two adjacent ventral roots per- and the focusing of signal transmission in the multi-
fused with isotonic sucrose. By comparing the ampli- channel muscle stretch system postulated by Wind-
tudes of these PSPPs with the entry points of the horst (351).
afferent fibers, the distribution of their terminals was
inferred. The la impulses entering a spinal segment Concluding Comments
evoked larger PSPPs in its motoneurons than in those
of the adjacent segment. The size of the PSPP was The action of a motor pool is determined to a large
correlated with the exact entry level of the afferent extent by the distribution of afferent input to its
fiber, indicating that la-fibers give off more endings to individual members of motoneurons. The best studied
motoneurons located near the entry point. Figure 44 afferent input to spinal a-motoneurons is provided by
reproduces records from a typical experiment that group la-fibers. The most striking result is that each
illustrate how the cord entry points of three afferent single la-fiber distributes the excitatory action to all
fibers influenced the amplitudes of the PSPPs their homonymous motoneurons even though these can
impulses evoked. The conduction velocity, entry point, number in the hundreds. In general the amount of
and muscle of origin of each of the three afferent fibers excitatory action (EPSP amplitude) is greatest in
are shown schematically in Figure 44A-C. The PSPPs small motoneurons. Consideration of the details of
recorded simultaneously from the lower part of LNR transmission at this synapse reveals a number of fac-
and from SIVR are reproduced. In each pair of records tors that might influence EPSP amplitude; these in-
the larger PSPP was recorded from the spinal segment clude presynaptic processes such as terminal invasion
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 485

A B c

L7 L7 L7 ,"L....
~MG LG
~-t ~ CV=90.2 m/sec ~{1 CV=98.5m/sec

f<r S, ¢:::=JMG
CV=80.5m/sec f<r S,
~ S,

i
I
,.o/'V]
~ 5.0mIMC ,
2 2 ~\

FIG. 44. Influence of entry level of la-fiber of cat on amplitude of postsynaptic population potentials
(PSPPs) its impulses elicit. Data are from a single experiment. Schematic drawings in top of figure
illustrate recording arrangements on ventral roots (VR); CV represents conduction velocity. Arrow
points at spinal segment that the afferent fiber enters. The PSPPs recorded from caudal part of 7th
lumbar ventral root (LNRl and 1st sacral ventral root (1) and 1st sacral (SNR) (2) are reproduced
below. A: la-fiber arising in medial' gastrocnemius (MG) enters through S, dorsal root. B: la-fiber
from MG enters spinal segment L,. C: la-fiber from lateral gastrocnemius (LG) enters spinal segment
L,. Each PSPP was averaged 4,096 sweeps. [From Luscher, Ruenzel, and Henneman (246).]

and release from boutons, both of which may be behaves in a uniform manner, indicating that the
subject to external control. Postsynaptic factors in- organization of these motoneuronal inputs is ex-
clude location of synapses on the somadendritic tree tremely important.
as well as motoneuron properties such as resistance. A
discussion of the possible mechanism generating the NONUNIFORMITY OF MOTONEURONS
inverse relation between EPSP amplitude and moto-
neuron size is deferred to section HOW SIZE OF MOTO- Motoneurons differ in their shape, size, connections,
NEURONS DETERMINES THEIR SUSCEPTIBILITY TO DlS- and electrical properties. Whether these differences
CHARGE,p.491. are sufficiently basic to classify motoneurons into dif-
The global projection of la input to homonymous ferent types or species is a major issue in this field.
a-motoneurons stands in marked contrast to the or- The facts that the contractile properties of muscle
ganization of all other segmental (heteronymous la, fibers are so strongly influenced by the discharge
interneuronal) and descending systems where single patterns of their motoneurons and that muscle fibers
fibers appear to have a much more restricted projec- are classified into at least three types, have led inves-
tion to a-motoneurons. At present it is not known tigators to look for corresponding types of motoneu-
whether these act to fractionate the motor pool, in rons. To date there is no clear basis, however, in their
effect preventing its collective action, or whether these histochemistry or ultrastructure for making sharp dis-
systems are themselves activated in a manner that tinctions between motoneurons.
ensures the stereotyped action of the pool expected The differences described below are essential for a
from the group la-fiber input. Information presented proper appreciation of the organization of the moto-
in section FIRING PATTERNS OF INDIVIDUAL MOTONEU- neuron pool. They are also of practical importance for
RONS AND MOTOR UNITS, p. 435, suggests that at least an understanding of certain neuromuscular disease
under most experimental conditions, the motor pool processes, as is noted below.
486 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

, - - - - - - - - - - , - - - - - - - - - , - - - - - - . - - - - - - + - - -•. - - - - - , 3 . 0

>~

c..
2.0 c..
rJ)
c..
u.
0
w
0
::::>
l-
to :::::i
c..
~
«

-2.0 -1.0 0.0 1.0 2.0

ENTRY POINT OF AFFERENT FIBER (mm)
FIG. 45. Three-dimensional graph illustrating relationship between amplitude of postsynaptic
population potentials (PSPPs) and conduction velocity and spinal entry point of afferent impulses
evoking them. Amplitudes were averaged over 4,096 sweeps and were recorded from 1st sacral ventral
root. Data were derived from a single experiment. All afferent fibers had origin in medial gastrocnemius
muscle. [From Liischer, Ruenzel, and Henneman (246).]

Early Classification of Tonic and discharge that went on for a long time, and 2) a phasic
Phasic Types of Motoneurons discharge, consisting of only one or two impulses dur-
ing the rising phase of the stretch. Intermediate units
As a result of earlier studies, Granit believed that were seen but were given little attention. The phasic
there were probably two types of motoneurons, those units generated larger spikes in general than the tonic
that subserve long-lasting "tonic" or postural motor ones in the ventral root recordings.
functions and those that initiate brief "phasic" re- Following this work, the terms tonic and phasic
sponses. To determine whether these tonic and phasic came into common use to distinguish two types of
motoneurons were basically different types of cells, motoneurons and their responses to stretch. In an
Granit and colleagues (129) devised tests to accentuate early paper Henneman (153) also noted that some
the differences between them. Recordings were made units tended to fIre tonically whereas others responded
from axons of ankle extensor motoneurons in ventral phasically, and that these fIring patterns were corre-
root filaments of decerebrate cats. After the response lated with motoneuron size. Later, Henneman and
of each unit had been recorded during three control colleagues (159) demonstrated, however, that moto-
stretches, the muscle nerve was stimulated electrically neurons should not be divided into fIxed tonic and
at 500 impulsesjs for 10 s with the myograph discon- phasic types, because, depending on the relation be-
nected. Beginning 20 s after the tetanus, the response tween their size and the existing level of excitatory
of each unit was reexamined every 10 s. In the post- and inhibitory input, they may change from one cat-
tetanic or potentiated state the responses to a constant egory to the other or behave in either mode.
10-mm stretch fell into two main categories: 1) a tonic Tonic and phasic are, therefore, appropriate terms
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 487

to characterize a given response, but not a particular

motoneuron.

Significance of Nonuniformity
of Muscle Fibers
Since the widespread use of histochemical tech-
niques, it has been apparent that most mammalian
muscles are composed of fibers differing widely in
diameter, mitochondrial density and distribution, cap-
illarity, myoglobin concentration, enzymatic content,
and background staining. Furthermore, when acti-
vated individually, the motor units in these muscles
show major differences in contractile tension and
speed, susceptibility to fatigue, and recruitment order.
There is considerable evidence that many of these
differences in muscle fibers and motor units are deter- 940ms
mined by the properties of the motoneurons that
FIG. 46. Maximal firing rates of 4 plantaris motoneurons of cat
supply them. Cross-innervation studies, initiated by with approximately equal critical firing levels recorded simultane-
Buller et al. (31) and repeatedly substantiated by ously during first 940 ms of their response to electrical stimulation
others, indicate that motoneurons play a dominant of plantaris nerve. Stimulus parameters adjusted to produce maxi-
role in determining the contractile properties of the mal firing rates. Bottom unit fires significantly slower than top 3
muscle fibers they innervate. The extent of this influ- units, giving impression of having been sampled from a different
population than the others. [From Harris and Henneman (145).]
ence is revealed strikingly by the appearance of new
types of muscle fibers in the previously homogeneous
soleus of the cat after reinnervation of the muscle by both of them discharged rhythmically (Fig. 47). In
the nerve to a mixed muscle (358). single experiments 10-15 of the smallest units in the
In conjunction, these findings suggest that the his- pool (all with CFLs in the 0%-8% range) were isolated
tochemical and physiological characteristics, which and compared. Statistical analyses and visual inspec-
have led to the identification of three types of muscle tion .of these small samples again suggested the exis-
fibers and motor units, may be due in part to differ- tence of two populations of motoneurons. Analysis
ences in the types of motoneurons supplying them. In also indicated that the FRs of units in single experi-
the past ten years Henneman and colleagues (159, 160, ments were not sampled from anyone of a variety of
253, 352) have demonstrated that many of the prop- parametric, single-modal distributions. This implies
erties of motor units and motoneurons are highly that the data were sampled from a distribution having
correlated with the size of the neural soma. Size-de- more than one mode, which indicates the existence of
pendent differences in motoneurons, however, would separate populations of motoneurons among the small
presumably lead to a continuous gradation in the units of the pool. Pooling of the normalized data from
histological appearance of muscle fibers rather than to different experiments revealed a bimodal histogram,
three distinct types. Gradations are apparent in the reinforcing the conclusion that there are two types of
appearance and properties of muscle fibers, but rec- small a-motoneurons in the plantaris pool. It is possi-
ognition of different fiber types indicates that factors ble that inputs to the motoneuron pool other than the
independent of cell size may also affect the differentia- group Ia volleys used in the determination of CFLs,
tion process. Le., inputs that might be effective only during repeti-
tive stimulation of the muscle nerve, could result in
Motoneuron Properties Independent of Size more synaptic drive on one cell than another. In a
recent study (58) six different sources of inhibitory
Harris and Henneman (145) analyzed this problem input did not alter the CFL of motoneurons or their
by isolating single motor axons of the plantaris pool in relative rank order during repetitive firing. This argues
ventral root filaments of decerebrate cats, measuring against the possibility that certain motoneurons in a
their cortical firing levels (CFLs), and comparing their pool have completely different sets of inputs than
firing rates (FRs) during repetitive stimulation of the others, but it does not entirely dispose of input as an
plantaris nerve. Units in the 0%-8% range of CFLs explanation for differences in FR. Any systematic
(thus varying little in size) were found to differ either difference in input to cells of similar size that was
very little or quite widely in maximal FR, as illustrated sufficient to account for fast- and slow-firing popula-
in Figure 46. Apparently these units were sampled tions of motoneurons might be considered an adequate
randomly from two populations, one firing rapidly, the basis for defining different types of cells based on their
other slowly. The ratio between the two rates re- connectivity. There is, however, reason to be doubtful
mained approximately constant regardless of the in- about drawing this conclusion. Chapple (55) has shown
tensity or rate of input the units received, as long as that pairs of bilaterally homologous motoneurons in
488 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

A
25 0/
000

....
'"
Q.
Q.
0
00
/
0

CO
~ • ~
.
LL. 0
o 15 • ••
/
0

/
W

«
/ ..
~

a:

• •
~
5 15 25 35
RATE OF A (pps)

.. .. .-
B

2.0
•
-.- •
~-.
.... -
••
_- •

« CO
LL. LL.
0 0
w w 0 0

<
0 0
~ - 0
-0-00 0 000 - 0 -
a: ~ 1.0
0 0

25 50 75 100%
RATE OF A / MAX RATE OF A
FIG. 47. Firing rates (averaged over fIrst second) of2 pairs of units of cat with similar critical fIring
levels, compared simultaneously. Data obtained by applying a wide range of stimulus intensities and
frequencies to plantaris nerve. A: for each pair of units fIring rate (FR) of 1 unit is plotted versus FR
of other unit. Dashed lines indicate linear regression. B: ratio of FRs of each pair plotted as a function
of the rate of 1 of the units expressed as a percentage of its maximum rate. Bottom pair (open circles)
had very similar FRs (ratio"" 1.0); top pair (closed circles) had widely differing FRs (ratio> 2.0).
Linear regression lines with slopes near zero indicate ratio of FRs remained constant for all these
inputs. [From Harris and Henneman (145).]

the hermit crab are physiologically asymmetrical in to establish control values and also during added stim-
their tonic frequencies. Those on the left side fIre at ulation of various inhibitory nerves. Criteria were de-
higher frequencies than those on the right. The con- veloped whereby pairs of units, previously classifIed
duction velocities of the motoneurons on the two sides by their FRs, could be independently classifIed ac-
did not differ appreciably and no signifIcant differ- cording to their responses to inhibitory inputs. The
ences in the axon or soma diameters of motoneurons ability to distinguish motoneurons consistently by
injected with cobalt were apparent on the two sides. more than one set of criteria reinforces the evidence
Chapple concluded that the source of the symmetry summarized above, that physiologically different pop-
was not in the motoneurons themselves but arose ulations of cells exist within a single motoneuron pool.
centrally in the absence of any major peripheral asym- The fIndings indicate that cells that respond unequally
metry. to inhibition either receive different amounts of input
Further evidence regarding differences in the prop- from certain inhibitory sources or react differentially
erties of motoneurons that are not dependent on size to the same amounts of these inputs.
has recently been obtained by Harris and Henneman Another indication that the motoneurons innervat-
(146). Motoneurons of similar size were compared in ing different types of motor units may vary in their
FR during repetitive stimulation of the plantaris nerve essential properties appears in a recent study by Burke
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 489

and his colleagues (46). The maximum amplitudes of fibers fired repetitively (i.e., more than once) to an
aggregate EPSPs produced in MG motoneurons were intracellular current step. All but two of the repeti-
found to be related to the type of muscle fibers they tively firing cells continued discharging for the dura-
supplied. On the average, EPSPs were the largest in tion of the current pulse (usually 1 s). The average
motoneurons of type S units, somewhat smaller in current applied to produce rhythmic firing was only
type FR units, and smallest in type FF units. Group Ia about 30% above rheobase for a single spike. Currents
EPSPs were also correlated inversely with motor unit of up to 175% above rheobase for a single spike were
tension production and directly with resistance to used in testing motoneurons that did not respond
fatigue. repetitively.
If different types of motoneurons are present in the All the motoneurons examined had resting mem-
same pool and the sizes of their aggregate EPSPs are brane potentials of 50 mV or greater. When the mo-
correlated with the types of muscle fibers they supply, toneurons were grouped according to the amplitude of
the order of recruitment should perhaps correlate their antidromic action potentials, those between 57
more closely with motor unit type than motoneuron and 78 mV had many more nonrepetitive cells than
size. Faden and Zajac (110) investigated recruitment those between 79 and 96 mV, indicating that injury
order in motor units of known type by dissecting out due to penetration correlates with nonrepetitive firing.
small, unsevered ventral root filaments. Axonal con- Nevertheless, all the motoneurons supplying slow-
duction velocity was measured for each unit and the twitch units fired repetitively whereas only one-half to
tension and EMG they produced when stimulated two-thirds of the fast units did, depending on the
were recorded. After identification of motor unit type minimum amplitude chosen.
according to the criteria of Burke et al. (41) the intact These groups could also be differentiated on the
filaments were cut and recruitment order was exam- basis of their rheobase. Cells with antidromic action
ined by direct comparisons of firing in different fila- potentials greater than 72 mV had rheobasic values as
ments. A variety of reflex maneuvers was used in these follows: fast, nonrepetitive units, 19.0 ± 9.3 nA; fast,
comparisons, including tendon taps, muscle stretches, repetitive units, 7.9 ± 6.6 nA; slow, repetitive units, 3.7
repetitive shocks to the muscle nerve, and single ± 2.9 nA.
shocks after posttetanic potentiation. In all of the 11 Mishelevich's results are consistent with the general
pairs of units studied in these difficult experiments, lack of accommodation in motoneurons supplying
the unit producing the largest tension, fatiguing most slow-twitch fibers and with the various degrees of
rapidly, and having the highest conduction velocity accommodation found in cells innervating fast-twitch
was recruited last. The unit with the higher conduction fibers.
velocity had the larger tetanic tension in 60 out of 76 Kernell (200-202) concluded that most cat motoneu-
comparisons, with the reverse in 4 cases. There were rons in good condition will respond repetitively to
12 "ties." current injection. The currents he used, however, often
The authors conclude that "these results are con- far exceeded those that may occur normally. With the
sistent with the hypothesis that recruitment order is more stringent criteria used in Mishelevitch's study,
solely based on motoneuron size, which, in turn, is distinctions in repetitive response can apparently be
highly correlated with the size and fatigability of the made between motoneurons innervating different
innervated motor unit." Units defined as FF, FR, and types of muscle fibers. These distinctions are not cor-
S were included in the study. related with any known differences in synaptic input
to motoneurons; they presumably depend on the mem-
Differential Responses of brane properties of the motoneurons.
Motoneurons to Injected Currents
The responses of motoneurons to current injected
Influence of Muscle on Developing
through an intracellular electrode have been studied
and Mature Motoneurons
by a number of investigators (114, 130,200-202,276). It has been known for some time that muscles exert
Although injected current is not a natural input, it is an important influence on the motoneurons that in-
a useful means of studying the membrane properties nervate them. The full extent of this influence has not
of motoneurons. None of the papers cited above at- been determined but is under active investigation.
tempted to relate the firing patterns of the motoneu- Many motoneurons die during early stages of devel-
rons to the properties of the muscle fibers they sup- opment. This probably occurs because they fail to
plied. Mishelevich (260) has investigated this problem. form adequate connections with muscle (see ref. 143).
Motoneurons ofthe LG, MG, SOL, and plantaris (PL) If the limb bud of an embryo is removed before motor
were designated as fast units if their motor unit twitch innervation is established, all the corresponding mo-
times were 30 ms or less; they were classified as slow toneurons degenerate (142). If an additional limb bud
if their twitch times were 40 ms or more. Slightly more is grafted onto the embryo, enlarging the peripheral
than one-half of the motoneurons supplying fast- target organ (162), fewer motoneurons degenerate
twitch muscle fibers and all those innervating slow than normally. Development of motoneurons appar-
490 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

