Cognitive Therapy and Research, VoL 2, No. 4, 1978, pp.

323-333

Anxiety Reduction Following Exercise and Meditation

Michael S. Bahrke ~ and W i l l i a m P. M o r g a n ~

University of Wisconsin--Madison

The purpose of this investigation was to compare the influence of acute

physical activity and meditation ("noncultic") on state anxiety. Seventy-
five adult male volunteers served as Ss with 25 Ss randomly assigned to
either an exercise, meditation, or control group. Physical activity was per-
formed at 70% of self-imposed maximal exercise heart rate for 20 minutes
by Ss in the exercise group; Ss assigned to the meditation group practiced
Benson "s Relaxation Response for 20 minutes; and Ss in the control group
simply rested quietly in a "'Lazyboy'" chair for 20 minutes. State anxiety
was measured with the Spielberger Scale, and it was assessed (1) prior to, (2)
immediately following, and (3) 10 minutes following each treatment.
Oxygen consumption, heart rate, skin temperature, and blood pressure
were also measured as confirmatory variables under selected conditions.
The data were analyzed by means of a two-way repeated measures A N O VA,
and this analysis revealed that a significant reduction in anxiety occurred
for each treatment. This held for both those Ss falling within the normal
range for state anxiety and those Ss regarded as high-anxious. It was also
noted that none o f the physiological variables differed significantly follow-
ing the control and meditation treatments. The present evidence suggests
that acute physical activity, noncultic meditation, and a quiet rest session
are equally effective in reducing state anxiety.

Psychological problems involving anxiety states have been estimated to be

from 2 to 4°70 in the general population and from 16 to 25% in psychiatric
patient samples (Lader, 1972; Pitts, 1969; Wilkinson & Latif, 1974).
Furthermore, it has been estimated that 30 to 70070 of all patients currently

'Present address: Department of Health, Physical Education and Recreation, University of

Kansas, Lawrence, Kansas 66045.
2Present address: Exercise and Sport Science Laboratory, Department of Physical Education,
University of Arizona, Tucson, Arizona 85721.
323
0147-5916/78/1200-0323505.00/0 © 1 9 7 8 Plenum Publishing Corporation
324 Bahrke and Morgan

being treated by physicians in general practice are suffering from conditions

which have their origins in unrelieved stress (Pitts, 1969). The prevalence of
anxiety states makes it readily apparent that anxiety represents one of con-
temporary society's major health problems.
Therapies of various types, including autogenic training, biofeed-
back, drugs, hypnosis, exercise, meditation, and various relaxation tech-
niques have been employed as a means of reducing anxiety. In this context it
is noteworthy that acute physical activity (Byrd, 1965; de Vries, 1968; de
Vries, 1972; Morgan, 1973; Morgan, 1978) and meditation or relaxation
techniques (Benson, 1975; Benson & Wallace, 1972, Benson, Steinert,
Greenwood, Klemchuk, & Peterson, 1975; Ferguson & Gowan, 1975;
Wallace, 1970; Wallace & Benson, 1972), two seemingly divergent ther-
apies, are both capable of reducing tension and improving psychological
states. This becomes an important theoretical point when one considers that
meditation and relaxation techniques produce quiescence, whereas exercise
produces the converse, arousal. Hence, two procedures representing
opposite ends of the quiescence-arousal continuum produce a similar effect
--anxiety reduction. The assumption made, however, that peripheral
neurophysical and biochemical changes reflect central alterations, is de-
batable and certainly not proven. In other words, it is possible that changes
occurring centrally are not reflected peripherally. Also while both of these
techniques produce a decrement in tension state there has been no direct
comparison of these two divergent therapies heretofore.
The primary purpose of this investigation was to compare the effects
of acute physical activity and noncultic meditation on state anxiety. The
secondary purpose was to assess the influence of differential anxiety states
on the anxiety response.

METHOD

Subjects

The subjects consisted of 75 regularly exercising adult male volun-

teers. Subjects' ages ranged between 22 and 71 years with a mean of 51.9
years. The mean height and weight of this sample were 177.6 cm (SD =
6.5) and 78.4 kg (SD = 9.7), respectively.

