Human auditory system response to

modulated
electromagnetic
energy
ALLAN
H. FREY
General Electric Advanced Electronics Center, Cornell University,
Ithaca, New York

FREY, ALLAN H.
electromagnetic
energy.

Human

auditory

system response

to modulated

J. Appl. Physiol. I ~(4) : 68g--692.

I g62.The intent of this paper is to bring a new phenomenon to the
attention of physiologists. Using extremely low average power
densities of electromagnetic energy, the perception of sounds
was induced in normal and deaf humans. The effect was induced several hundred feet from the antenna the instant the
transmitter was turned on, and is a function of carrier frequency and modulation.
Attempts were made to match the
sounds induced by electromagnetic
energy and acoustic
energy. The closest match occurred when the acoustic amplifier
was driven by the rf transmitters modulator.
Peak power
density is a critical factor and, with acoustic noise of approximately 80 db, a peak power density of approximately 275 mw/
cm2 is needed to induce the perception at carrier frequencies
of 425 mc and I ,31 o mc. The average power density can be at
least as low as 400 pw/cm2. The evidence for the various possible sites of the electromagnetic
energy sensor are discussed
and locations peripheral to the cochlea are ruled out.

SIGNIFICANT
AMOUNT
OF RESEARCH
has been concerned with the effects of radio-frequency
(rf) energy on
I kc and
organisms
(electromagnetic
energy between
I oo Gc). Typically,
this work has been concerned
with
determining
damage resulting
from body temperature
increase. The average power densities used have been on
the order of o. I-I w/cm2 used over many minutes to
several hours.
In contrast, using average power densities measured in
microwatts
per square centimeter,
we have found that
other effects, which are transient,
can be induced with
this energy. Further,
these effects occur the instant the
transmitter
is turned on. With appropriate
modulation,
the perception
of various sounds can be induced
in
clinically
deaf, as well as normal, human subjects at a
distance of inches up to thousands of feet from the transmitter. With somewhat different transmitter
parameters,
we can induce the perception
of severe buffeting
of the
head, without
such apparent
vestibular
symptoms
as
dizziness or nausea. Changing
transmitter
parameters
again, one can induce a pins-and-needles
sensation.

Received

for publication

29 September

1961.

Experimental
work with these phenomena
may yield
information
on auditory
system functioning
and, more
generally,
information
on nervous system function.
For
example, this energy could possibly be used as a tool to
explore nervous system coding, possibly using Neider and
Neffs procedures
(I), and for stimulating
the nervous
system without
the damage caused by electrodes.
Since most of our data have been obtained
on the rf
sound and only the visual system has previously
been
shown to respond to electromagnetic
energy, this paper
will be concerned
only with the auditory effects data. As
a further restrictian,
only data from human subjects will
be reported,
since only these data can be discussed
meaningfully
at the present time. The long series of
studies we performed
to ascertain that we were dealing
with a biologically
significant
phenomenon
(rather than
broadcasts from sources such as loose fillings in the teeth)
are summarized
in another paper (2), which also reports
on the measuring
instruments
used in this work.
The intent of this paper is to bring this new phenomenon
to the attention
of physiologists.
The data reported are intended
to suggest numerous lines of experimentation
and indicate necessary experimental
controls.
Since we were dealing with a significant phenomenon,
we decided to explore the effects of a wide range of transmitter
parameters
to build up a body of knowledge
which would allow us to generate hypotheses and determine what experimental
controls would
be necessary.
Thus, the numbers given are conservative; they should not
be considered
precise, since the transmitters
were never
located in ideal laboratory
environments.
Within
the
limits of our measurement,
the orientation
of the subject
in the rf field was of little consequence.
Most of the transmitters
used to date in the experimentation
have been pulse modulated
with no information placed on the signal. The rf sound has been described as being a buzz, clicking,
hiss, or knocking,
depending
on several transmitter
parameters,
i.e., pulse
width
and pulse-repetition
rate (PRF). The apparent
source of these sounds is localized by the subjects as being
within,
or immediately
behind,
the head. The sound
always seems to come from within or immediately
behind

ALLAN

690
the head, no matter how the subject twists or rotates in
the rf field.
Our early experimentation,
performed
using transmitters with very short square pulses and high pulserepetition
rates, seemed to indicate that we were dealing
with harmonics of the PR.F. However, our later work has
indicated
that this is not the case; rather, the rf sound
appears to be the incidental
modulation
envelope on each
pulse, as shown in Fig. I.
Some difficulty
was experienced
when the subjects
tried to match the rf sound to ordinary
audio. They reported that it was not possible to satisfactorily
match the
rf sound to a sine wave or to white noise. An audio
amplifier was connected to a variable bandpass filter and
pulsed by the transmitter
pulsing mechanism.
The subjects, when allowed to control the filter, reported a fairly
satisfactory match. The subjects were fairly well satisfied
when all frequencies below 5-kc audio were eliminated
and the high-frequency
audio was extended as much as
possible. There was, however, always a demand for more
high-frequency
components.
Since our tweeter
has a
rather good high-frequency
response, it is possible that
we have shown an analogue of the visual phenomenon
in
which people see farther into the ultraviolet
range when
the lens is eliminated
from the eye. In other words, this
may be a demonstration
that the mechanical
transmission system of the ossicles cannot respond to as high a
frequency as the rest of the auditory system. Since the rf
bypasses the ossicle system and the audio given the sub-

TRANSMITTER ELECTRONICNOISE
( INCIDENTAL MODULATION)

ON OFF
Fro. I. Oscilloscope
time (pulse-modulated).