ently requires the presence of, or connections with, from mild to maximal forces; 4) acceleration of firing
skeletal muscle. The mechanisms underlying these rates with stronger contractions to a maximum of 18-
developmental effects are obscure. 20/s. In this group of cases open biopsy disclosed that
Partial denervation of the soleus muscle in adult 95%-99% of the muscle fibers of these normally het-
cats alters the electrical properties of the soleus mo- erogeneous muscles (biceps brachii or quadriceps fe-
toneurons whose axons are intact as well as those moris) were small and red.
whose axons are severed (170). Section oflumbosacral Motor units from another group of patients had
dorsal roots does not produce these effects, hence they completely different characteristics: 1) sudden or vig-
are not mediated by sensory impulses (221). Czeh et orous contraction required for activation; 2) either
al. (84) pursued this problem by showing that transec- discharge of brief bursts lasting 0.5-5.0 s with no
tion of the thoracic cord resulted in a significant capacity to sustain them longer (during this time they
decrease in the duration of the afterhyperpolarization fired steadily at 1O-25/s or sputtered irregularly) or
(AHP) following action potentials. This change, which discharge only once or twice; 3) peak firing frequencies
occurred within 8 days, was prevented by daily stim- reaching 16-50/s at the height of bursts accompanying
ulation of the sciatic nerve. The AHP of soleus moto- maximal contractions; 4) more rapid initial and peak
neurons was also shortened appreciably within 8 days discharge frequencies with increasingly vigorous con-
after blocking conduction in the soleus nerve with a traction. Biopsies from these patients contained 96%-
tetrodotoxin cuff. This change was prevented by 100% pale (type A) fibers. In brief, the normal recruit-
chronic stimulation of the soleus nerve peripheral to ment pattern, with its wide range of potential sizes
but not central to the cuff. It was concluded that and frequencies, was lacking in both these groups.
motoneuron properties in an adult probably depend It is a reasonable, but unproven, inference that the
partly upon some factors associated with activity of motoneurons supplying one type of muscle fiber in
the innervated muscle and that such factors are car- these cases were susceptible to a disease process to
ried in the motor axons by retrograde axoplasmic flow which the motoneurons supplying the other type of
rather than by nerve impulses. In line with this rea- muscle fibers were resistant. The muscle fibers sup-
soning is the finding that disuse atrophy of muscle, plied by the susceptible motoneurons were denervated
which occurred after cord section in these experi- as a result. The motoneurons innervating the healthy
ments, has been shown to be associated with a de- muscle fibers apparently remained relatively intact.
crease in the rate of protein synthesis and an increase These cells presumably put out axonal collaterals that
in its rate of degradation (126, 127). Other mecha- supplied the denervated muscle fibers and in time
nisms, however, are also considered in the authors' converted them to the type characteristic of their own
discussions. activity pattern, as in cross-innervation experiments.
To date there is no evidence indicating whether There are puzzling features in these studies. The
different muscle fibers exert type-specific retrograde findings suggest that the motoneurons supplying het-
effects on their motoneurons. The results of cross- erogeneous muscles may differ significantly in their
innervation experiments, in which motor nerves orig- susceptibility to disease. The extreme uniformity of
inally supplying fast muscles cause the development the muscle population that eventually results is sur-
of three types of muscle fibers in the homogenous prising and difficult to explain. It is unlikely that the
soleus muscle, suggest that the influence of motoneu- motoneurons themselves are as homogeneous as the
rons on muscles may be more powerful than the ret- appearance of the muscle fibers suggests. The con-
rograde effects of the muscle on the motoneuron. The verted muscle fibers appear to be normal as judged by
latter, however, may simply be more subtle influences their electrical properties. In five cases showing selec-
that appear less potent. tive, widespread, moderately severe atrophy of large,
pale fibers, and in four cases with milder involvement,
Evidence From Human Disease the amplitude and duration of the low-threshold uni-
tary potentials appeared normal, suggesting that they
When an abnormal pattern of recruitment is re- were derived from normal red fibers not then involved
corded from a muscle, biopsy is almost certain to in a disease process. In seven of these nine cases the
reveal pathology. Electromyography (EMG) of the only recognizable EMG abnormality was failure to
most superficial parts of the muscle followed by an achieve a full "interference pattern" of large and small
open biopsy of 2-5 rom of muscle around the tip of the spikes with maximal effort. These studies seem to
electrodes may lead to a diagnosis of long-standing require the existence of at least two types of motoneu-
myopathy even in muscles that appear grossly normal. rons. Variations in motoneuron properties or in the
In a survey of 110 selected patients, Warmoltz and properties of the muscle fibers they innervate may
Engel (344) discovered a small group of patients with account for the differences in susceptibility to disease
the following characteristics: 1) initial activation on that apparently exist. However, the recent work al-
minimal to mild voluntary contraction; 2) well-main- ready described on motoneurons with similar critical
tained rhythmic firing with little effort at frequencies fIring levels but widely differing fIring rates suggests
of 6-10 s; 3) sustained discharge through all ranges that there are at least two populations of motoneurons
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 491

with size-independent properties sufficiently different aggregate EPSP = synaptic current

to account for different susceptibility to disease. X input resistance of cell (1)
In the next subsection some of the properties of
Concluding Comments motoneurons that influence Equation 1 are described.
The various types of evidence described in this
section clearly indicate that the motoneurons in a pool Properties of Motoneurons That
are not uniform in their basic properties. Tentatively, Influence Susceptibility to Discharge
one may conclude that there are at least three types
As Equation 1 suggests, the input resistance of the
of motoneurons in a pool. Differences in their fIring
motoneuron is an important factor in determining the
characteristics presumably account for the widely dif-
size of its aggregate EPSPs and its susceptibility to
fering properties of the muscle fIbers they supply.
discharge. This resistance varies inversely with the
Until clear distinctions between the histochemistry,
surface area of the cell's membrane, directly with its
ultrastructure, or membrane properties of motoneu-
specifIc resistance per unit area and to some extent
rons can be demonstrated, however, it seems prema-
perhaps with the density of the synaptic boutons that
ture to classify motoneurons into specifIc types as has
partly cover the surface. According to Conradi (72)
been done for muscle fIbers. This is an area with
the boutons may cover more than 50% of the moto-
important clinical implications, as indicated in subsec-
neuron surface.
tion Evidence From Human Disease, p. 490.
The importance of input resistance in determining
the size of synaptic potentials was demonstrated by
HOW SIZE OF MOTONEURONS DETERMINES Katz and Thesleff (199). When miniature endplate
THEIR SUSCEPTIBILITY TO DISCHARGE potentials (MEPPs) were recorded from different frog
muscle fIbers, their mean amplitudes varied by factors
Analysis of the fIring patterns of motoneurons leaves of more than 10 to 1, although the mean quantum of
little doubt that susceptibility to discharge is highly acetylcholine eliciting the MEPPs was approximately
correlated with motoneuron size. This conclusion constant. By determining the input resistances of the
poses a difficult question. Why are small motoneurons individual muscle fIbers, which are a function of their
fIred more readily than large ones by almost all types diameters, Katz and Thesleff found that the greater
of excitatory inputs, and why are they more resistant the resistance, the larger the MEPP.
to silencing by inhibition? This result was important to subsequent consider-
Kuno (214) has shown that the resting membrane ations of the relationship between cell size and suscep-
potentials of small and large motoneurons do not differ tibility to discharge. The application of this fInding to
and that approximately the same degree of depolari- the motoneuron, however, has been difficult, because
zation is required to discharge them. If their fIring the surface area, which determines the input resist-
thresholds are not signifIcantly different, then the ance, also determines the number of boutons on the
properties of the motoneurons themselves, the input cell (122) and, thus, influences the total synaptic cur-
to them, or both of these factors in combination must rent generated under various conditions.
somehow result in larger depolarizations (EPSPs) in Although the input resistance of the motoneuron
small cells than in large ones. A simple experiment does not per se account for its susceptibility to dis-
illustrates that this inference is correct, at least under charge, it has been shown to be a good index to its
special experimental conditions. If a muscle nerve is size. Kernell (203) showed experimentally that the
stimulated so that a synchronous volley is conducted input resistance of a cell is inversely related to its
centrally in all of its la-fIbers, the effect of this stand- axonal conduction velocity and, thus, to its dimen-
ard input can be examined by recording the "compos- sions. Barrett and Crill (13) and Cullheim (80) ex-
ite" or "aggregate" EPSP it produces in motoneurons tended Cajal's original fIndings by demonstrating spe-
of the homonymous pool. These aggregate EPSPs cifIcally in motoneurons the close relationship between
vary inversely in amplitude with the size of the mo- axon diameter in the white matter and at the initial
toneuron from which they are recorded (36, 99). An segment, and between axonal conduction velocity and
explanation for this particular correlation has recently cell size. Barrett and Crill (13) and Lux et al. (247) also
been found (245), which also accounts for the more confIrmed Kernell's (203) finding that axonal conduc-
general relationship between the size of motoneurons tion velocity and input resistance are inversely corre-
and their susceptibility to discharge. This is described lated by carrying out electrical measurements on im-
in the subsection entitled Role of Input in Determin- paled motoneurons before filling them with a dye and
ing Susceptibility to Discharge, p. 493. later examining them morphologically.
According to Ohm's Law (E = IR), the aggregate In view of the preceding observations, it is not
EPSP (E) is the product of the total synaptic current surprising that several investigators have shown that
(I) produced by all of the active la-endings on a the aggregate EPSP of a motoneuron is directly pro-
motoneuron and the resistance (R), which the mem- portional to its input resistance. Eccles et al. (99) were
brane of the soma and dendrites offers to this current. the fIrst to note that small motoneurons have larger
492 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

aggregate EPSPs than bigger ones. The five tables in which determines the input resistanc~, also determines
which their observations are compiled summarize a the number of boutons on a cell. The additional num-
great deal of data concerning the size of aggregate ber of boutons on large cells may exactly compensate
EPSPs for many different species of motoneurons. for the decrease in input resistance, as Zucker's (360)
Kernell (203) reasoned that with synapses of similar calculations indicate. Barrett and Crill (13) have dem-
size and location the current generated by a given onstrated that small motoneurons have a slightly
conductance change should alter the soma membrane higher specific resistance than large ones. This sug-
potential more in a cell with a high than in one with gests that input resistance is disproportionately higher
a low input resistance, as Katz and Thesleffs (199) in small cells and might explain their larger EPSPs.
findings predicted. Thus, he argued, less current (for Other systematic differences between large and
example, a smaller number of active synapses) would small motoneurons might also help to explain the
presumably be needed for eliciting a repetitive dis- observed relationship between EPSP amplitude and
charge in a cell with a high input resistance. Using cell size. For example, if small motoneurons had a
four different methods of measuring input resistance, shorter total electrotonic length (L), la-synapses
he found that cells with slowly conducting axons had might be closer on the average to the soma, thereby
higher resistances than those with rapidly conducting causing larger EPSPs. Burke and ten Bruggencate (48)
axons. have shown, however, that there is no systematic
Accepting "the notion that there is a relation be- relationship between cell size and L and it has already
tween motoneuron size, cell input impedance and am- been noted that la-terminals on small cells are not
plitude of synaptic potentials," Burke (36) set out to closer to the soma (subsection PHYSIOLOGICAL
examine the basis for these relationships by studying ANALYSIS OF LOCATION OF BOUTONS, p. 468).
the characteristics of group la synaptic input to triceps Before concluding this section, it should be noted
surae motoneurons. A significant positive correlation that there is strong experimental support (see review
was found between the maximum amplitudes of ho- in ref. 44) for the concept that the all-or-none action
monymous aggregate EPSPs evoked electrically and potential of a motoneuron is generated in the initial
the input resistance values across the entire popula- segment of the axon (axon hillock plus nonmedullated
tion of units studied. A similar correlation, though segment). According to Eccles et al. (99), the aggregate
somewhat less strong, was found for heteronymous EPSP "will be virtually identical with the depolariza-
EPSPs. The amplitudes of "unitary miniature EPSPs tion produced in the initial segment." This belief was
(mEPSPs) of presumed group la origin," elicited by consistent with the observation that "homonymous
small static stretches were also examined. A significant and heteronymous EPSPs are indistinguishable with
positive correlation was found between the median respect to the level of depolarization at which they
amplitudes of the mEPSP distributions and the input initiate the discharge of impulses" (99). It is possible
resistance values. Positive correlations were also noted that the threshold of the initial segment may influence
between the amplitudes of the median mEPSPs and the susceptibility of a motoneuron to discharge, but
the maximum homonymous aggregate EPSPs in the Kuno's (214) finding that the same degree of depolar-
cells for which both data points were available. The ization is required to fire large and small cells does not
duration of the aggregate EPSPs in high-resistance support this notion.
(type S) cells tended to be significantly longer than in Although differences in the amplitudes of aggregate
lower resistance (type F) cells. This observation could EPSPs of small and large cells are undoubtedly useful
not be entirely accounted for by the relatively small in explaining susceptibility to discharge, it is also
difference in the mean time constants of the two types important to keep in mind that motoneuron threshold
of cells. For these reasons and because of a theoretical is sometimes considered to be that level of membrane
analysis he carried out, Burke concluded that "the potential at which the cell begins to discharge repeti-
density of la synaptic terminals tends to be higher on tively. A property of motoneurons that may be crucial
type S motoneurons than on type F cells." This valu- in allowing them to fire repetitively in response to
able set of data and the theoretical analysis accom- steady current is the slow potassium conductance that
panying it led Burke to the conclusion that small is turned on following an impulse (205). Using com-
motoneurons with a high input resistance are more puter modeling techniques, Traub (338) has shown
susceptible to discharge because they receive a higher that the time constant governing the decay of this
density of la-endings than large cells, although addi- slow potassium conductance may be smaller in large
tional factors were believed to be important as well cells than in small ones. This postulate is apparently
(36). necessary to obtain the correct impulse frequency vs.
Despite the generally agreed facts that input resist- injected current relationships in small and large cells.
ance is a function of motoneuron size and surface area It emphasizes the importance of inherent membrane
and, as Equation 1 indicates, is an important factor properties in addition to geometrical considerations
governing the size of EPSPs and susceptibility to and organization of synaptic input in determining how
discharge, its role vis-a.-vis other factors is difficult to susceptibility to discharge is established. Burke and
define precisely. This is because the surface area, Nelson (43) have demonstrated that large motoneu-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 493