Procedure

The experimental protocol was described to each subject with the

understanding that random assignment to one of the following three groups
Anxiety Reduction Following Exercise and Meditation 325

would follow: (1) exercise, (2) meditation, or (3) quiet rest (control).'Sub-
jects were asked to sign an informed consent document and complete a
24-hour history dealing with consummatory behavior and activity during
the previous day as well as their general state of health. Prior to being
randomly assigned to a group, anxiety was assessed by completion of the
STAI X-1 (State) and STAI X-2 (Trait) (Spielberger, Gorsuch, & Lushene,
1970). State anxiety (STAI X-I) was also measured immediately following
and 10 minutes following each experimental session. Electrodes for moni-
toring and recording cardiac frequency were placed in a C-5 position.

Treatment Conditions

Subjects in the exercise group (N = 25) walked for 20 minutes on a

motor-driven treadmill at 70% of their self-imposed maximal heart rate
(MHR) following 3 minutes of warm-up. Seventy percent of MHR was
calculated from each subject's most recent modified Balke Treadmill Test
(Balke & Ware, 1959). Subjects rated their perceived exertion at 5 minute
intervals during exercise according to Borg's revised psychophysical scale
(Borg, 1973). A 3 minute "warm-down" followed each exercise walk. Heart
rate was recorded during the final 15 seconds of each minute using a Gilson
Recorder. Postexercise state anxiety was obtained immediately and 10
minutes following exercise. Systolic and diastolic blood pressure were also
measured 10 minutes after cessation of exercise.
Subjects in the meditation group (N = 25) received tape-recorded
instructions describing the Relaxation Response (Benson, 1975), and they
practiced the technique while seated in a standard "Lazyboy" recliner.
Resting metabolism (oxygen consumption) was continuously monitored by
means of a Webb Metabolic Rate Monitor (MRM), and skin temperature
was measured throughout meditation with a United Systems Digitec digital
thermometer. Electrode placement for skin temperature was 1 inch proxi-
mal to the right wrist on the dorsal side, and skin temperature was recorded
at 2 minute intervals. Heart rate was continuously recorded on a Gilson
polygraph recorder with the final 15 seconds of each minute used to deter-
mine cardiac frequency. As in the exercise session state anxiety was
measured immediately following a n d again 10 minutes following the
session. Blood pressure was obtained at the completion of the 10 minute
time interval as well.
The control group (N = 25) rested quietly for 20 minutes in the
recliner and subjects were provided a current issue of the Reader's Digest to
read during this period if they so desired. The magazine contained no
articles relating to relaxation, exercise, or cardiovascular health. Oxygen
consumption, heart rate, and skin temperature were monitored and
326 Bahrke and Morgan

recorded as described earlier for the meditation procedure. State anxiety

levels were assessed immediately and 10 minutes following this session by
means of the STAI (Spielberger et al., 1970).
Means, standard deviations and standard errors were computed for
each variable. A series of two-way repeated measures ANOVAs for blood
pressure, state anxiety, skin temperature, heart rate, and oxygen consump-
tion was performed. In cases where significant F ratios ( p < .05) were
observed, a probe of the means was carried out utilizing the Newman-Keuls
procedure (Winer, 1971).
The results of the primary analysis are summarized in Figure 1. State
anxiety decreased across time for the exercise, meditation, and control
groups. A two-way repeated measures analysis of variance (Winer, 1971)
demonstrated a significant decrease (p < .05) in state anxiety across time (F
= 48.94). No significant differences (p< .05) were demonstrated between
groups (F = 1.82) or for the interaction between time and group (F = .32).
When the three groups were combined and subjects were divided into
high- (N = 10) and low- (N = 10) anxious categories based upon their initial
level of trait anxiety, mean state anxiety values decreased significantly
(p< .05) in both the high trait anxious (41.7 to 31.8) and in the low trait