THRESHOLDS

As shown in Table I, we have used a fairly wide range

of transmitter
parameters.
We are currently
experimenting
with transmitters
that radiate energy at frequencies below 425 mc, and are using different types of
modulation,
e.g., pulse-repetition
rates as low as 3 and
41 sec.
In the experimentation
reported
in this section, the
ordinary
noise level was To-go db (measured
with a
General R.adio Co. model I 55 I-B sound-level
meter). In
order to minimize
the rf energy used in the experimentation, subjects wore Flent antinoise ear stopples whenever
measurements
were made. The ordinary noise attenuation
of the Flents is indicated
in Fig. 3. Although
the rf
sounds can be heard without
the use of Flents, even

Transmitter

I. Transmitter
Frequency,
mc

1,310
2,982
425
425
425
425
425
8,900

ON OFF
representation

of transmitter

output

FREY

ject for matching

does not, this may explain the dissatisfaction of our subjects in their matching.
At one time in our experimentation
with deaf subjects,
there seemed to be a clear relationship
between
the
ability to hear audio above 5 kc and the ability to hear
rf sounds. If a subject could hear above 5 kc, either by
bone or air conduction,
then he could hear the rf sounds.
For example, the threshold
of the subject whose audiogram appears in Fig. 2 was the same average power
density as our normal subjects. Recently,
however, we
have found people with a notch around 5 kc who do not
perceive the rf sound generated
by at least one of our
transmitters.

TABLE

H.

over

parameters
Wavelength, cm

Pulse

Width,
pSeC

22.9

10.4
70.6
70.6
70.6
70.6
70.6
3.4

I25
250
5oo
1000
2000

2.5

Pulses/Set

Duty

244
4oo

27
27
27
27
27
4oo

Cycle
.0015

.0004
.0038

,007
.014
.028
.056
.OOI

20

= 40
z
3
60
80
125

FIG.

normal

2.

Audiogram
rf sound

250

of deaf
threshold.

500 1000 2000 4000 8000

FREQUENCY (cps)
subject

(otosclerosis)

who

had

100

1000

l0,000

FREQUENCY
a

PIG.

stopples

3. Attenuation
of ambient
(collated
from Zwislocki

sound with Flent antinoise

(3) and Von Gierke
(4).

ear

AUDITORY

SYSTEM

AND

ELECTROMAGNETIC

ENERGY

2. Threshold for perception of rf sound

(ambient noise level To-go db)

TABLE

Transmitter

A
B
c
D
E
F
10,000,

Frequency,
mc

Duty
Cycle

1,310
2,982
425
425
425
425

.0015
.0004
.0038
.007
.014
.028

Avz
Power
Density,
mw/cm2

o-4
2.1
1.0

I 09
3*2
7J

Peak
Power
Density,
mw/cm2

Peak
Electric
Field,
v/cm

Peak.
Maw&

ampl
turns/m

267
5,250
263
271
129
254

I4
63
I5
14
I3
4

4
17
4
4
3
4

691
instruments
used in that experiment
did not read high
enough to give an accurate reading.
The energy from
transmitter
H was not perceived,
even when the peak
power density was as high as 25 w/cm2.
When the threshold energy is plotted as a function of
the rf energy (Fig. 4), a curve is obtained
which is suggestive of the curve of penetration
of rf energy into the
head. Figure 5 shows the calculated
penetration,
by frequency of rf energy, into the head. Our data indicate
that the calculated
penetration
curve may well be accurate at the higher frequencies
but the penetration
at
the lower frequencies
may be greater than that calculated on this model.
As nreviously noted, the thresholds were obtained in a
ABSORBED
FREQUENCIES

REFLECTED
FREQUENCIES

CORTICAL
TISSUE

10%

OF INCIDENT

BONE

/
1000
t

I,

100

FIG. 5. Microwave
neglecting
resonance
(from Nieset et al.

100

200
FIG.

magnetic

FREQUENCY
1
1000

( mc)
2000

4. Threshold
energy as a function
of frequency
energy
(ambient
noise level To-go db).

I0,000

1000

FREQUENCY

I00,000

(MC)

power
distribution
effects and considering
(5), modified).

in

a forehead
model
only first reflections

I
3000
of electro-

above an ambient noise level of go db, it appears that

the ambient
noise to some extent masked
the rf
sound.
of the
Table 2 gives the thresholds for the perception
rf sounds. It shows fairly clearly that the critical factor in
the perception
of the rf sound is the peak power density,
rather than the average power density. The relatively
high value for transmitter
B was expected and will be
discussed below. Transmitter
G has been omitted from
this table since the 2o-mw/cm2
reading for it can be considered only approximate.
The field-strength-measuring

FIG.