rons have a tendency to exhibit greater accommoda- possibility that impulses in them will die out at certain
tion than small ones, another example of a size-related points and the smaller will be the percentage of la-
property that may contribute to the functioning of the endings that are activated by maximal volleys in its
size principle. Although motoneuron size is not the la-fibers. Anything than facilitates invasion of la-ter-
only variable influencing recruitment order, it appears minals and activation of synaptic endings should lead
that other important factors are also functions of cell to larger aggregate EPSPs, and this tendency should
size. be greater in large than in small motoneurons.
Finally, in attempting to evaluate the general sig- To test this hypothesis, intracellular recordings were
made from motoneurons of anesthetized cats. First,
nificance of input resistance, it is pertinent to consider
whether it bears the same relationship to inhibitory as the input resistance (IR) of each motoneuron was
to excitatory inputs. The data needed to answer this measured, which provided a reliable estimate of its
question are less numerous and not as easy to interpret size (13, 203). Then the medial gastrocnemius nerve
for inhibitory as for excitatory inputs, because inhibi- was stimulated with brief shocks (l/s) that set up
tory inputs are not as precisely controlled, the re- afferent volleys in all of its la-fibers. The monosyn-
sponses are harder to measure, and the afferent cir- aptic EPSPs evoked by the volleys were recorded and
cuits are generally at least disynaptic. Nevertheless, ita series of 16 responses was averaged electronically
has been found in a number of experimental circum- and written out on an x-y plotter. The muscle nerve
stances that IPSPs are larger in smaller motoneurons was then tetanized at 500 impulsesjs for 10 s and a
(46, 131, 152, 214). In the study by Burke et al. (46) posttetanic series of EPSPs was recorded and aver-
the same motoneurons that yielded large aggregate aged. Lloyd (237) showed that conditioning tetani of
EPSPs also exhibited large disynaptic IPSPs. Hence this frequency and duration caused hyperpolarization
it is likely that input resistance or cell size influences
of the la-fibers and a dramatic increase in the ampli-
the responses to all inputs, both excitatory and inhib- tude of the monosynaptic reflexes elicited by volleys
itory. in them. This "posttetanic potentiation" (PTP) grad-
ually declined back to control levels in 2-3 min.
Role of Input in Determining Figure 48 illustrates typical effects of PTP on the
Susceptibility to Discharge EPSPs of three motoneurons of different sizes. The
lower traces in the three pairs of records illustrate that
A number of theoretical analyses have been made the average size of the aggregate EPSPs before poten-
to explain the size-dependent susceptibility of moto- tiation varied directly with the input resistance of the
neurons to discharge by la volleys (36, 329, 338, 360). motoneuron. This confirmed previous observations
These formulations relied on the experimental data (36). The upper traces are the averages of 16 EPSPs
that were available, but they also depended on a following the tetanus. The percent potentiations in A,
number of stated and unstated assumptions that were B, and C were 123, 92, and 11, respectively.
of uncertain validity. It was apparent that until these Figure 49A plots the relationship between input
theories could be tested experimentally they had to be resistance and amplitude of control EPSPs in data
regarded chiefly as speculative exercises. pooled from nine experiments. Figure 49B illustrates
Recently, however, a hypothesis was developed that how percent potentiation was related to input resist-
could be tested experimentally (245). The essential ance in the same experiment. As the hypothesis pre-
features of this hypothesis are as follows. Since mo- dicts, the greatest potentiation occurred in motoneu-
toneurons of different sizes reportedly have equal den- rons with the lowest IR, i.e., in the largest cells. The
sities of synaptic boutons (122), those with large sur- least potentiation and the largest control EPSPs were
face areas must have proportionally more boutons. In found in motoneurons with the highest IRs. Less
particular, if large motoneurons supplying the medial branching would be expected in la-fibers approaching
gastrocnemius have more endings from la-fibers com- these small cells, and few inactive endings should be
ing from that muscle, this is due not to receiving available to increase synaptic input during PTP. It
projections from more la-fibers, but to more extensive has generally been assumed that PTP is due to release
arborization of individual la-fibers on these cells, be- of more transmitter from a fixed number of active
cause each motoneuron receives terminals from essen- endings. This assumption, however, predicts that the
tially all la-fibers coming from its muscle (255). The effect of PTP is independent of cell size, so it is no
more numerous branch points, terminal boutons, and longer tenable.
boutons en passant in extensive arborizations presum- If the hypothesis is correct, PTP might be expected
ably lower the safety factor for propagation (177). to cause small increases in the conduction time re-
Experimental evidence indicates that an impulse in a quired for la volleys to invade additional terminals
motor axon may fail to invade all of its branches (213) and to activate more synaptic knobs. The flow of
and that transmission failures may occur in la-syn- synaptic currents at the various endings might be less
apses on motoneurons (105-107). It follows that the synchronous, i.e., more prolonged due to these delays
larger the motoneuron, the more extensive are the (177), but it is difficult to predict the effects of PTP
terminal arborizations on it and the greater is the on the time course of EPSPs without precise infor-
494 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

A B c
~I ~I N

L 5.0 m sec

IR:0.3 Mn
IR:3.0Mn
FIG. 48. Examples of aggregate excitatory postsynaptic potentials (EPSPs) recorded from 3 medial
gastrocnemius motoneurons of cat to illustrate effects of posttetanic potentiation on cells of different
input resistance (IR) and size. Bottom traces illustrate that average size of aggregate EPSPs before
potentiation varied directly with IR of motoneuron. Top traces are averages of 16 EPSPs following
tetanus. Percent potentiations in A-C were 123,92, and 11, respectively. [From Luscher, Ruenzel, and
Henneman (245).]

mation about the geometry of the terminal arboriza- fore, represents compromises between conflicting
tion of la-fibers. Figure 50 is a plot of the changes in needs. Its characteristics suggest that four require-
rise time in EPSPs versus the changes in their half- ments took precedence and played major roles in
widths. With a few exceptions PTP increased both rise determining the final organization. Survival, with its
time and half-width, but the magnitudes of these emphasis on speed, endurance, and agility of move-
changes do not seem to be related. ment, apparently placed a premium on 1) maximal
These experimental findings are consistent in all sensitivity in the control of muscular tension and 2)
respects with the hypothesis that has been proposed. maximal economy in expenditure of contractile en-
They suggest that invasion ofla-terminals is a graded ergy. Less obvious, but increasingly essential as the
process that is generally more complete in the terminal complexity of the central nervous system increased,
arborizations on small cells because they have fewer was the need for a mechanism that provided 3) collec-
branch points. The percentage of la-boutons that are tive action from a large population of neurons and 4)
activated by afferent impulses depends on this process. as nearly automatic operation as possible at the final
The surface area that a motoneuron presents to ap- output stage. These aspects of the neuromuscular
proaching la-fibers may influence their tendency to system are discussed briefly in the concluding sections
branch and, thus, indirectly it may determine its own of this chapter in order to provide further insight into
susceptibility to discharge. the organization of the motoneuron pool.

SOME PRINCIPLES UNDERLYING

How Sensitivity in Gradation
ORGANIZATION OF MOTONEURON POOLS
of Tension is Achieved
The experimental results in section FIRING PAT-
Although this chapter has been devoted primarily TERNS OF INDIVIDUAL MOTONEUTRONS AND MOTOR
to the organization of the motoneuron pool, it should UNITS, p. 435, reveal the basic principle involved in the
be apparent that the properties of motoneurons and grading of· muscular tension. The first step in the
of the muscle fibers they supply are so interdependent grading process is that the motoneuron pool functions
that neither can be understood in isolation from the incrementally by adding progressively larger units to
other. In the course of evolution all of the features of its total neural output as the net excitatory input
the combined neuromuscular system have been sub- increases. As excitation decreases or inhibition in-
jected to selection pressures. Some capacities that creases, the pool operates decrementally by subtract-
survival might tend to optimize, however, are mutually ing progressively smaller units from its output. These
exclusive. The control system that has evolved, there- operations depend on the distribution of cell sizes in
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 495

the highest ranking cell discharged by a particular

~ A input.
Equation 2 is a concise formulation of a basic law of
a.. 10 combination that specifies how the activities of mo-
en
a.. toneurons are combined in monosynaptic reflexes. The
w same law applies when the monosynaptic reflex is
...J conditioned by various mixtures of excitation and in-
oa: hibition. Figure 51 illustrates some of the features of
!z 5
pool organization schematically. Each vertical line in
o the figure represents a motoneuron whose relative size
o is indicated by its amplitude. Small motoneurons,
u. which are most numerous, are finely graded in size
o and highly susceptible to discharge. Large cells are far
less numerous, differ more widely in size, and require
~ o'-- -'--- -'- ....I.- ...J
much larger inputs to be fired. Resistance to inhibition
~
« 0.0 2.0 3.0 4.0 is distributed in the same general pattern, but in the
reverse order, the largest motoneurons being most
readily silenced by inhibitory inputs. Figure 51 should
be regarded as a simplified illustrative scheme rather
150 B than as an accurate quantitative representation of
experimental findings.

.~ .
As described in section FIRING PATTERNS OF INDI-
VIDUAL MOTONEURONS AND MOTOR UNITS, p. 435, re-
cruitment follows the same order during repetitive

..
ZIOO
firing as it does in monosynaptic reflexes. Since rate of
o fIJ firing is a function of size and other factors, a more
I- .\
.:\:
« general form of Equation 2 is required to represent the
I- output of a pool during repetitive activity
Z
W OT = f10 1 + f20 2 + f30 3 + ... fxO x (3)
I- 50
oa..

-:c
70
#
60
O'-- -'- --L- ---'- ---' I-
o
0.0 1.0 2.0 3.0 4.0
~I •
INPUT RESISTANCE (mQ)
LL
FIG. 49. Relation between input resistance and amplitude of -J
aggregate excitatory postsynaptic potentials (EPSPs) before (A) «
and during (B) posttetanic potentiation in cat. Circles, medial :c
gastrocnemius motoneurons; triangles, semitendinosus motoneu-
rons. [From Luscher, Ruenzel, and Henneman (245).] z
w
the pool and certain correlates of cell size that deter- (!)
z
mine susceptibility to discharge. The pool is organized «
as a graded hierarchy. Rank order in this hierarchy is :c
defined in terms of susceptibility to discharge by phys- ()
iological inputs. The lowest ranking cell requires the
smallest net excitatory input to be fired; the highest
ranking cell requires the largest. Experimental analysis -10 10------20 30 40 50 60 70
has shown that whenever a particular motoneuron ~.
responds monosynaptically, all the lower ranking cells -10
in the pool discharge with it. Thus, all the monosyn- CHANGE IN RISE-TIME (%)
aptic reflex outputs (0) have the form 50. Plot of changes in rise times of excitatory postsynaptic
FIG.
potentials (EPSPs) versus changes in their half-widths following a
(2) tetanus to cat muscle nerve at 500ls for 10 s. Arrows indicate
directions of change. [From Luscher, Ruenzel, and Henneman
where OT is the total monosynaptic output and Ox is (245).]
496 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

where fl-fx represent firing rates detennined by input The small spread of the data points of each pair (8
to the pool and the size and other properties of the pairs are shown) reveals that firing rate variations
motoneurons. In general, the firing rate of a motoneu- among the units of each pair were highly correlated
ron at recruitment and its maximal rate of discharge and maintained an orderly pattern. The dashed
increase with its size. Monster and Chan (261) re- straight lines, indicating the relative rates of change
corded the firing rates of pairs of motor units in the in the firing rates of the units in each pair, had linear
extensor digitorum muscle of humans during volun- regression coefficients of > 0.9 in 80% of the pairs.
tary isometric contractions with values of force up to Nolliinearities were observed in several pairs involving
500 g. The simultaneously recorded values were plot- very low threshold units. Analysis of the slopes of the
ted against one another, as illustrated in Figure 52. dashed lines showed that the higher threshold (i.e.,
larger) unit of a pair generally increased its firing rate
more rapidly than the lower threshold unit, as the
results of animal experiments had also revealed. The
slopes of the regression lines were found to depend on
the threshold differences between members of a pair.
Rate modulation occurs systematically throughout the
pool, enhancing the sensitivity of the gradation proc-
ess.
Basis for Relation Between
Motoneuron Size
FIG. 51. Some features of a motoneuron pool. Heights of vertical and the Force Its Motor Unit Develops
lines represent sizes of motoneurons. Their spacing indicates rela-
tive numbers of cells of different sizes. Horizontal line below base Recruitment of motoneurons results in the activa-
line shows combination of motoneurons fIring in response to a given tion of motor units developing progressively larger
input. Largest cell discharged denoted by X. [From Henneman
(l54a). Organization of the motoneuron pool: the size principle. In: twitch and tetanic tensions. The total force produced
Medical Physiology (l4th ed.), edited by V. B. Mountcastle. St. by the muscle grows in direct proportion to the output
Louis, MO: The C. V. Mosby Co., 1980.] of the motoneuron pool. Two factors account for the

24 A B

I
...
20

16
/
.,.
.
,'(1)

s)'··
® ,."
'J-.
}'

Z
12 ". ~~
.,.0•••
••
-, .~.

...0- .
~
• ,• ••••
z
::J
8
,
• .
u..
0 4
w 24 C D
...'-®
~
<t
a:::
~ @,
®
20 '(

J~
(.:J
r Yo' _..
~
a:::
~.Jts~f

'if.
u.. 16

.,.
12 . 00 7

••
8

~
,I
.. ,r ,
~.
~..
J.
I
J

4 8 12 16 20 24 4 8 12 16 20 24
FIRING RATE OF UNIT I (IN Hz)
FIG. 52. Concomitant changes in mean fIring rates of units i and j (fi and fj ) during slow volitional
changes in an isometric contraction in human. The f;-fj relationship of each pair is approximated by
1 or 2 straight (dashed) lines. The same i unit was used in pairs 5-8. [From Monster and Chan (261).]
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 497

correspondence between neural and contractile out- the relation of the various possible increments of te-
put. 1) The scaling principle, described in MORPHO- tanic tension to the total tension already developed
LOGICAL CONSIDERATIONS, p. 424, implies that larger prior to each new addition must be calculated. To do
motoneurons have larger axons that give off more this, all of the maximal tetanic tensions of MG motor
terminals and innervate more muscle fibers. All of the units measured experimentally were grouped in dec-
muscle fibers of a motor unit in mammals normally ades. Line 1 in Table 8 gives the sums of the unit
participate in its activities. This purely dimensional tensions in each decade. Line 2 contains the total
aspect of the size principle probably explains the linear cumulative tensions, i.e., the sum of all the unit ten-
relationship between the size of a motoneuron and the sions in each column, plus the totals of all columns to
tension capacity of its motor unit in the eat's soleus the left. The percent of tension contributed by new
(Fig. 6). In this "homogeneous" red muscle all of the units to the tensions already developed by all units of
muscle fibers have approximately the same cross-sec- smaller size is shown in line 3. For example, 6.4% in
tional areas. Judging from their uniform histochemical the left-hand column was obtained by dividing 10 g
reactions and fine structure, each fiber presumably (the smallest unit in the next column) by 156.3 g (the
has the same capacity for developing tension. The cumulative total of all units below 10 g). If all incre-
maximal force produced by a motor unit, therefore, ments of tension are viewed as percentages of the
depends on the number of fibers it comprises, although preexisting total tension in this manner, there is no
other factors are not excluded. 2) In heterogeneous loss of fine control as the total tension approaches a
muscles the types of muscle fibers, with their various maximum.
sizes and contractile properties, are related to the sizes This treatment of the data is intended only as an
of their motoneurons. In these muscles force also example of a basic principle in the control of muscle
increases monotonically with neural output. The slope tension. A number of factors combine to make the
of the relationship between these variables, however, actual control much finer and more precise than the
becomes steeper as larger motoneurons are recruited. data in the table indicate. There are about three times
The presence of different types of fibers endows mus- as many motor units in an average MG muscle as
cles with a much wider range of contractile properties, there were in the experimental sample used to con-
making them more versatile servants of the motoneu- struct Table 8. Hence, each addition to the total
ron pool. tension is an even smaller percentage of the total
tensions. With more units the spectrum of sizes is
Actual Sensitivity in Grading more nearly continuous and the thresholds of the
Muscular Tension motoneurons at the low end of the scale are much
It is instructive to consider how much sensitivity closer together than at the high end. Finally and most
over control of muscular tension would be afforded by importantly, units are actually recruited in overlap-
recruitment alone without considering the contribu- ping fashion. New units are added to the total dis-
tion of rate modulation. Henneman and Olson (157) charge well before the last recruits reach their maximal
made such calculations, based on data obtained from tension levels. Rate modulation adds enormously to
the medial gastrocnemius muscle of the cat (Table 8). the fineness of the grading process. The result is a
It is obvious that a large number of small motor units, smooth crescendo of total tension.
varying only slightly in size, could provide extremely
Mathematical Derivation of a "Principle
fine control at the low end of the tension scale. What
of Maximum Grading Sensitivity"
happens to fine control at the other end of the scale
where the largest units recruited may contribute up to Henneman et al. (159) pointed out that the grading
120 g to the total tension? To answer this question, of output in a muscle is somewhat analogous to the