34
T
32

30
C~

>- 2 8
I--
hi

26 "''"1~ 0
hi
24
0,-..O MEDITATION
1...........................
!
i
O'~O EXERCISE
22
D - - D CONTROL

I I I
PRE POST-I POST-2

EXPERIMENTAL PERIOD

Fig. 1. State anxiety before and following exercise, meditation,

and control treatments.
Anxiety Reduclion Following Exercise and Meditation 327

anxious subjects (27.4 to 22.0) across trials ( F = 27.02). An additional

analysis was carried out in which the state anxiety response of four high
trait anxious subjects from each group was evaluated. A significant
( p < .05) decrement in state anxiety was observed for the three groups
across trials ( F = 9.79), but the group (F = 1.20) and group by trials inter-
action ( F = .29) effects were not significant. An evaluation o f state anxiety
responsivity in high ( N = 11) and low ( N = 11) state anxious subjects re-
vealed that state anxiety decreased significantly ( p < .05) for the high-
anxious (44.7 to 32.1) but remained virtually unchanged for the low-anxious
subjects. This resulted in significant F r a t i o s for groups ( F = 102.53), trials
( F = 16.80), and groups by trials interaction ( F = 14.72). The nature o f
this interaction rules out statistical regression since the high-anxious de-
creased substantially, whereas the low-anxious did not change.
A small increase in skin temperatures occurred for both the control
and meditation groups. A two-way repeated measures ANOVA (25)
demonstrated a significant difference (p < .05) across time within groups ( F
= 6.73) but no significant differences between groups ( F = .71), nor inter-
action for time and group ( F = 1.28). The temperature of the test room was
stable, having a standard error of .99°C across the 20 minutes for the
control session and a standard error o f .73°C for the meditation session.

120

115
A

:E ~lo
n
in e---O MEDITATION (N,25)
io5
11,1 £3 .... O CONTROL (NI25)
I--
< 0 0 EXERCISE (N=25)

I,- 75
~g
,~ "to
I,i,I
65

60

55

I I I I I I I I I I I I I I I I I I [ I
2 4 6 8 I0 12 14 16 18 20

MINUTE
Fig. 2. Heart rate in beats per minute during exercise, meditation, and control treatments.
328 B a h r k e and Morgan

4.2
4.1 / ~ •--• MEDITATION(N:II)
4.0 ~
/

~
t

O---DCONTROL(N:II)
3.9
_.~ 3.8
X
• 3.7

! _ _ /.•

/tk\',e__i . ~ .
o (~
~ 3.4
', ,, ',i ,
3.3 ',, ; .

\ /
3.2
V'
3.1

, i i * I I I I I I I I I I I I I I I
2 4 6 8 I0 12 14 16 18 20
MINUTE
Fig. 3. Oxygen consumption during meditation and control treatments.

The increased skin temperature can be taken to reflect a state of relaxation,

and as such offers physiologic support for the self-report or state anxiety
data. However, it should be emphasized that the meditation and control
groups did not differ in this respect.
The response of heart rate to the three sessiOns is illustrated in Figure
2. A relatively constant heart rate occurred in each session. Initial values for
the meditation group were lower than those o f the control and exercise
groups and remained lower throughout the sessions. Final values were also
lower in the meditation group than those of the control and exercise groups.
A two-way repeated measures A N O V A demonstrated a significant differ-
ence ( p < .05) between the three groups ( F -- 261.77), and the Newman-
Keuls procedure showed the meditation and control groups to differ signifi-
cantly (p < .05) from the exercise group. However, no significant difference
was demonstrated between the control and meditation groups. No signifi-
cant difference (p < .05) was demonstrated across time for the groups ( F =
1.74), but the interaction of time and group was significant (p < .05). This
interaction was due to the increase in heart rate o f approximately 5 b p m
across the first 6 minutes for the exercise group. It will be noted in Figure 2,
however, that the exercise group achieved a virtual steady state after the
Anxiety Reduction Following Exercise and Meditation 329