6. Area

portion).

most

sensitive

to electromagnetic

energy

(shaded

ALLAN

692
high ambient
noise environment.
This is an unusual
situation as compared
to obtaining
thresh01 .ds of regul ar
audio sound. Our recen t experimentation
leads us to
believe that, if the ambient noise level were not so high,
these threshold
field strengths would
be much lower.
Since one purpose of this paper is to suggest experiments,
it might be appropriate
to theorize
as to what the rf
sound threshold might be if we assume that the subject
is in an anechoic chamber.
It is also assumed that there
is no transducer
noise.
Given:
As a threshold
for the rf sound, a peak power
density
of 275 mw/cm2
determined
in an ambient
noise environment
of 80 db. Earplugs
attenuate
the ambient
noise 30 db.
If: I mw/cm2
is set equal to o db, then 275 mw/cm2
is equal to
24 db.
* Then: We can reduce the rf energy
50 db to - 26 db as we
reduce the noise level energy from 50 db to o db. We find that
-26
db rf energy is approximately
3 pw/cm2.
Thus:
In an anechoic
room, rf sound could
theoretically
be
induced
by a peak power density of 3 pw/cm2
measured
in free
space. Since only 10%
of this energy
is likely
to penetrate
the
skull, the human
auditory
system and a table radio may be one
order of magnitude
apart in sensitivity
to rf energy.
RF

DETECTOR

IN

AUDITORY

SYSTEM

One possibility that seems to have been ruled out in our

experimentation
is that of a capacitor-type
effect with the
tympanic
membrane
and oval window
acting as plates
of a capacitor.
It would seem possible that these membranes, acting as plates of a capacitor,
could be set in
motion by rf energy. There are, however, three points of
evidence against this possibility.
First, when one rotates
a capacitor in an rf field, a rather marked change occurs
in the capacitor
as a function
of its orientation
in the
field. When our subjects rotate or change the positions
of their heads in the field, the loudness of the rf sound
does not change appreciably.
Second, the distance between these membranes
is rather small, compared
with
the wavelengths
used. As a third point, we found that
one of our subjects who has otosclerosis
heard the rf
sound.

H. FREY

Another possible location for the detecting mechanism

is in the cochlea. We have explored this possibility
with
nerve-deaf people, but the results are inconclusive
due to
factors such as tinnitus, We are currently
exploring
this
possibility with animal preparations.
The third likely place for the detection mechanism
is
the brain. Burr and Mauro
(6) presented evidence that
indicates that there is an electrostatic field about neurons.
Morrow
and Sepiel (7) presented evidence that indicates
the existence of a magnetic field about neurons. Becker
(personal communication)
has done some work indicating that there is longitudinal
flow of charge carriers in
neurons. Thus, it is reasonable
to suspect that possibly
the electromagnetic
field could interact
with neuron
fields. As yet, evidence of this possibility
is inconclusive.
The strongest point against it is that we have not found
visual effects although
we have searched for them. On
the other hand, we have obtained
other nonauditory
effects and have found that the sensitive area for detecting
rf sounds is a region over the temporal lobe of the brain.
One can shield, with a a-in.2 piece of fly screen, a portion
of the stippled area shown in Fig. 6 and completely
cut
off the rf sound.
Another possibility
should also be considered.
There is
no good reason to assume that there is only one detector
site. On the contrary,
the work of Jones et al. (8), in
which they placed electrodes in the ear and electrically
stimulated
the subject, is sufficiently relevant to suggest
the possibility
of more than one detector
site. Also,
several sensations
have been elicited
with
properly
modulated
electromagnetic
energy. It is doubtful that all
of these can be attributed
to one detector.
As mentioned
earlier, the purpose of this paper is to
focus the attention
of physiologists
on an unusual area
and stimulate additional
work on which interpretations
can be based. Interpretations
have been deliberately
omitted from this paper since additional
data are needed
before a clear picture can emerge. It is hoped that the
additional
exploration
will also result in an increase in
our knowledge
of nervous system functions.

REFERENCES
I. NEIDER,

P. C., AND W. D. NEFF. Science I 33 : I OI o, 1961.

2. FREY, A. H. Aero Space Med. 32 : I I 40, I g6 I.
3. ZWISLOCKI,
J. Noise Control 4: 42, 1958.
4. VON GIERKE,
H. Noise Control 2 : 37, 1956.
5.

NIESET,

R., L. PINNEO,

R. BAUS, J. FLEMING,

AND

R. M~AFEE.

Ann.
65,

Rept.

USAF

Rome

Air Development

Command,

TR-61

1961.

H., AND A. MAURO.

Yale J. Biol.
and Med. 21 : 455, 1949.
7. MORROW,
R., AND J. SEIPEL. J. Wash. Acad. Sci. 50: I, x 960.
8. JONES, R. C., S. S. STEVENS, AND M. H. LURIE. J. Acoust.
Sot.
Am.
12: 281, 1940.

6. BURR,