TABLE 8. Percent Contribution of Single Motor Units to Tension Developed

by All Smaller Motor Units of Medial Gastrocnemius
10.0- 20.0- 30.0- 40.0- 50.0- 60.0- 70.0- 80.0- 90.0- 100.0- 110.0-
Grouping of Units <lOg
19.9 29.9 39.9 49.9 59.9 69.9 79.9 89.9 99.9 109.9 120.0

1. Total tensions, g 156 152 387 252 358 323 443 519 587 278 107 345
2. Cumulative totals, g 156 308 695 947 1,305 1,628 2,071 2,590 3,177 3,455 3,562 3,907
3. Percent contribution 6.4 6.5 4.3 4.2 3.8 3.7 3.4 3.1 2.8 2.9 3.1 3.1
of 1 unit to cumu-
lative tensions of
samples
4. Percent contribution 2.1 2.2 1.4 1.4 1.3 1.2 1.1 1.0 0.9 1.0 1.0 1.0
of 1 unit to cumu-
lative tensions, es-
timated for whole
muscle
Adapted from Henneman and Olson (157).
498 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

grading of sensation. The smallest increment that can synchrony throughout the pool. This arrangement
be added to the force exerted by a muscle becomes provides some insight into the organizational principle
greater as the total force increases, just as the smallest underlying the collective action of the pool. The extent
increment of sensory stimulation required to produce to which this model applies to other inputs remains to
a "just noticeable difference" in sensation becomes be seen.
larger as the sensation itself increases in intensity.
Starting with the experimental results of Milner- The Size Principle in Ia and
Brown et al. (258), Hatze (149) has derived mathe- Group II Sensory Fibers
matically a "motor unit size law" that is synonymous
with the size principle of recruitment. From this he The organization of the motoneuron pool depends
has also derived an expression relating the proportion heavily on its input. Recently the size principle has
of active muscle mass to the sequential order number been shown (244) to apply, at least in part, to sensory
of a newly recruited unit. He concludes that his motor fibers that impinge directly on motoneurons. The
unit size law is a consequence of the minimum entropy larger the Ia or group II fiber, the greater are the
principle and "since a minimization of the fluctuations excitatory effects on its impulses on motoneurons.
in the recruitment sequence implies a maximization of Some indication of how the size of afferent fibers
the grading sensitivity of the muscle's force output, it governs the amplitude of synaptic effects presynapti-
appears more appropriate to speak of a principle of cally as well as postsynaptically is emerging. The size
maximum grading sensitivity" (150). of a sensory neuron not only determines the magnitude
of the effect its impulses produce, but also influences
Recruitment Order and Minimum the locations where these effects are exerted postsyn-
Energy Principle aptically.
The size of a primary stretch afferent from muscle
Hatze and Buys (151) have shown mathematically is also correlated with its responsiveness to stretch.
that the sequential order of motor unit recruitment Differences in the dynamic properties of primary and
found experimentally coincides with the order pre- secondary endings have been reported by several in-
dicted when the total energy expenditure of the muscle vestigators (148, 231, 250). Furthermore, Matthews
is minimized at all levels of force output. If the se- (250) showed that the dynamic responses of primary
quence of recruitment is determined by the size prin- and secondary endings were related to the conduction
ciple, as discussed in the previous subsection, but the velocities of the afferent fibers, and that sensitivity to
same sequence can independently be predicted from velocity of stretch increased with fiber size.
a minimum energy hypothesis, then it follows that the The evidence that the synapses from group II and
rate of energy expenditure per unit of muscle mass la afferents are located on different parts of the so-
must be a monotonically increasing function of the matodendritic membrane, in conjunction with a sug-
relative motor unit size (as Fig. 4 in ref. 151 illustrates). gestion of Rall (285), indicates a functional distinction
Studies of heat production in contracting muscle (123, between dendritic and somatic activity. Dendritic syn-
346) tend to support this conclusion, but detailed apses would be more appropriate for the regulation of
analysis of the relative energy cost of tension produc- the general excitatory state of the motoneuron pool.
tion in motor units of different size and type is lacking. They would be well suited to process input from
secondary endings, which are most sensitive to static
Collective Action of Motoneuron changes in muscle length. More proximally located
Pool: Role of Input input would be more apt to cause rapid triggering of
reflex discharge, an appropriate response to input from
One of the properties of a motoneuron pool that primary endings that are most sensitive to rapid
enables it to perform its complex functions is its ca- changes in muscle lengths. Thus, both ends of these
pacity to operate as a collective entity. What organi- sensory fibers follow the size principle and enable the
zational principle enables the individual motoneurons motoneuron pool to respond to the various types of
to work together smoothly as an ensemble? Unques- changes occurring within its muscle.
tionably, the arrangement of the afferent fibers to the
pool provides this cohesion. The distribution of the How Does the Central Nervous System
terminals of single la-fibers to all the homonymous Use the Motoneuron Pool?
motoneurons is an example of how this is accom-
plished in the case of one afferent system. As a result, No definite answer to this question can be given at
the signals in a single la-fiber affect every cell in the present, but it is surely pertinent to ask "what are the
pool at approximately the same time. Although the 60 functional implications of the precise and nearly in-
la-fibers from the eat's MG muscle usually fire asyn- variant rank order that has been demonstrated for the
chronously, the 300 motoneurons in the MG pool higher centers where commands originate?" To con-
receive the same combinations of la inputs from the sider this question in broad perspective, imagine that
MG in approximately the same temporal sequences. no rank order exists. The central nervous system has
Thus, the total la input waxes and wanes in near at its disposal motor units with a wide range of con-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 499

tractile properties. How should it select those it needs considerable neural delay. Consider the number (n) of
for a particular task, and how should it combine them possible combinations that might occur. For a pool
to produce the muscle tension it desires? It is conceiv- consisting of 300 cells, such as the MG pool, n is
able that the central nervous system might use the 300 calculated as follows
motoneurons in the cat's medial gastrocnemius pool 300 300!
like the keys of a large adding machine, picking any n = L
k-l k!
-0---,-------:-
(300 - k)!
combination of cells to yield the proper total. This
notion is implicit in suggestions that motoneurons can In this example n > 10 • 90

be activated in a selective manner. However, this Feedback signals from muscle could not replace
proposition would require 300 separately activable circuits to precalculate the total tension. They would
inputs, so that each of the 300 motoneurons could be be too slow and, in any case, they would be needed to
fired selectively, a very demanding requirement. It determine the size of the feedback correction. The
would also necessitate circuits to do something much solution that has evolved is the rank-ordered pool,
more difficult, namely to calculate what combinations which relieves the central nervous system of the ne-
of active units would produce the correct total ten- cessity for selective activation of motoneurons and
sions. This would require formidable circuitry and provides a simple rule for their combination.

REFERENCES
1. ABRAHAMS, V. C., R. RICHMOND, AND P. K. ROSE. Absence of apses in cat motoneurons. J. Physiol. London 239: 325-345,
monosynaptic reflex in dorsal neck muscles of the cat. Brain 1974.
Res. 92: 130-131, 1975. 16. BARRETT, E. F., AND C. F. STEVENS. Quanta! independence
2. ABZUG, C., M. MAEDA, B. W. PETERSON, AND V. J. WILSON. and uniformity of presynaptic release kinetics at the frog
Branching of individual lateral vestibulospinal axons at differ- neuromuscular junction. J. Physiol. London 227: 665-689,
ent spinal cord levels. Brain Res. 56: 327-330, 1973. 1972.
3. ADRIAN, E. D., AND D. W. BRONK. The discharge of impulses 17. BASMAJIAN, J. V. Control and training of individual motor
in motor nerve fibers. II. The frequency of discharge in reflex units. Science 141: 440, 1963.
and voluntary contractions. J. Physiol. London 67: 119-151, lB. BASMAJIAN, J. V. Control of individual motor units. Am. J.
1929. Phys. Med. 46: 480, 1967.
4. ANDERSON, M. E. Segmental reflex inputs to motoneurons 19. BASMAJIAN, J. V. Electromyography comes of age. Science
innervating dorsal neck musculature in the cat. Exp. Brain 176: 603, 1972.
Res. 28: 175-187, 1977. 20. BESSOU, P., F. EMONET-DENAND, AND Y. LApORTE. Relation
5. ANZENBACHER, H., AND W. ZENKER. Uber die Grossenbezie- entre la vitesse de conduction des fibres nerveuses motrices et
hung der Muskelfasern zu ihren motorischen Endplatten und Ie temps de contraction de leurs unites motrices. C. R. Acad.
Nerven. Z. Zellforsch. Mikrosk. Anat. 60: 860-B71, 1963. Sci. 256: 5625-5627, 1963.
6. ApPELBERG, B., AND F. EMONET-DENAND. Motor units of the 21. BIGLAND, B., AND O. C. J. LIPPOLD. Motor unit activity in the
first superficial lumbrical muscle of the cat. J. Neurophysiol. voluntary contraction of human muscle. J. Physiol. London
30: 154-160, 1967. 125: 322-335, 1954.
7. ApPENTENG, K., M. J. O'DONOVAN, G. SOMJEN, J. A. STE- 22. BLANKENSHIP, J. E., AND M. KUNO. Analysis of spontaneous
PHENS, AND A. TAYLOR. The projection of jaw elevator muscle subthreshold activity in spinal motoneurons of the cat. J.
spindle afferents to fifth nerve motoneurons in the cat. J. Neurophysiol. 31: 195-209, 1968.
Physiol. London 279: 409-423, 1978. 23. BORG, J., L. GRIMBY, AND J. HANNERZ. Axonal conduction
B. ARAKI, T., K. ENDO, Y. KAWAI, K. ITO, AND Y. SHIGENAGA. velocity and voluntary discharge properties of individual short
Supraspinal control of slow and fast spinal motoneurons of the toe extensor motor units in man. J. Physiol. London 277: 143-
cat. In: Progress in Brain Research. Understanding the 152,1978.
Stretch Reflex, edited by S. Homma. Amsterdam: Elsevier, 24. BRACCHI, F., M. DECANDIA, AND T. GUALTIEROTTI. Frequency
1976, vol. 44, p. 413-432. stabilization in the motor centers of spinal cord and caudal
9. BAGINSKY, R. G. Voluntary control of motor unit activity by brain stem. Am. J. Physiol. 210: 1170-1177, 1966.
visual and aural feed-back. Electroencephalogr. Clin. Neuro- 25. BROCK, L. G., J. S. COOMBS, AND J. C. ECCLES. The recording
physiol. 27: 724-725, 1969. of potentials from motoneurons with an intracellular electrode.
10. BAGUST, J. Relationships between motor nerve conduction J. Physiol. London 117: 431-460, 1952.
velocities and motor unit contraction characteristics in a slow 26. BROWN, A. G., AND R. E. W. FYFFE. The morphology of group
twitch muscle of the cat. J. Physiol. London 238: 269-278, Ia afferent fiber collaterals in the spinal cord of the cat. J.
1974. Physiol. London 274: 111-127, 1978.
11. BAGUST, J., S. KNOTT, D. M. LEWIS, J. C. LUCK, AND R. A. 27. BROWN, A. G., AND R. E. W. FYFFE. The morphology of Group
WESTERMAN. Isometric contractions of motor units in a fast Ib muscle afferent fiber collaterals. J. Physiol. London 277:
twitch muscle ofthe cat. J. Physiol. London 231: 87-104,1973. 44-45, 1978.
12. BARRETT, J. N. Motoneuron dendrites: role in synaptic inte- 28. BROWN, M. C., J. K. S. JANSEN, AND D. VAN ESSEN. Polyneu-
gration. Federation Proc. 34: 1398-1407, 1975. ronal innervation of skeletal muscle in newborn rats and its
13. BARRETT, J. N., AND W. E. CRILL. Specific membrane resistiv- elimination during maturation. J. Physiol. London 261: 387-
ity of dye-injected cat motoneuron. Brain Res. 28: 556-561, 422,1976.
1971. 29. BUCHTHAL, F., AND H. SCHMALBRUCH. Contraction times and
14. BARRETT, J. N., AND W. E. CRILL. Specific membrane prop- fibre types in intact human muscle. Acta Physiol. Scand. 79:
erties of cat motoneurons. J. Physiol. London 239: 301-324, 435-452, 1970.
1974. 30. BUDINGEN, H. J., AND H. J. FREUND. The relationship between
15. BARRETT, J. N., AND W. E. CRILL. Influence of dendritic the rate of rise of isometric tension and motor unit recruitment
location and membrane properties on the effectiveness of syn- in a human forearm muscle. Pfluegers Arch. 362: 61-67,1976.
500 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