fifth minute of exercise. The failure of heart rate to decrease following the
meditation or control period can be explained by the observation that
cardiac frequency was quite low in the two groups from the very outset
(61-67 bpm). While heart rate is often employed in psychophysiology as a
somatic indicant of anxiety, it often proves to be an inappropriate measure.
The present case is an example.
The consumption of oxygen across the 20 minute period for the con-
trol and meditation groups is illustrated in Figure 3. Data are presented for
only 11 subjects in each of these groups owing to equipment malfunction
during the testing of 14 subjects from each group. Following a moderate
increase in oxygen consumption, a decrease and then a plateau occurred for
each group. The subjects in the meditation group, however, initially con-
sumed less oxygen than did subjects in the control group, but after approxi-
mately 10 minutes the oxygen consumption values of both groups were very
similar. A two-way repeated measures ANOVA (Winer, 1971) demonstrated
no significant differences (p> .05) between groups (F = .01), across time
for each group (F = 1.00), nor for the interaction of time and group (F =
.86).

DISCUSSION

The results of the present inquiry support the expectation that acute
physical activity can reduce state anxiety. This study further corroborates
the existing literature which presents evidence that decreases in state anxiety
follow acute physical activity of a vigorous nature (Morgan, 1973; Morgan
& Horstman, 1976).
It is important to note that in the present investigation state anxiety
was also reduced in those subjects judged to be either high trait or high state
anxious. This result is in agreement with the anxiety reduction reported by
Morgan (1973) in that those subjects falling within the normal range on
anxiety, as well as those classified as high-anxious, experienced decrements.
In contrast, Sime (1977) was unable to demonstrate a significant decrease in
anxiety following acute exercise. However, previous research has shown
that light or moderate exercise does not have an influence on state anxiety
(Morgan, Roberts, & Feinerman, 1971) and, therefore, decrements in state
anxiety would not have been predicted in Sime's study. The exercise stim-
ulus used in the present study required a strenous level of activity (70% self-
imposed MHR).
The present data support the earlier work of Ferguson and Gowan
(1975) and Lazar, Farwell, & Farrow (1975), who reported that meditation
(TM) resulted in reduced anxiety. These data, however, are in disagreement
with the recent report by Sime (1977), who failed to observe a decrement in
330 Bahrke and Morgan

state anxiety following the same form of "noncultic" meditation (Benson,

1975) employed in the present investigation. The apparent disagreement
may be due to the fact that Sime (1977) employed meditation within a stress
paradigm involving test-induced anxiety, whereas the present study was pre-
sumably devoid of pretreatment stress.
Physiological response to the meditation treatment does not support
the findings of Wallace (1970), Wallace and Benson (1972), Benson and
Wallace (1972), and Benson (1975). These investigators have reported de-
creases in oxygen consumption and heart rate following both Transcen-
dental Meditation (TM) (Wallace, 1970; Wallace & Benson, 1972) and
"noncultic" meditation (Benson et al., 1975), which has been labeled the
relaxation response (RR) by Benson (1975). In the present investigation a
decrease in oxygen consumption occurred during the meditation period, but
the decrease was small and nonsignificant. It will also be recalled that heart
rate was quite constant throughout the 20 minutes of meditation in this
study. It should be pointed out that every effort was made in the present
investigation to obtain accurate as opposed to pseudo baselines from the
outset. In other words, the subjects rested quietly for a period of 20 minutes
before the meditation or control procedure began. Had the experimental
period begun as soon as they assumed a resting position in the chair it is
likely that a substantial decrease in metabolism and heart rate would have
occurred.
Perhaps the most important observation in this investigation, and a
truly serendipitous finding in our view, was the decrement in state anxiety
which accompanied the control treatment. The psychophysiologic respon-
sivity of the control subjects was identical with that of the meditation
group. Earlier research has demonstrated that both acute physical activity
(Morgan, 1973; Morgan & Horstman, 1976) and meditation (Ferguson &
Gowan, 1975; Lazar et al., 1975) are capable of reducing anxiety. The
primary purpose of the present investigation was to evaluate the compara-
tive efficacy of vigorous exercise (70% self-imposed MHR) and a popular
meditation procedure known as the relaxation response (Benson, 1975) in
their ability to reduce state anxiety. In other words, two seemingly different
procedures, one designed to provoke arousal (e.g., physical activity), and
the other quiescence (e.g., meditation), are commonly used to reduce
tension states. A rather basic and important question in our view related to
the comparative effectiveness of the two procedures. The use of a control or
placebo 3 group was simply employed in order to quantify the extent to