31. BULLER, A. J., J. C. ECCLES, AND R. M. ECCLES. Interactions rons. In: Integration in the Nervous System, edited by H.
between motoneurons and muscles in respect of the character- Asanuma and V. J. Wilson. Toyko: Igaku-Shoin, 1979, p. 27-
istic speeds of their responses. J. Physiol. London 150: 417- 45.
439,1960. 52. BUSH, B. M. H. Proprioceptive reflexes in the legs of Carcinus
32. BULLOCK, T. H. Comparative aspects of some biological trans- maenas (L.). J. Exp. Bioi. 39: 89, 1962.
ducers. Federation Proc. 12: 666-672, 1953. 53. CANGIANO, E., W. A. COOK, AND O. POMPEIANO. Primary
33. BULLOCK, T. H., AND G. A. HORRIDGE. Structure and Function afferent depolarization in the lumbar cord evoked from the
in the Nervous Systems of Invertebrates. San Francisco: Free- fastigial nucleus. Arch. Ital. Bioi. 107: 321-340, 1969.
man, 1965. 54. CARPENTER, D., I. ENGBERG, AND A. LUNDBERG. Primary
34. BURKE, R. E. Composite nature ofthe monosynaptic excitatory afferent depolarization evoked from the brain stem and the
postsynaptic potential. J. Neurophysiol. 30: 1114-1137, 1967. cerebellum. Arch. Ital. Bioi. 104: 73-85, 1966.
35. BURKE, R. E. Motor unit types of cat triceps surae muscle. J. 55. CHAPPLE, W. D. Motoneurons innervating the dorsal superfi-
Physiol. London 193: 141-160, 1967. cial muscles of the hermit crab, Pagurus pollicarus, and their
36. BURKE, R. E. Group la synaptic input to fast and slow twitch reflex asymmetry. J. Compo Physiol. 121: 413-431, 1977.
motor units of cat triceps surae. J. Physiol. London 196: 605- 56. CHIN, N. K, M. COPE, AND M. PANG. Number and distribution
630,1968. of spindle capsules in seven hindlimb muscles of the cat. In:
37. BURKE, R. E. On the central nervous control of fast and slow Symposium on Muscle Receptors, edited by D. Barker. Hong
twitch motor units. In: New Developments in Electromyogra- Kong: Hong Kong Univ. Press, 1962, p. 241-248.
phy and Clinical Neurophysiology. Human Reflexes, Patho- 57. CLAMANN, H. P. Activity of single motor units during isometric
physiology of Motor Systems, Methodology of Human Re- tension. Neurology 20: 254-260, 1970.
flexes, edited by J. E. Desmedt. Basel: Karger, 1973, vol. 3, p. 58. CLAMANN, H. P., J. D. GILLIES, AND E. HENNEMAN. Effects of
69-94. inhibitory inputs on critical firing level and rank order of
38. BURKE, R. E., L. FEDlNA, AND A. LUNDBERG. Spatial synaptic motoneurons. J. Neurophysiol. 37: 1350-1360, 1974.
distribution of recurrent and group la inhibitory systems in cat 59. CLAMANN, H. P., J. D. GILLIES, R. D. SKINNER, AND E.
spinal motoneurons. J. Physiol. London 214: 305-326, 1971. HENNEMAN. Quantitative measures of output of a motoneuron
39. BURKE, R. E., E. JANKOWSKA, AND G. TEN BRUGGENCATE. A pool during monosynaptic reflexes. J. Neurophysiol. 37: 1328-
comparison of peripheral and rubrospinal synaptic input to 1337, 1974.
slow and fast twitch motor units of triceps surae. J. Physiol. 60. CLAMANN, H. P., AND E. HENNEMAN. Electrical measurement
London 207: 709-732, 1970. of axon diameter and its use in relating motoneuron size to
40. BURKE, R. E., D. N. LEVINE, M. SALCMAN, AND P. TSAIRIS. critical firing level. J. Neurophysiol. 39: 844-851, 1976.
Motor units in cat soleus muscle: physiological, histochemical 61. CLAMANN, H. P., AND C. G. KUKULKA. Evidence for selective
and morphological characteristics. J. Physiol. London 238: inputs to a motoneuron pool. Neurosci. Abst. 2: 539, 1976.
503-514, 1974. 62. CLAMANN, H. P., AND C. G. KUKULKA. The relation between
41. BURKE, R. E., D. N. L~~VINE, P. TSAIRIS, AND F. E. ZAJAC III. size of motoneurons and their position in the cat spinal cord.
Physiological types and histochemical profJJ.es in motor units J. Morpho I. 153: 461-466, 1977.
of the cat gastrocnemius. J. Physiol. London 234: 723-748, 63. CLAMANN, H. P., AND C. G. KUKULKA. Reversals of recruit-
1973. ment order in medial gastrocnemius produced by stimulation
42. BURKE, R. E., AND P. G. NELSON. Synaptic activity in moto- of Deiters' nucleus. Neurosci. Abst. 3: 269, 1977.
neurons during natural stimulation of muscle spindles. Science 64. CLARK, D. A. Muscle counts of motor units: a study in inner-
151: 1088-1091, 1966. vation ratios. Am. J. Physiol. 96: 296-304, 1931.
43. BURKE, R. E., AND P. G. NELSON. Accommodation to current 65. CLOUGH, J. F. M., D. KERNELL AND C. G. PHILLIPS. The
ramps in motoneurons of fast and slow twitch motor units. Int. distribution of monosynaptic excitation from the pyramidal
J. Neurosci. 1: 347-356, 1971. tract and from primary spindle afferents to motoneurons of the
44. BURKE, R. E., AND P. HUDOMIN. Spinal neurons and synapses. baboon's hand and forearm. J. Physiol. London 198: 145-166,
In: Handbook of Physiology. The Nervous System. Cellular 1968.
Biology of Neurons, edited by E. Kandel. Bethesda, MD: Am. 66. COERS, C. Les variations structurelles normales et patholo-
Physiol. Soc., 1977, sect. 1, vol. I, pt. 2, chapt. 24, p. 877-944. giques de la jonction neuromusculaire. Acta Neurol. Psychiatr.
45. BURKE, R. E., W. Z. HYMER, AND J. V. WALSH, JR. Functional Belg. 55: 741-866, 1955.
specialization in the motor unit population of cat medial gas- 67. COHEN, L. A. Localization of stretch reflex. J. Neurophysiol.
trocnemius muscle. In: Control of Posture and Locomotion, 16: 272-285, 1953.
edited by R. B. Stein, K B. Pearson, R. S. Smith, and J. B. 68. COHEN, L. A. Organization of stretch reflex into two types of
Hedford. New York: Plenum, 1974. (Adv. Behav. BioI. Ser., direct spinal arcs. J. Neurophysiol. 17: 443-453, 1954.
vol. 7.) 69. COLE, W. V. Structural variations of nerve endings in the
46. BURKE, R. E., W. Z. RYMER, AND J. V. WALSH, JR. Relative striated muscle of rat. J. Compo Neurol. 108: 445-464, 1957.
strength of synaptic input from short-latency pathways to 70. COLLATOS, T. C., A. NIECHAJ, S. J. NELSON, AND L. M.
motor units of defined type in cat medial gastrocnemius. J. MENDELL. Fluctuations in time of onset of la-motoneuron
Neurophysiol. 39: 447-458, 1976. EPSPs in the cat. Brain Res. 160: 514-518, 1979.
47. BURKE, R. E., P. L. STRICK, K KANDA, C. C. KIM, AND B. 71. CONRADI, S. Ultrastructure and distribution of neuronal and
WALMSLEY. Anatomy of medial gastrocnemius and soleus mo- glial elements on the motoneuron surface in the lumbosacral
tor nuclei in cat spinal cord. J. Neurophysiol. 40: 667-680, 1977. spinal cord of the adult cat. Acta Physiol. Scand. Suppl. 332:
48. BURKE, R. E., AND G. TEN BRUGGENCATE. Electrotonic char- 5-48, 1969.
acteristics of alpha motoneurons of varying size. J. Physiol. 72. CONRADI, S. Ultrastructure of dorsal root boutons on lumbo-
London 212: 1-20, 1971. sacral motoneurons of the adult cat. Acta Physiol. Scand.
49. BURKE, R. E., AND P. TSAIRIS. Anatomy and innervation ratios Suppl. 332: 85-115, 1969.
in motor units of cat gastrocnemius. J. Physiol. London 234: 73. CONRADI, S. Functional anatomy of the anterior horn motor
749-765, 1973. neuron. In: The Peripheral Nerve, edited by D. H. Landon.
50. BURKE, R. E., AND P. TSAIRIS. The correlation of physiological London: Chapman Hall, 1976, p. 279-329.
properties with histochemical characteristics in single muscle 74. CONRADI, S., AND L. O. RONNEVI. Spontaneous elimination of
units. Ann. NY Acad. Sci. 228: 145-159, 1974. synapses on cat spinal motoneurons after birth: do half of the
51. BURKE, R. E., B. WALMSLEY, AND J. A. HODGSON. Structural- synapses on the cell bodies disappear? Brain Res. 92: 505-510,
functional relations in monosynaptic action on spinal motoneu- 1975.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 501

75. COOMBS, J. S., D. R CURTIS, AND J. C. ECCLES. The generation 97. ECCLES, J. C., R M. ECCLES, A. IGGO, AND M. ITO. Distribution
of impulses in motoneurons. J. Physiol. London 139: 232-249, of recurrent inhibition among motoneurons. J. Physiol. Lon-
1957. don 159: 479-499, 1961.
76. CORDA, M., G. EKLUND, AND C. V. EULER. External intercostal 98. ECCLES, J. C., R M. ECCLES, A. IGGO, AND A. LUNDBERG.
and phrenic alpha motor responses to changes in respiratory Electrophysiological studies on gamma motoneurons. Acta
load. Acta Physiol. Scand. 63: 391-400, 1965. Physiol. Scand. 50: 32-40, 1960.
77. COWAN, W. M. The development of the brain. Sci. Am. 241: 99. ECCLES, J. C., R M. ECCLES, AND A. LUNDBERG. The con-
113-133, 1979. vergence of monosynaptic excitatory afferents onto many dif-
78. CREED, R S., D. DENNy-BROWN, J. C. ECCLES, E. G. T. ferent species of alpha motoneurons. J. Physiol. London 137:
LIDDELL, AND C. S. SHERRINGTON. Reflex Activity of the 22-50, 1957.
Spinal Cord. London: Oxford Univ. Press, 1932. 100. ECCLES, J. C., R M. ECCLES, AND A. LUNDBERG. Synaptic
79. CREED, R S., AND C. SHERRINGTON. Observations on concur- actions on motoneurons caused by impulses in Golgi tendon
rent contraction of flexor muscles in the flexion reflex. Proc. R. organ afferents. J. Physiol. London 138: 227-252, 1957.
Soc. London Ser. B 100: 258-267, 1926. 101. ECCLES, J. C., R M. ECCLES, AND F. MAGNI. Central inhibitory
80. CULLHEIM, S. Relations between cell body size, axon diameter action attributable to presynaptic depolarization produced by
and axon conduction velocity of cat sciatic alpha-motoneurons muscle afferent volleys. J. Physiol. London 159: 147-166, 1961.
stained with horseradish peroxidase. Neurosci. Letters 8: 17- 102. ECCLES, J. C., AND C. S. SHERRINGTON. Numbers and contrac-
20,1978. tion values of individual motor-units examined in some muscles
81. CULLHEIM, S., AND J. O. KELLERTH. A morphological study of of the limb. Proc. R. Soc. Lond. Ser. B 106: 326, 1939.
the axons and recurrent axon collaterals of cat alpha-motoneu- 103. ECCLES, R M., AND A. LUNDBERG. Integrative pattern of Ia
rons supplying different functional types of muscle unit. J. synaptic actions of motoneurons of hip and knee muscles. J.
Physiol. London 281: 301-313, 1978. Physiol. London 144: 271-298, 1958.
82. CULLHEIM, S., AND J. O. KELLERTH. A morphological study of 104. ECCLES, R M., AND A. LUNDBERG. Synaptic actions in moto-
the axons and recurrent axon collaterals of cat sciatic a-moto- neurons by afferents which may evoke the flexion reflex. Arch.
neurons after intracellular staining with horseradish peroxi- Ital. Biol. 97: 199-221, 1959.
dase. J. Compo Neurol. 178: 537-558, 1978. 105. EDWARDS, F. R, S. J. REDMAN, AND B. WALMSLEY. Statistical
83. CULLHEIM, S., J. O. KELLERTH, AND S. CONRADI. Evidence for fluctuations in charge transfer at Ia synapses on spinal moto-
direct synaptic interconnections between cat spinal a-moto- neurons. J. Physiol. London 259: 665-688, 1976.
neurons via the recurrent axon collaterals: a morphological 106. EDWARDS, F. R, S. J. REDMAN, AND B. WALMSLEY. Non-
study using intracellular injection of horseradish peroxidase. quantal fluctuations and transmission failures in charge trans-
Brain Res. 132: 1-10, 1977. fer at Ia synapses on spinal motoneurons. J. Physiol. London
84. CZEH, G., R GALLEGO, N. KUDO, AND M. KUNO. Evidence for 259: 689-704, 1976.
the maintenance of motoneuron properties by muscle activity. 107. EDWARDS, F. R, S. J. REDMAN, AND B. WALMSLEY. The effect
J. Physiol. London 281: 239-252, 1978. of polarizing currents on unitary Ia excitatory postsynaptic
85. DAVIS, W. J. Functional significance of motoneuron size and potentials evoked in spinal motoneurons. J. Physiol. London
soma position in swimmeret system of the lobster. J. Neuro- 259: 705-723, 1976.
physiol. 34: 274-288, 1971. 108. EMONET-DENAND, F., L. JAMI, AND Y. LAPORTE. Skeleto-fu-
86. DEITERS, O. Untersuchung iiber Gehirn and Riickenmark des simotor axons in hind-limb muscles of the cat. J. Physiol.
Menschen und der Siiugethiere. Braunschweig, 1865. London 249: 153-166, 1975.
87. DEKKER, J. J., D. G. LAWRENCE, AND H. G. J. M. KUYPERS. 109. ERLANGER, J., AND H. S. GASSER. Electrical Signs of Nervous
The location of longitudinally running dendrites in the ventral Activity. Philadelphia: Univ. Pennsylvania Press, 1937.
horn of the cat spinal cord. Brain Res. 51: 319-325, 1973. 110. FADEN, J. S., AND F. F. ZAJAC. Direct comparison of recruit-
88. DEL CASTILLO, J., AND B. KATZ. Quantal components of the ment order with neural and muscular properties of motor units.
end plate potential. J. Physiol. London 124: 560-573, 1954. Soc. Neurosci. Abstr. 3: 271, 1977.
89. DENNy-BROWN, D. Nature of postural reflexes, Proc. R. Soc. Ill. FATT, P. Sequence of events in synaptic activation of a moto-
London Ser. B 104: 252-301, 1929. neurone. J. Neurophysiol. 20: 61-80, 1957.
90. DESMEDT, J. E., AND E. GODAUX. Ballistic contractions in man: 112. FATT, P., AND B. KATZ. An analysis of the end-plate potential
characteristic recruitment pattern of single motor units of the recorded with an intracellular electrode. J. Physiol. London
tibialis anterior muscle. J. Physiol. London 264: 673-693, 115: 320-370, 1951.
1977. 113. FEDDE, M. R, P. D. DE WET, AND R L. KITCHELL. Motor unit
91. DESMEDT, J. E., AND E. GODAUX. Critical evaluation of the recruitment pattern and tonic activity in respiratory muscles
size principle of human motoneuron recruitment in ballistic of Gallus domesticus. J. Neurophysiol. 32: 995-1004, 1969.
movements and in vibration-induced inhibition or potentiation. 114. FRANK, K., AND M. G. F. FUORTES. Accommodation of spinal
Trans. Am. Neurol. Assoc. 102: 104-108,1977. motoneurons of cats. Arch. Ital. Biol. 98: 165-170, 1960.
92. DESMEDT, J. E., AND E. GODAUX. Fast motor units are not 115. FREUND, H.-J., H. J. BUDINGEN, AND V. DIETZ. Activity of
preferentially activated in rapid voluntary contractions in man. single motor units from human forearm muscles during vol-
Nature London 267: 717-719, 1977. untary isometric contractions. J. Neurophysiol. 38: 933-946,
93. DESMEDT, J. E., AND E. GODAUX. Ballistic skilled movements: 1975.
load compensation and patterning of the motor commands. In: 116. Fu, T. C., H. HULTBORN, R LARSSON, AND A. LUNDBERG.
Progress in Clinical Neurophysiology. Cerebral Motor Con- Reciprocal inhibition during the tonic stretch reflex in the
trol in Man: Long Loop Mechanisms, edited by J. E. Desmedt. decerebrate cat. J. Physiol. London 284: 345-369, 1978.
Basel: Karger, 1978, vol. 4, p. 21-55. 117. Fu, T. C., AND E. D. SCHOMBURG. Electrophysiological inves-
94. DESMEDT, J. E., AND E. GODAUX. Ballistic contractions in fast tigation of the projection of secondary muscle spindle afferents
or slow human muscles: discharge patterns of single motor in the cat spinal cord. Acta Physiol. Scand. 91: 314-329, 1974.
units. J. Physiol. London 285: 185-196, 1978. 118. Fuortes, M. G. F., K. Frank, and M. C. Becker. Steps in the
95. DESMEDT, J. E., AND E. GODAUX. Mechanism of the vibration production of motoneuron spikes. J. Gen. Physiol. 40: 732-752,
paradox: excitatory and inhibitory effects of tendon vibration 1957.
on single soleus muscle motor units in man. J. Physiol. London 119. GARNETT, R, AND J. A. STEPHENS. Changes in the recruitment
285: 197-207, 1978. threshold of motor units in human first dorsal interosseous
96. ECCLES, J. C. The Physiology of Synapses. New York: Aca- muscle produced by skin stimulation. J. Physiol. London 282:
demic, 1964. 13-14, 1978.
502 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