q n view of the fact that subjects in the " c o n t r o l " group were treated exactly the same as sub-
jects in the meditation group, with the exception that they were not introduced to the medita-
tion procedure, it might be more appropriate to regard the " c o n t r o l " group as a "placebo"
group. The " c o n t r o l " subjects were also seated in a sound-filtered room for a 20 minute
period in a comfortable reclining chair while oxygen consumption, heart rate, and skin tem-
perature were monitored.
Anxiety Reduction Following Exercise and Meditation 331

which the Hawthorne effect, or any treatment vs. no treatment (Morgan,

1972), was operative.
The present results and the recent reports by Michaels, Huber, &
McCann (1976) and de Vries, Burke, Hopper, & Sloan (1977) suggest that
simply taking "time out" is just as effective as exercise, biofeedback, or
meditation (cultic or noncultic) in reducing anxiety. Furthermore, this
appears to be the case whether anxiety is operationalized in a traditional
psychometric sense (present study), biochemically (Michaels et al., 1976), or
neurophysiologically (de Vries et al., 1977).
In many respects the present findings, along with those of Michaels et
al. (1976) and de Vries et al. (1977), raise far more questions about anxiety
states than they provide answers. First of all, these investigations support
the widely held beliefs that exercise, biofeedback, and meditation are
capable of reducing anxiety, but these studies finally, in a compelling
fashion, force the basic and fundamental issue of why. If these procedures
reduce anxiety, and they are clearly associated with anxiety decrements, is
there evidence of causality? In other words, do these treaments cause the
tension reduction to occur, or are they simply associated with the lowered
anxiety? While the assumption appears to be one of a causal function in
both the exercise and meditation literature, there has surprisingly been no
apparent evidence advanced in support of causality.
The question of causality has major theoretical, as well as applied
implications. The basis of beneficience, for example, resulting from exercise
and meditation may not rest in the physiologic changes provoked by such
procedures, but rather in the diversional aspects of such activities. It is con-
ceivable that "time-out" therapy, whether in the form of exercise, medi-
tation, biofeedback, or simple rest breaks, may represent the effective
ingredient in therapies designed to reduce anxiety. There are, of course,
many questions which should be addressed prior to adopting the view that
"time-out" therapy is just as effective as more traditional therapies. The
time course or decay of the response following various treatments must cer-
tainly be explored. It is possible, for example, that while "time-out" is just
as effective as exercise in reducing anxiety, the decrement following exercise
may be sustained for a longer period. In other words, the quantity of the
shift may be similar, whereas the quality may differ. In addition the long-
term or chronic effects of such acute changes may differ owing to the addi-
tive nature of such divergent therapies. Finally, another important con-
sideration relates to the target population; that is, "time-out," meditation,
and exercise may differentially influence individuals who have adopted
various lifestyles.
The present evidence suggests that acute physical activity, noncultic
meditation, and a quiet rest are equally effective in reducing state anxiety.
The finding that no difference exists between physical activity, meditation,
and control treatments in terms of anxiety reduction has implications for
332 Bahrke andMorgan

therapeutic practice. For those unable or unwilling to exercise, meditation

or a quiet rest could be effective therapy in reducing tension. Conversely,
individuals opposed to meditation feel that physical activity possesses
inherent physiological advantages not found in sedentary methods. The
underlying cause for the reduction of anxiety following exercise, noncultic
meditation, and quiet rest is open to speculation. There may be common
internal or central mechanisms operating during exercise, meditation, and
quiet rest. These mechanisms may go undetected by tools presently em-
ployed to assess tension levels. Perhaps taking "time-out" from the daily
activities is in itself sufficient in reducing anxiety. It is concluded that
vigorous physical activity of an acute nature, noncultic meditation, and
quiet rest periods are equally effective in reducing state anxiety, and this
holds for individuals who fall within the normal range on anxiety, as well as
those who are elevated on state and/or trait anxiety.