120. GASSER, H. The classification of nerve fibers. Ohio J. Sci. 41: motor column of the chick embryo. J. Compo Neurol. 160: 535-
145, 1941. 546,1975.
121. GELFAN, S. Neuron and synapse populations in the spinal cord: 144. HARRIS, C. The morphology of the myoneural junction as
indication ofrole in total integration. Nature London 198: 162, influenced by neurotoxic drugs. Am. J. Pathol. 30: 501-519,
1963. 1954.
122. GELFAN, S., AND A. G. RAPISARDA. Synaptic density on spinal 145. HARRIS, D. A., AND K HENNEMAN. Identification of two spe-
neurons of normal dogs and dogs with experimental hindlimb cies of alpha motoneurons in cat's plantaris pool. J. Neuro-
rigidity. J. Compo Neurol. 123: 73-95, 1964. physiol. 40: 16-25, 1977.
123. GIBBS, C. L., AND W. R GIBSON. Energy production of rat 146. HARRIS, D. A., AND K HENNEMAN. Different species of alpha
soleus muscle. Am. J. Physiol. 223: 864-871, 1972. motoneurons in same pool: further evidence from effects of
124. GOGAN, P., J. P. GEURITAUD, G. HORCHELLE-BoSSAVIT, AND inhibition on their firing rates. J. Neurophysiol. 42: 927-935,
S. TYc-DuMONT. Direct excitatory interactions between spinal 1979.
motoneurons ofthe cat. J. Physiol. London 272: 755-767, 1977. 147. HARRISON, V. F., AND O. A. MORTENSON. Identification and
125. GOLDBERG, L. J., AND B. DERFLER. Relationship among re- voluntary control of single motor unit activity in the tibialis
cruitment order, spike amplitude, and twitch tension of single anterior muscle. Anat. Rec. 144: 109, 1962.
motor units in human masseter muscle. J. Neurophysiol. 40: 148. HA~VEY, R J., AND P. B. C. MATTHEWS. The response of
879-890, 1977. d~fferentedmuscle spindle endings in the cat's soleus to slow
126. GOLDSPINK, D. F. The influence of immobilization and stretch l~~ension of the muscle. J. Physiol. London 157: 370-392,
on protein turnover of rat skeletal muscle. J. Physiol. London /1961.
264: 267-282, 1977. 149. HATZE, H. The relative contribution of motor unit recruitment
127. GOLDSPINK, D. F. The influence of activity on muscle size and and rate coding to the production of static isometric muscle
protein turnover. J. Physiol. London 264: 283-296, 1977. .force. Biol. Cybernetics 27: 21-25, 1977.
128. GRANIT, R Neuromuscular interaction in postural tone of the 150.HATZE, H. A teleological explanation of Weber's Law and the
eat's isometric soleus muscle. J. Physiol. London 143: 387-402, motor unit size law. Bull. Math. Biol. 41: 407-425, 1979.
1958. 151. HATZE, H., AND J. D. Buys. Energy-optimal controls in the
129. GRANIT, R, H. D. HENATSCH, AND G. STEG. Tonic and phasic mammalian neuromuscular system. Biol. Cybernetics 27: 9-20,
ventral hom cells differentiated by post-tetanic potentiation in 1977.
cat extensors. Acta Physiol. Scand. 37: 114-126, 1956. 152. HENATSCH, H. D., AND F. J. SCHULTE. Reflexerregung und
130. GRANIT, R, D. KERNELL, AND G. K. SHORTESS. Quantitative Eigenhemmung tonischer und phossicher Alpha-motoneurone
aspects of repetitive firing of mammalian motoneurons, caused wahrend chemischer Dauererregung der Muskelspindeln.
by injected currents. J. Physiol. London 168: 911-931, 1963. Pfluegers Arch. 268: 134-147, 1958.
131. GRANIT, R, J. E. PASCOE, AND G. STEG. The behaviour of 153. HENNEMAN, K Relation between size of neurons and their
tonic alpha and gamma motoneurons during stimulation of susceptibility to discharge. Science 126: 1345-1347, 1957.
recurrent collaterals. J. Physiol. London 138: 381-400, 1957. 154. HENNEMAN, K Skeletal muscle: the servant of the nervous
132. GRILLNER, S. Muscle stiffness and motor control-forces in system. In: Medical Physiology (14th ed.), edited by V. B.
the ankle during locomotion and standing. In: Motor Control, Mountcastle. St. Louis: Mosby, 1980, p. 674-702.
edited by A. A. Gydikov, N. T. Tankov, and D. S. Kosarov. 154a.HENNEMAN, E. Organization of the motoneuron pool: the size
New York: Plenum, 1973, p. 195-215. (Proc. Int. Symp. Motor principle. In: Medical Physiology (14th ed.), edited by V. B.
Control "Zlatni Pyassatsi," 2nd, near Varni, Bulgaria, 1973.) Mountcastle. St. Louis: Mosby, 1980, p. 718-741.
133. GRILLNER, S., T. HONGO, AND S. LUND. The vestibulospinal 155. HENNEMAN, K, H. P. CLAMANN, J. D. GILLIES, AND R D.
tract. Effects on alpha-motoneurons in the lumbosacral spinal SKINNER. Rank-order of motoneurons within a pool: law of
cord in the cat. Exp. Brain Res. 10: 94-120, 1970. combination. J. Neurophysiol. 37: 1338, 1974.
134. GRILLNER, S., T. HONGO, AND S. LUND. Convergent effects on 156. HENNEMAN, K, H.-R LUSCHER, AND P. RUENZEL. Terminal
alpha motoneurons from the vestibulospinal tract and a path- distribution of Ia and group II spindle afferent fibers in the
way descending in the medial longitudinal fasciculus. Exp. spinal cord as revealed by postsynaptic population potentials
Brain Res. 12: 457-479, 1971. recorded from two adjacent spinal segments. Neurosci. Abstr.
135. GRILLNER, S., AND M. UDO. Recruitment in the tonic stretch 5: 724, 1979.
reflex. Acta Physiol. Scand. 81: 571-573, 1971. 157. HENNEMAN, E., AND C. B. OLSON. Relations between structure
136. GRIMBY, L., AND J. HANNERZ. Recruitment order of motor and function in the design of skeletal muscles. J. Neurophysiol.
units in voluntary contraction: changes induced by propriocep- 28: 581-598, 1965.
tive afferent activity. J. Neurol. Neurosurg. Psychiatry 31: 158. HENNEMAN, K, B. T. SHAHANI, AND R R YOUNG. Voluntary
565-573, 1968. control of human motor units. In: The Motor System: Neuro-
137. GRIMBY, L., AND J. HANNERZ. Differences in recruitment order physiology and Muscle Mechanisms, edited by M. Shahani.
and discharge pattern of motor units in the early and late Amsterdam: Elsevier, 1976, p. 73-78.
flexion reflex components in man. Acta Physiol. Scand. 90: 159. HENNEMAN, K, G. SOMJEN, AND D. O. CARPENTER. Functional
555-564, 1974. significance of cell size in spinal motoneurons. J. Neurophysiol.
138. GRIMBY, L., AND J. HANNERZ. Disturbances in voluntary re- 28: 560-580, 1965.
cruitment order of low and high frequency motor units on 160. HENNEMAN, K, G. SOMJEN, AND D. O. CARPENTER. Excitabil-
blockade of proprioceptive afferent activity. Acta Physiol. ity and inhibitibility of motoneurons of different sizes. J. Neu-
Scand. 96: 207-216, 1976. rophysiol. 28: 599-620, 1965.
139. GRIMBY, L., AND J. HANNERZ. Firing rate and recruitment 161. HODES, R, S. M. PEACOCK, JR., AND D. BODIAN. Selective
order of toe extensor motor units in different modes of volun- destruction of large motoneurons by poliomyelitis virus. II.
tary contraction. J. Physiol. London 264: 865-879, 1977. Size of motoneurons in the spinal cord of rhesus monkeys. J.
140. GRINELL, A. D. A study of the interaction between motoneu- Neuropathol. Exp. Neurol. 8: 400-410, 1949.
rons in the frog spinal cord. J. Physiol. London 182: 612-648, 162. HOLLYDAY, M., AND V. HAMBURGER. Reduction of the natu-
1966. rally occurring motor neuron loss by enlargement of the pe-
141. GROSSMAN, Y., M. K SPRIA, AND I. PARNAS. Differential flow riphery. J. Compo Neurol. 170: 311-320, 1976.
of information into branches of a single axon. Brain Res. 64: 163. HONGO, T., N. ISHIZUKA, H. MANNEN, AND S. SASAKI. Axonal
379-386, 1973. trajectory of single group Ia and Ib fibers in the cat spinal cord.
142. HAMBURGER, V. Regression vs. peripheral control of differen- Neurosci. Letters 8: 321-328, 1978.
tiation in motor hypoplasia. Am. J. Anat. 102: 365-410, 1958. 164. HONGO, T., E. JANKOWSKA, AND A. LUNDBERG. The rubro-
143. HAMBURGER, V. Cell death in the development of the lateral spinal tract. I. Effects on alpha motoneurons innervating hind-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 503

limb muscles in cats. Exp. Brain Res. 7: 344-364, 1969. 1971.

165. HORRIDGE, G. A., AND M. BURROWS. Tonic and phasic systems 187. JANKOWSKA, E., Y. PADEL, AND R. TANAKA. Projections of
in parallel in the eyecup responses of the crab Carcinus. J. pyramidal tract cells to alpha motoneurons innervating hind
Exp. Biol. 49: 269, 1968. limb muscles in the monkey. J. Physiol. London 249: 637-667,
166. HouK, J., AND E. HENNEMAN. Responses of Golgi tendon 1975.
organs to active contractions of the soleus muscle of the cat. J. 188. JANKOWSKA, E., AND W. J. ROBERTS. An electrophysiological
Neurophysiol. 30: 466-481, 1967. demonstration of the axonal projections of single spinal inter-
167. HOWLAND, B., J. Y. LETTVIN, W. S. MCCULLOCH, W. PITTS, neurons in the cat. J. Physiol. London 222: 597-622, 1972.
AND P. D. WALL. Reflex inhibition by dorsal root interaction. 189. JANKOWSKA, E., AND W. J. ROBERTS. Synaptic actions of single
J. Neurophysiol. 18: 1-17, 1955. interneurons 'mediating reciprocal Ia inhibition of motoneu-
168. HOYLE, G. Exploration of neuronal mechanisms underlying rons. J. Physiol. London 222: 623-642, 1972.
behaviour in insects. In: Neural Theory and Modeling: Pro- 190. JANKOWSKA, E., AND D. O. SMITH. Antidromic activation of
ceedings, edited by R. F. Reiss. Stanford, CA: Stanford Univ. Renshaw cells and their axonal projections. Acta Physiol.
Press, 1964, p. 346-376. (Ojai Symp., 1962.) Scand. 88: 198-214, 1973.
169. HUGHES, G. M., AND C. M. BALLINTIJN. Electromyography of 191. KAISER, E. I., AND I. PETERSEN. Muscle action potentials
the respiratory muscles and gill water flow in the dragonet. J. studied by frequency analysis and duration measurements.
Exp. Biol. 49: 583, 1968. Acta Neurol. Scand. 41: (Suppl. 13) 213-236, 1965.
170. HUIZAR, P., N. KUDO, M. KUNO, AND Y. MIYATA. Reaction of 192. KANDA, K, R. E. BURKE, AND B. WALMSLEY. Differential
intact spinal motoneurons to partial denervation of the muscle. control of fast and slow twitch motor units in the decerebrate
J. Physiol. London 265: 175-191, 1977. cat. Exp. Brain Res. 29: 57-74,1977.
171. HULTBORN, H. Transmission in the pathway of reciprocal Ia 193. KATO, M., AND K FUKUSHIMA. Effect of differential blocking
inhibition to motoneurons and its control during the tonic of motor axons on antidromic activation of Renshaw cells. Exp.
stretch reflex. In: Progress in Brain Research. Understanding Brain Res. 20: 135-143, 1974.
the Stretch Reflex, edited by S. Homma. Amsterdam: Elsevier, 194. KATO, M., AND J. TANJI. Volitionally controlled single motor
1976, vol. 44, p. 235-255. units in human finger muscles. Brain Res. 40: 345-357, 1972.
172. HULTBORN, H., E. JANKOWSKA, AND S. LINDSTROM. Recurrent 195. KATZ, B. The transmission of impulses from nerve to muscle,
inhibition from motor axon collaterals of transmission in the and the subcellular unit of synaptic action. Proc. R. Soc.
Ia inhibitory pathway to motoneurons. J. Physiol. London London Ser. B 155: 455-477,1962.
215: 591-612, 1971. 196. KATZ, B., AND R MILEDI. A study of spontaneous miniature
173. HUNT, C. C. Monosynaptic reflex response of spinal motoneu- potentials in spinal motoneurons. J. Physiol. London 168: 389-
rons to graded afferent stimulation. J. Gen. Physiol. 38: 813- 422,1963.
832,1955. 197. KATZ, B., AND R MILEDI. The measurement of synaptic delay
174. HURSH, J. B. Conduction velocity and diameter of nerve fibers. and the time course of acetylcholine released at the neuromus-
Am. J. Physiol. 127: 131-139, 1939. cular junction. Proc. R. Soc. London Ser. B 161: 483-495, 1965.
175. IANSEK, R, AND S. J. REDMAN. An analysis of the cable 198. KATZ, B., AND R MILEDI. A study of synaptic transmission in
properties of spinal motoneurons using a brief intracellular the absence of nerve impulses. J. Physiol. London 192: 407-
current pulse. J. Physiol. London 234: 613-636, 1973. 436,1967.
176. IANSEK, R, AND S. J. REDMAN. The amplitude, time course 199. KATZ, B., AND S. THESLEFF. On the factors which determine
and charge of unitary excitatory post-synaptic potentials the amplitude of the miniature end plate potential. J. Physiol.
evoked in spinal motoneuron dendrites. J. Physiol. London London 137: 267-278, 1957.
234: 665-688, 1973. 200. KERN ELL, D. The adaptation and the relation between dis-
177. ILES, J. F. Central terminations of muscle afferents on moto- charge frequency and current strength of cat lumbosacral
neurons in the cat spinal cord. J. Physiol. London 262: 91-117, motoneurons stimulated by long-lasting injected currents. Acta
1976. Physiol. Scand. 65: 65-73, 1965.
178. ILES, J. F. The speed of passive dendritic conduction of syn- 201. KERNELL, D. High-frequency repetitive firing of cat lumbosac-
aptic potentials in a model motoneuron. Proc. R. Soc. London ral motoneurons stimulated by long-lasting injected currents.
Ser. B 197: 225-229, 1977. Acta Physiol. Scand. 65: 74-86, 1965.
179. ILLIS, L. Spinal cord synapses in the cat: the normal appear- 202. KERNELL, D. The limits of firing frequency in cat lumbosacral
ances by the light microscope. Brain 87: 543-554, 1964. motoneurons possessing different time course of afterhyper-
180. ILLIS, L. The relative densities of monosynaptic pathways to polarization. Acta Physiol. Scand. 65: 87-100, 1965.
the cell bodies and dendrites of the cat ventral horn. J. Neurol. 203. KERN ELL, D. Input resistance, electrical excitability, and size
Sci. 4: 259-270, 1967. of ventral horn cells in cat spinal cord. Science 152: 1637-1640,
181. JACK, J. J. B., S. MILLER, R. PORTER, AND S. J. REDMAN. The 1966.
distribution of group Ia synapses on lumbrosacral spinal mo- 204. KERNELL, D. Recruitment, rate modulation and the tonic
toneurons in the cat. In: Excitatory Synaptic Mechanisms, strength reflex. Progress in Brain Research. Understanding
edited by P. Anderson and J. K S. Jansen. Oslo: Universitets- the Stretch Reflex, edited by S. Homma. Amsterdam: Elsevier,
forlaget, 1970, p. 199-205. 1976, vol. 44, p. 257-266.
182. JACK, J. J. B., S. MILLER, R PORTER, AND S. J. REDMAN. The 205. KERN ELL, D., AND H. SJOHOLM. Repetitive impulse firing:
time course of minimal excitatory post-synaptic potentials comparisons between neuron models based on "voltage clamp
evoked in spinal motoneuron by group Ia afferent fibers. J. equations" and spinal motoneurons. Acta Physiol. Scand. 87:
Physiol. London 215: 353-380, 1971. 40-56, 1973.
183. JACK, J. J. B., D. NOBLE, AND R W. TSIEN. Electric Current 206. KERNELL, D., AND H. SJOHOLM. Recruitment and firing rate
Flow in Excitable Cells. Oxford: Clarendon, 1975, p. 502. modulation of motor unit tension in a small muscle of the cat's
184. JACK, J. B., AND R PORTER. Rapidly and slowly rising com- foot. Brain Res. 98: 57-72,1975.
ponents of monosynaptic excitatory postsynaptic potentials in 207. KIRKWOOD, P. A., AND T. A. SEARS. Monosynaptic excitation
spinal motoneurons. J. Physiol. London 186: 106-107, 1966. of motoneurons from secondary endings of muscle spindles.
185. JAMI, L., AND J. PETIT. Correlation between axonal conduction Nature London 252: 243-244,1974.
velocity and tetanic tension of motor units in four muscles of 208. KIRKWOOD, P. A., AND T. A. SEARS. Monosynaptic excitation
the cat hind limb. Brain Res. 96: 114-118, 1975. of motoneurons from muscle spindle secondary endings of
186. JANKOWSKA, E., AND S. LINDSTROM. Morphological identifi- intercostal and triceps surae muscles in the cat. J. Physiol.
cation of Renshaw cells. Acta Physiol. Scand. 81: 428-430, London 245: 64-66, 1975.
 504 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

209. KIRKWOOD, P. A., AND T. A. SEARS. The synaptic connexions 234. LILEY, A. W. The effects of presynaptic polarization on the
to intercostal motoneurons as revealed by the average common spontaneous activity of the mammalian neuromuscular junc-
excitation potential. J. Physiol. London 275: 103-134, 1978. tion. J. Physiol. London 134: 427-443, 1956.
210. KNOTT, S., D. M. LEROIS, AND J. C. LUCK. Motor unit area in 235. LLOYD, D. P. C. Reflex action in relation to pattern and
a cat limb muscle. Exp. Neurol. 30: 475-483, 1971. peripheral source of afferent stimulation. J. Neurophysiol. 6:
211. KNOX, C. K. Cross-correlation functions for a neuronal model. 111-119, 1943.
Biophys. J. 14: 567-582, 1974. 236. LLOYD, D. P. C. Integrative pattern of excitation and inhibition
212. KOZIOL, J. A., AND H. C. TUCKWELL. Analysis and estimation in two-neuron reflex arcs. J. Neurophysiol. 9: 439-444, 1946.
of synaptic densities and their spatial variation on the moto- 237. LLOYD, D. P. C. Post-tetanic potentiation of response in mono-
neuron surface. Brain Res. 150: 617-624, 1978. synaptic reflex pathways of the spinal cord. J. Gen. Physiol.
213. KRNJEVIC, K., AND R. J. MILEDI. Failure of neuromuscular 33: 147-170, 1949.
propagation in rats. J. Physiol. London 140: 440-461, 1958. 238. LOEB, G. Decreased conduction velocity in the proximal pro-
214. KUNO, M. Excitability following antidromic activation in spinal jections of myelinated dorsal root ganglion cells in the cat.
motoneurons supplying red muscles. J. Physiol. London 149: Brain Res. 103: 381-385, 1976.
374-393, 1959. 239. LORENTE DE No, R. Limits of variation of the synaptic delay
215. KUNO, M. Quantal components of excitatory synaptic poten- of motoneurons. J. Neurophysiol. 1: 187-194, 1938.
tials in spinal motoneurons. J. Physiol. London 175: 81-99, 240. LUCAS, M. E., AND W. D. WILLIS. Identification of muscle
1964. afferents which activate interneurons in the intermediate nu-
216. KUNO, M. Mechanism of facilitation and depression of the cleus. J. Neurophysiol. 37: 282-293, 1974.
excitatory synaptic potential in spinal motoneurons. J. Physiol. 241. LUND, S., AND O. POMPEIANO. Monosynaptic excitation of
London 175: 100-112, 1964. alpha motoneurons from supraspinal structures in the cat.
217. KUNO, M. Quantum aspects of central and ganglionic synaptic Acta Physiol. Scand. 73: 1-21, 1968.
transmission in vertebrates. Physiol. Rev. 51: 647-678, 1971. 242. LUNDBERG, A., K. MALMGREN, AND E. D. SCHOMBURG. Role
218. KUNO, M., AND R. LUNAS. Enhancement of synaptic trans- of joint afferents in motor control exemplified by effects on
mission by dendritic potentials in chromatolyzed motoneurons reflex pathways from Ib afferents. J. Physiol. London 284:
of the cat. J. Physiol. London 210: 807-821, 1970. 327-343, 1978.
219. KUNO, M., AND J. T. MIYAHARA. Non-linear summation of 243. LUNDBERG, A., U. NORSELL, AND P. VOORHOEVE. Pyramidal
unit synaptic potentials in spinal motoneurons of the cat. J. effects on lumbo-sacral interneurons activated by somatic af-
Physiol. London 201: 465-477, 1969. ferents. Acta Physiol. Scand. 56: 220-229, 1962.
220. KUNO, M., AND J. T. MIYAHARA. Analysis of synaptic efficacy 244. LUSCHER, H.-R., P. RUENZEL, E. FETZ, AND E. HENNEMAN.
in spinal motoneurons from "quantum" aspects. J. Physiol. Postsynaptic population potentials recorded from ventral roots
London 201: 479-493, 1969. perfused with isotonic sucrose: connections of groups Ia and II
221. KUNO, M., Y. MIYATA, AND E. J. MUNOZ-MARTINEZ. Differ- spindle afferent fibers with large populations of motoneurons.
ential reactions of fast and slow alpha motoneurons to axo- J. Neurophysiol. 42: 1146-1164, 1979.
tomy. J. Physiol. London 240: 725-739, 1974. 245. LUSCHER, H.-R., P. RUENZEL, AND E. HENNEMAN. How the
222. KUNO, M., E. J. MUNOZ-MARTINEZ, AND M. RANDlC. Synaptic size of motoneurons determines their susceptibility to dis-
action in Clarke's column neurons in relation to afferent ter- charge. Nature London 282: 859-861, 1979.
minal size. J. Physiol. London 228: 343-360, 1973. 246. LUSCHER, H.-R., P. RUENZEL, AND E. HENNEMAN. Topo-
223. KUNO, M., S. A. TURKANIS, AND J. N. WEAKLY. Correlation graphic distribution of terminals Ia and Group II fibers in the
between nerve terminal size and transmitter release at the spinal cord, as revealed by post-synaptic population potentials.
neuromuscular junction of the frog. J. Physiol. London 213: J. Neurophysiol. 43: 968-985, 1980.
545-566, 1971. 247. Lux, H. D., P. SCHUBERT, AND G. W. KREUTZBERG. Direct
224. KUNO, M., AND J. N. WEAKLY. Quantal components of the matching of morphological and electrophysiological data in cat
inhibitory synaptic potential in spinal motoneurons of the cat. spinal motoneurons. In: Excitatory Synaptic Mechanisms, ed-
J. Physiol. London 224: 287-303, 1972. ited by P. Anderson and J. K. S. Jansen. Oslo: Universitetsfor-
225. LAMB, A. The projection patterns of the ventral horn to the laget, 1970, p. 189-198.
hind limb during development. Dev. Biol. 54: 82-99, 1976. 248. MAGHERINI, P. C., W. PRECHT, AND P. C. SCHWINDT. Evidence
226. LAMB, A. H. Neuronal death in the development of the soma- for electrotonic coupling between frog motoneurons in the in
totopic projections of the ventral horn in Xenopus. Brain Res. situ spinal cord. J. Neurophysiol. 39: 474-483, 1976.
134: 145-150, 1977. 249. MATTHEWS, P. B. C. The dependence oftension upon extension
227. LANDMESSER, L. The development of motor projection pat- in the stretch reflex of the soleus muscle of the decerebrate
terns in the chick hind limb. J. Physiol. London 284: 391-414, cat. J. Physiol. London 147: 521-546, 1959.
1978. 250. MATTHEWS, P. B. C. The response of de-efferented muscle
228. LANDMESSER, L., AND D. G. MORRIS. The development of spindle receptors to stretching at different velocities. J. Phys-
functional innervation in the hind limb of the chick embryo. J. iol. London 168: 660-678, 1963.
Physiol. London 249: 301-326, 1975. 251. MATTHEWS, M. A., W. D. WILLIS, AND V. WILLIAMS. Dendrite
229. LAPORTE, Y., AND D. LLOYD. Nature and significance of the bundles in lamina IX of cat spinal cord: a possible source for
reflex connections established by large afferent fibers of mus- electrical interaction between motoneurons. Anat. Rec. 171:
cular origin. Am. J. Physiol. 169: 609-621, 1952. 313-328, 1971.
230. LAVAIL, J. H., AND M. M. LAVAIL. The retrograde intra-axonal 252. McLAUGHLIN, B. J. Dorsal root projections to the motor nuclei
transport of horseradish peroxidase in the chick visual system: in the cat spinal cord. J. Compo Neurol. 144: 461-474, 1972.
a light and electron microscopic study. J. Compo Neurol. 157: 253. MCPHEDRAN, A. M., R. B. WUERKER, AND E. HENNEMAN.
303-358, 1974. Properties of motor units in a homogeneous red muscle (soleus)
231. LENNERSTRAND, G. Position and velocity sensitivity of muscle of the cat. J. Neurophysiol. 28: 71-84, 1965.
spindles in the cat. I. Primary and secondary endings deprived 254. MENDELL, L. M., AND E. HENNEMAN. Terminals of single Ia
of fusimotor activation. Acta Physiol. Scand. 73: 281-299, 1968. fibers: distribution within a pool of 300 homonymous motor
232. LETBETTER, W. D., W. D. WILLIS, AND W. M. THOMPSON. neurons. Science 160: 96-98, 1968.
Monosynaptic excitation of motoneurons by single dorsal root 255. MENDELL, L. M., AND E. HENNEMAN. Terminals of single Ia
ganglion cells. Federation Proc. 27: 749, 1968. fibers: location, density, and distribution within a pool of 300
233. LIDDELL, E. G. T., AND C. S. SHERRINGTON. Reflexes in homonymous motoneurons. J. Neurophysiol. 34: 171-187,
response to stretch (myotatic reflexes). Proc. R. Soc. London 1971.
Ser. B 96: 212-242, 1924. 256. MENDELL, L. M., AND R. WEINER. Analysis of pairs of individ-
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 505

ual Ia-EPSPs in single motoneurons. J. Physiol. London 255: 280. PETERSON, R. P. Cell size and rate of protein synthesis in
81-104, 1976. ventral horn neurons. Science 153: 1413-1414, 1966.
257. MILL, P. J. Neural patterns associated with ventilatory move- 281. PETTE, D., M. KLINGENSBERG, AND TH. BUCHER. Comparable
ments in dragonfly larvae. J. Exp. Biol. 52: 67, 1970. and specifIc proportions in the mitochondrial enzyme activity
258. MILNER-BROWN, H. S., R. B. STEIN, AND R. YEMM. The pattern. Biochem. Biophys. Res. Commun. 7: 425-429, 1962.
orderly recruitment of human motor units during voluntary 282. PORTER, R., AND J. HORE. Time course of minimal cortico-
isometric contractions. J. Physiol. London 230: 359-370, 1973. motoneuronal excitatory postsynaptic potentials in lumbar
259. MILNER-BROWN, H. S., R. B. STEIN, AND R. YEMM. Changes motoneurons of the monkey. J. Neurophysiol. 32: 443-451,
in fIring rate of human motor units during linearly changing 1969.
voluntary contractions. J. Physiol. London 230: 371-390, 1973. 283. PROSKE, U., AND P. M. E. WAITE. Properties of types of motor
260. MISHELEYICH, D. J. Repetitive fIring to current in cat moto- units in the medial gastrocnemius muscle of the cat. Brain
neurons as a function of muscle unit twitch type. Exp. Neurol. Res. 67: 89-101, 1974.
25:401-409,1969. 284. RACK, P. M. H., AND D. R. WESTBURY. The effects of length
261. MONSTER, A. W., AND H. CHAN. Isometric force production by and stimulus rate on tension in the isometric cat soleus muscle.
motor units of extensor digitorum communis muscle in man. J. d. Physiol. London 204: 443-460, 1969.
Neurophysiol. 40: 1432-1443, 1977. 285. RALL, W. Branching dendritic trees and motoneuron mem-
262. MOSHER, C. G., R. L. GERLACH, AND D. G. STUART. Soleus brane resistivity. Exp. Neurol. 1: 491-527, 1959.
and anterior tibial motor units of the cat. Brain Res. 44: 1-11, 286. RALL, W. Theoretical signifIcance of dendrites for neuronal
1972. input-output relations. In: Neuronal Theory and Modeling:
263. MUNSON, J. B., AND G. W. SYPERT. Latency-rise time relation- Proceedings, edited by R. Reiss. Stanford, CA: Stanford Univ.
ship in unitary postsynaptic potentials. Brain Res. 151: 404- Press, 1964, p. 73-97. (Ojai Symp., 1962.)
408,1978. 287. RALL, W. Distinguishing theoretical synaptic potentials com-
264. MUNSON, J. B., AND G. W. SYPERT. Properties of central Ia puted for different soma-dendritic distributions of synaptic
afferent fIbers projecting to motoneurons. Neurosci. Abstr. 4: input. J. Neurophysiol. 30: 1138-1168, 1967.
569,1978. 288. RALL, W. Core conductor theory and cable properties of neu-
265. MUNSON, J. B., AND G. W. SYPERT. Properties of single central rons. In: Handbook of Physiology. The Nervous System. Cel-
Ia fIbers projection to motoneurons. J. Physiol. London 296: lular Biology of Neurons, edited by E. Kandel. Bethesda, MD:
315-328, 1979. Am. Physiol. Soc., 1977, sect. I, vol. I, pt. 1, chapt. 3, p. 39-97.
266. MUNSON, J. B., AND G. W. SYPERT. Properties of single fIber 289. RALL, W., R. E. BURKE, T. G. SMITH, P. G. NELSON, AND K.
excitatory post-synaptic potentials in triceps surae motoneu- FRANK. Dendritic location of synapses and possible mecha-
rons. J. Physiol. London 296: 329-342, 1979. nisms for the monosynaptic EPSP in motoneurons. J. Neuro-
267. NELSON, P. G. Interaction between spinal motoneurons of the physiol. 30: 1169-1193, 1967.
cat. J. Neurophysiol. 29: 275-287, 1966. 290. RALL, W., AND J. RINZEL. Branch input resistance and steady
268. NELSON, S. G., T. C. COLLATOS, A. NIECHAJ, AND L. M. attenuation for input to one branch of a dendritic neuron
MENDELL. Immediate increase in la-motoneuron synaptic model. Biophys. J. 13: 648-688, 1973.
transmission caudal to spinal cord transection. J. Neurophys- 291. RAMON Y CAJAL, S. Histologie du Systeme de l'Homme et des
iol. 42: 655-664, 1979. Vertebres. Paris: Maloine, 1909.
269. NELSON, S. G., AND L. M. MENDELL. Projection of single knee 292. RAPAPORT, S., A. SUSSWEIN, Y. UCHINO, AND V. J. WILSON.
flexor Ia fIbers to homonymous and heteronymous motoneu- Synaptic actions of individual vertibular neurons on cat neck
rons. J. Neurophysiol. 41: 778-787, 1978. motoneurons. J. Physiol. London 272: 367-382, 1977.
270. NELSON, S. G., AND L. M. MENDELL. Enhancement in Ia- 293. REDFERN, P. A. Neuromuscular transmission in newborn rats.
motoneuron synaptic transmission caudal to chronic spinal ,J. Physiol. London 209: 701-709,1970.
cord transection. J. Neurophysiol. 42: 642-654, 1979. 294. REDMAN, S. J. A quantitative approach to the integrative
271. NYSTROM, B. Postnatal development of motor nerve terminals function of dendrites. In: International Review of Physiology:
in "slow-red" and "fast-white" cat muscles. Acta Neurol. Neurophysiology II, edited by R. Porter. Baltimore: Univ.
Scand. 44: 363-383, 1968. Park, 1976, vol. 10, p. 1-36.
272. OLSON, C. B., D. O. CARPENTER, AND E. HENNEMAN. Orderly 295. REDMAN, S. J. Junctional mechanisms at group Ia synapses.
recruitment of muscle action potentials: motor unit threshold Prog. Neurobiol. NY 12: 33-83, 1979.
and EMG amplitude. Arch. Neurol. Chicago 19: 591-597,1968. 296. RENSHAW, B. Activity in the simplest spinal reflex pathways.
273. OLSON, C. B., AND C. P. SWETT, JR. A functional and histo- d. Neurophysiol. 3: 373-387, 1940.
chemical characterization of motor units in a heterogeneous 297. RENSHAW, B. Influence of discharge of motoneurons upon
muscle (flexor digitorum longus) of the cat. J. Compo Neurol. excitation of neighboring motoneurons. J. Neurophysiol. 4:
128:475-498,1966. 167-183, 1941.
274. PADYKULA, H. A., AND G. F. GAUTHIER. The ultrastructure of 298. RENSHAW, B. Central effects of centripetal impulses in axons
the neuromuscular junctions of mammalian red, white and of spinal ventral roots. d. Neurophysiol. 9: 191-204, 1946.
intermediate skeletal muscle fIbers. J. Cell Biol. 46: 27-41, 299. RICHMOND, F. J. R., AND V. C. ABRAHAMS. Morphology and
1970. enzyme histochemistry of dorsal muscles of the cat neck. d.
275. PARNAS, I. Differential block at high frequency of branches of Neurophysiol. 38: 1312-1321, 1975.
a single axon innervating two muscles. J. Neurophysiol. 35: 300. RINZEL, J., AND W. RALL. Transient response in a dendritic
903-914, 1972. neuron model for current injected at one branch. Biophys. d.
276. PASCOE, J. E. The responses of anterior horn cells to applied 14: 759-790, 1974.
currents. Acta Physiol. Scand. 42: (Suppl. 145) 112-113, 1957. 301. ROEDER, K. D., L. TOZIAN, AND E. A. WEIANT. Endogenous
277. PEARSON, K. G., AND J. F. ILES. Discharge patterns of coxal nerve activity and behavior in the mantis and cockroach. J.
levator and depressor motoneurons of the cockroach, Peripla- Insect Physiol. 4: 45, 1960.
neta americana. J. Exp. Biol. 52: 139, 1970. 302. ROMANES, G. J. The motor cell columns of the lumbo-sacral
278. PEARSON, K. G., R. B. STEIN, AND S. K. MALHOTRA. Properties spinal cord of the cat. d. Compo Neri~ol. 94: 313-363, 1951.
of action potentials from insect motor nerve fIbers. J. Exp. 303. RONNEYI, L. 0., AND S. CONRADI. Ultrastructural evidence for
Biol. 53: 299-316, 1970. spontaneous elimination of synaptic terminals on spinal motor
279. PENNY, J. E., J. R. KUKUMS, M. J. EADIE, AND J. H. TYRER. neurons in the kitten. Brain Res. 80: 335-339, 1974.
Quantitative oxidative enzyme histochemistry of the spinal 304. ROWELL, C. H. F., AND J.M. McKAY. An acridid auditory
cord. Part 1. Distribution of enzyme activity in anterior horn interneuron. I. Functional connexions and response to single
cells. J. Neurol. Sci. 26: 179-185, 1975. sounds. d. Exp. Biol. 51: 231, 1969.
 506 HANDBOOK OF PHYSIOLOGY - THE NERVOUS SYSTEM II