REFERENCES

Balke, B., & Ware, R. W. An experimental study of physical fitness of Air Force personnel.
U.S. Armed Forces Medical Journal, 1959, 10, 675-688.
Benson, H. The relaxation response. New York: William Morrow and Company, 1975.
Benson, H., & Wallace, R. K. Decreased blood pressure in hypertensive subjects who practice
meditation. Supplement II to Circulation 45 and 46, 1972.
Benson, H., Steinert, R. F., Greenwood, M. M., Klemchuck, H. M., & Peterson, N. H. Con-
tinuous measurement of oxygen consumption and carbon dioxide elimination during a
wakeful hypometabolic state. Journal of Human Stress, 1975, 1, 37-44.
Borg, G. Perceived exerticn: A note on "history and methods." Medicine and Science in
Sports, 1973, 5, 90-93.
Byrd, O. E. The relief of tension by exercise: A survey of medical viewpoints and practices.
Journal of School Health, 1965, 43, 239-240.
de Vries, H. A. Immediate, and long-term effects of exercise upon resting muscle action poten-
tial level. Journal of Sports Medicine and Physical Fitness, 1968, 8, 1-11.
de Vries, H. A., & Adams, G. M. Electromyographic comparison of single dose of exercise and
meprobamate as to effect on muscular relaxation. American Journal of Physical Medi-
cine, 1972, 51, 130-141.
de Vries, H. A., Burke, R. K., Hopper, R. T., & Sloan, J. H. Efficacy of EMG biofeedback in
relaxation training. American Journal o f Physical Medicine, 1977, 56, 75-81.
Ferguson, P. C., & Gowan, J. C. Psychological findings on Transcendental Meditation. Scien-
tific Research on the Transcendental Meditation Program, Switzerland, MERU Press
Publication, 1975 (1, Number S 180).
Lader, M. The nature of anxiety. British Journal of Psychiatry, 1972, 121, 481-491.
Lazar, Z., Farwell, W., & Farrow, J. The effects of the Transcendental Meditation program
on anxiety, drug abuse, cigarette smoking, and alcohol consumption. Scientific
Research on the Transcendental Meditation program, Switzerland, MERU Press
Publication, 1975 (1, Number S 180).
Michaels, R. R., Huber, M. J., & McCann, D. S. Evaluation of Transcendental Meditation as
a method of reducing stress. Science, 1976, 192, 1242-1244.
Morgan, W. P. Basic considerations. In W. P. Morgan (Ed.), Ergogenic aids and muscular
performance. New York: Academic Press, 1972.
Morgan, W. P. Influence of acute physical activity on state anxiety. Proceedings of the
National College Physical Education Association for Men, January 1973, 114-121.
Anxiety Reduction Following Exercise and Meditation 333

Morgan, W. P. Sport personology: The credulous-skeptical argument in perspective. In A. H.

Ismail (Ed.), Psycho-social behavior of sport and play. Indianapolis: Indiana State
Board of Health, 1978.
Morgan, W. P., & Horstman, D. H. Anxiety reduction following acute physical activity.
Medicine and Science in Sports, 1976, 8, 62.
Morgan, W. P., Roberts, J. A., & Feinerman, A. D. Psychologic effect of acute physical
activity. Archives of Physical Medicine and Rehabilitation, 1971, 52, 422-425.
Pitts, F. N. The biochemistry of anxiety. Scientific American, 1969, 220, 69-75.
Sime, W. E. A comparison of exercise and meditation in reducing physiological response to
stress. Medicine and Science in Sports, 1977, 9, 55.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. Manual for the State-Trait Anxiety
Inventory (STAI). Palo Alto, California: Consulting Psychologists Press, 1970.
Wallace, R. K. Physiological effects of transcendental meditation. Science, 1970, 167,
1751-1754.
Wallace, R. K., & Benson, H. The physiology of meditation. Scientific American, 1972, 226,
84-90.
Wilkinson, J. C. M., & Latif, K. Common neuroses in general practice. Bristol, England:
John Wright and Sons, 1974.
Winer, B. J. Statistical principles in experimental design. New York: McGraw-Hill, 1971.