305. RUDOMfN, P., R K BURKE, R NUNEZ, J. MADRID, AND H. fiber patterns and connections. Proc. Natl. Acad. Sci. USA.
DUTTON. Control by presynaptic correlation: a mechanism 50: 703-710, 1951.
affecting information transmission from Ia fibers to motoneu- 327. SPRAGUE, J. M. The distribution of dorsal root fibers on motor
rons. J. Neurophysiol. 38: 267-284, 1975. cells in lumbosacral spinal cord of the cat and the site of
306. RYALL, R W. Renshaw cell mediated inhibition of Renshaw excitatory and inhibitory terminals in monosynaptic pathways.
cells: patterns of excitation and inhibition from impulses in Proc. R. Soc. London Ser. B 149: 534-556, 1958.
motor axon collaterals. J. Neurophysiol. 33: 257-270, 1970. 328. STAUFFER, K K, D. G. D. WATT, A. TAYLOR, R M. REINKING,
307. RYALL, R W., M. F. PIERCEY, C. POLOSA, AND J. GOLDFARB. AND D. G. STUART. Analysis of muscle receptor connections
Excitation of Renshaw cells in relation to orthodromic and by spike triggered averaging. 2. Spindle group II afferents. J.
antidromic excitation of motoneurons. J. Neurophysiol. 35: Neurophysiol. 39: 1393-1402, 1976.
137-148, 1972. 329. STEIN, R B., AND R BERTOLDI. The size principle: a synthesis
308. SALMONS, S., AND F. A. SRETER. Significance of impulse activ- of neurophysiological data. In: Progress in Clinical Neuro-
ity in the transformation of skeletal muscle type. Nature physiology. Recruitment Patterns of Motor Units and the
London 263: 30-34, 1976. Gradation of Muscle Force, edited by J. K Desmedt. Basel:
309. SCHEIBEL, M. K, AND A. B. SCHEIBEL. Spinal motoneurons, Karger, In press.
interneurons and Renshaw cells. A Golgi study. Arch. Ital. 330. STEPHENS, J. A., R GARNETT, AND N. P. BULLER. Reversal of
Biol. 104: 328-353, 1966. recruitment order of single motor units produced by cutaneous
310. SCHEIBEL, M. E., AND A. B. SCHEIBEL. Terminal patterns in stimulation during voluntary muscle contraction in man. Na-
cat spinal cord. III. Primary afferent collaterals. Brain Res. 13: ture London 272: 362-364, 1978.
417-443, 1969. 331. STERLING, P., AND H. G. J. M. KUYPERS. Simultaneous dem-
311. SCHEIBEL, M. E., AND A. B. SCHEIBEL. Organization of spinal onstration of normal boutons and degenerating nerve fibers
motoneuron dendrites in bundles. Exp. Neurol. 28: 106-112, and their terminals in the spinal cord. J. Anat. 100: 723-732,
1970. 1966.
312. SCHEIBEL, M. E., AND A. B. SCHEIBEL. Developmental rela- 332. STERLING, P., AND H. G. J. M. KUYPERS. Anatomical organi-
tionship between spinal motor neuron dendrite bundles and zation of the brachial spinal cord of the cat. II. The motoneuron
patterned activity in the forelimb of cats. Exp. Neurol. 30: 367- plexus. Brain Res. 4: 16-32, 1967.
373, 1971. 333. SZEKELY, G., G. CZEH, AND G. VOROS. The activity pattern of
313. SCHEIBEL, M. E., AND A. B. SCHEIBEL. Inhibition and the limb muscles in freely moving normal and deafferented newts.
Renshaw cell: a structural critique. Brain Behav. Evol. 4: 53- Exp. Brain Res. 9: 53, 1969.
93, 1971. 334. SZENTAGOTHAI, J. The anatomical basis of synaptic transmis-
314. SCHMIDT, R F. Presynaptic inhibition in the vertebrate ner- sion of excitation and inhibition of motoneurons. Acta Mor-
vous system. Ergeb. Physiol. Biol. Chem. Exp. Pharmakol. 63: phol. Acad. Sci. Hung. 8: 287-309, 1958.
20-101, 1971. 335. TANJI, J., AND M. KATO. Recruitment of motor units in vol-
315. SCOTT, J. G., AND L. M. MENDELL. Individual EPSPs produced untary contraction of a finger muscle in man. Exp. Neurol. 40:
by single triceps surae Ia afferent fibers in homonymous and 759-770, 1973.
heteronymous motoneurons. J. Neurophysiol. 39: 679-692, 336. TEIG, E. Tension and contraction times of motor units of the
1976. middle ear muscles in the cat. Acta Physiol. Scand. 84: 11-21,
316. SEARS, T. A. Some properties and reflex connections of respi- 1972.
ratory motoneurons of the eat's thoracic spinal cord. J. Physiol. 337. THOMAS, J. S., K M. SCHMIDT, AND F. T. HAMBRECHT. Facility
London 175: 386-403,1964. of motor unit control during tasks defined directly in terms of
317. SEARS, T. A., AND D. STAGG. Short-term synchronization of unit behaviors. Exp. Neurol. 59: 384-395, 1978.
intercostal motoneuron activity. J. Physiol. London 263: 357- 338. TRAUB, R D. Motoneurons of different geometry and the size
381, 1976. principle. Biol. Cybernetics 25: 163-176, 1977.
318. SHAPOVALOV, A. I., AND N. R GUREVICH. Monosynaptic and 339. VAN BUREN, J. M., AND K FRANK. Correlation between the
disynaptic reticulospinal actions on lumbar motoneurons of morphology and potential field of a spinal motor nucleus in
the cat. Brain Res. 21: 249-263, 1970. the cat. Electroencephalogr. Clin. Neurophysiol. 19: 112-126,
319. SHAPOVALOV, A. I., O. A. KARAMYAN, G. G. KURCHAVYI, AND 1965.
Z. A. REPINA. Synaptic actions evoked from the red nucleus 340. VAN DE GRAAFF, K M., K C. FREDERICK, R G. WILLIAMSON,
on the spinal alpha-motoneurons in the Rhesus monkey. Brain AND G. K GOSLOW, JR. Motor units and fiber types of primary
Res. 32: 325-348, 1971. ankle extensors of the skunk (Mephitis mephitis). J. Neuro-
320. SHAPOVALOV, A. I., AND G. G. KURCHAVYI. Effects of trans- physiol. 40: 1424-1431, 1977.
membrane polarization and TEA injection on monosynaptic 341. WAGMAN, I. H., D. S. PIERCE, AND R K BURGES. Propriocep-
actions from motor cortex, red nucleus and group Ia afferents tive influence in volitional control of individual motor units.
on lumbar motoneurons in the monkey. Brain Res. 82: 49-67, Nature London 207: 957-958, 1965.
1974. 342. WALCOTT, B., AND M. BURROWS. The ultrastructure and phys-
321. SHERRINGTON, C. S. The Integrative Action of the Nervous iology of the abdominal airguide retractor muscles in the giant
System. New Haven, CT: Yale Univ. Press, 1906. water bug, Lethocerus. J. Insect Physiol. 15: 1865, 1969.
322. SHERRINGTON, C. S. Flexion-reflex of the limb, crossed exten- 343. WANI, A. M., AND S. K GUHA. A model for gradation of
sion-reflex and reflex stepping and standing. J. Physiol. Lon- tension-recruitment and rate coding. Med. Biol. Eng. 13: 870-
don 40: 28-121, 1910. 874,1975.
323. SHINODA, Y, A. P. ARNOLD, AND H. ASANUMA. Spinal branch- 344. WARMOLTZ, J. R, AND W. K ENGEL. Open-biopsy electro-
ing of corticospinal axons in the cat. Exp. Brain Res. 26: 215- myography. I. Correlation of motor unit behavior with histo-
234,1976. chemical muscle fiber type in human limb muscle. Arch. Neu-
324. SNOW, P. J., P. K ROSE, AND A. G. BROWN. Tracing axons rol. 27: 512-517, 1972.
and axon collaterals of spinal neurons using intracellular injec- 345. WATT, D. G. D., K K STAUFFER, A. TAYLOR, R M. REINKING,
tion of horseradish peroxidase. Science 191: 312-313, 1976. AND D. G. STUART. Analysis of muscle receptor connections
325. SOMJEN, G., D. O. CARPENTER, AND E. HENNEMAN. Responses by spike-triggered averaging. 1. Spindle primary and tendon
of motoneurons of different sizes to graded stimulation of organ afferents. J. Neurophysiol. 39: 1375-1392, 1976.
supraspinal centers of the brain. J. Neurophysiol. 28: 958-965, 346. WENDT, I. R, AND C. L. GIBBS. Energy production of rat
1965. extensor digitorum longus muscle. Am. J. Physiol. 224: 1081-
326. SPERRY, R W. Chemoaffinity in the orderly growth of nerve 1086,1973.
 CHAPTER 11: ORGANIZATION OF MOTONEURON POOL 507

347. WILLIS, W. D., R. NUNEZ, AND P. RUDOMIN. Excitability gastrocnemius) of the cat. J. Neurophysiol. 28: 85-99, 1965.
changes of terminal arborizations of single Ia and Ib afferent 354. WYCKOFF, R. W. G., AND J. Z. YOUNG. The motor neuron
fibers produced by muscle and cutaneous conditioning volleys. surface. Proc. R. Soc. London Ser. B 144: 440-450, 1955.
J. Neurophysiol. 39: 1150-1159, 1976. 355. WYMAN, R. J. Somatotopic connectivity or species recognition
348. WILLIS, W. D., AND J. C. WILLIS. Properties ofinterneurons in connectivity? In: Control of Posture and Locomotion, edited
the ventral spinal cord. Arch. Ital. Biol. 104: 354-386, 1966. by R. B. Stein, K. B. Pearson, R. S. Smith, and J. B. Redford.
349. WILSON, V. J., AND M. YOSHIDA. Comparison of effects of New York: Plenum, 1974, p. 45-53. (Adv. Behav. BioI. Ser., vol.
stimulation of Deiters' nucleus and medial longitudinal fascic- 7.)
ulus on neck, forelimb, and hindlimb motoneurons. J. Neuro- 356. WYMAN, R. J., I. WALDRON, AND G. W. WACHTEL. Lack of
physiol. 32: 743-758, 1969. fixed order of recruitment in cat motoneuron pools. Exp. Brain
350. WILSON, V. J., AND M. YOSHIDA. Monosynaptic inhibition of Res. 20: 101-114, 1974.
neck motoneurons by the medial vestibular nucleus. Exp. 357. YATSUKI, Y., S. YOSHINO, AND J. CHEN. Action current of the
Brain Res. 9: 365-380, 1969. single lateral line nerve of the fish. II. On the discharge due to
351. WINDHORST, V. Cross correlations between discharge patterns stimulation. Jpn. J. Physiol. 1: 179-194, 1951.
of primary muscle spindle endings in active triceps-surae mus- 358. YELLIN, H. Neural regulation of enzymes in muscle fibers of
cles of the cat. Neurosci. Lett. 5: 63-67, 1977. red and white muscle. Exp. Neurol. 19: 92-103, 1967.
352. WUERKER, R. B., AND E. HENNEMAN. Reflex regulation of 359. YEMM, R. The orderly recruitment of motor units of the
primary (annulospiral) stretch receptors via gamma motoneu- masseter and temporal muscles during voluntary isometric
rons in the cat. J. Neurophysiol. 26: 539-550, 1963. contraction in man. J. Physiol. London 265: 163-174, 1977.
353. WUERKER, R. B., A. M. MCPHEDRAN, AND E. HENNEMAN. 360. ZUCKER, R. S. Theoretical implications ofthe size principle of
Properties of motor units in a heterogeneous pale muscle (M. motoneuron recruitment. J. Theor. Biol. 38: 587-596, 1973.