Fairbanks, Grant. Experimental Phonetics – T18

TTS Produced by Low-Level Tones and the Effects
of Testing on Recovery **1

Donald W. Bell and Grant Fairbanks *2
Research Center, Subcommittee on Noise, Los Angeles, California
(Received 15 July 1963)

To evaluate TTS from exposure frequencies presented at low level and the effects of threshold testing
on recovery, sensation levels of 10, 20, 40, and 60 dB and frequencies of 1,2, and 4 kc/sec were presented for
60 sec, followed by 0, 15, 30, or 60 sec of silence before testing. Twenty audiometrically normal subjects
were used. At all frequencies, TTS was found to be a monotonic function of level if measured within 5 sec
after exposure. Recovery from TTS induced at 60 dB SL appeared faster than at 40 dB SL. Early recovery
rate increased directly with level of exposure when recovery was measured by continuous tracing, and,
with a few exceptions, the same relation was found when recovery was measured by testing after intervals
of silence. At 40 and 60 dB SL, TTS varied directly with frequency when measured immediately after exposure.
Continuous postexposure threshold testing significantly retarded recovery produced by all frequency-by-level
combinations. Thirty seconds after exposure, this retardation averaged about 3.5 dB. Recovery
after 15 sec of silence was approximately equivalent to recovery after 60 sec of continuous testing.

In recent years, a number of experiments have
investigated temporary shift of the auditory threshold
following exposure to pure frequencies and thermal
noise. Although most effort has been invested in investigation
of the effects of comparatively high-exposure
levels in the range of 80 to 120 dB above threshold (e.g.,
Davis, Morgan, Hawkins, Galambos, and Smith 13
Hood, 24 Ward, Glorig, and Sklar, 35 some experiments
have used lower levels. Studies using low-level pure
tones as the exposure stimuli have revealed that TTS
is produced by a sensation level as low as 10 dB if
measurement is made within a few seconds (15 sec)
following termination of exposure.

Reger and Lierle, 46 using an uninterrupted 1-kc/sec
tone for both exposure and threshold testing, found
little difference in TTS whether the 1-min exposure was
20 or 80 dB SL. Actually, they reported that, for the
20 subjects, exposure to the 20-dB SL appeared to
produce slightly more TTS than exposure to the 80-dB
SL. This was an unexpected observation; other studies,
using levels above 80 dB SL, had clearly shown that
TTS increases as level increases if exposure duration
and frequency were constant.

Hirsh and Bilger 57 reported that Hirsh and William
Burns, in an unpublished experiment using six subjects,
had repeated the conditions of the Reger and Lierle
experiment, and had found about the same amount of
TTS at 20 and 80 dB SL. They concluded that TTS is
not related to level if both exposure and threshold measurement
are at the same frequency and the exposure is
in the range of 10 to 80 dB SL. If the threshold test is at
one-half octave above the exposure frequency (e.g.,
exposure at 1 kc/sec and test at 1.4 kc/sec), TTS is then
clearly related to level. To test if TTS produced by
such low levels of exposure is real and not an artifact
of the testing procedure, Hirsh and Burns were also
reported to have exposed subjects to 1 min of silence,
after which interval they discovered no significant
shifts of threshold.151

The complex recovery curves, i.e., curves describing
return to pre-exposure sensitivity as a function of time,
found by Hirsh and Bilger led them to hypothesize two
recovery processes, which they call the “R-1” and
“R-2” processes. R-2 appears to be the basic recovery
process; it is relatively slow and stable, and the time
required to complete the process is clearly dependent
upon the level of exposure. They conjecture that the
R-2 process may be recovery from chemical depletion
of the receptor cells. When some minimum level of
exposure is exceeded, another process, termed the R-1
process, is activated. The R-1 process is characterized
by rapid recovery; the threshold becomes hypersensitive
45-60 sec following exposure, then returns to hyposensitivity
about 30 sec later. The R-1 process is most
noticeable at the exposure frequency. They suggest that
the R-1 process may be related to nerve excitability.
According to this view, recovery from TTS produced
by 20 dB SL, as in the Reger and Lierle study, 4 8would
appear to be pure R-2 process, while recovery from TTS
produced by 80 dB SL apparently would be a combination
of R-1 and R-2 processes. TTS after the 20-dB SL
exposure appears greater than TTS after the 80-dB SL
exposure, only if measurement is made during about the
first minute after termination of exposure.

Reger and Lierle tended to favor a different explanation
for their observations. They proposed that contraction
of the stapedius muscle might account for the lack
of difference in TTS following the two exposure conditions,
20 vs 80 dB SL. Stapedius reflex can attenuate
sounds entering the inner ear and this contraction can
be elicited by pure tones (Jepsen, 69 Klockhoff, 710 Kobrak, 811
Ward 912). Reger and Lierle hypothesized that 80 dB SL
is sufficient to cause reflex, but that 20 dB SL is not.
Therefore, the intensity of the exposure reaching the
inner ear may be reduced by the reflex in the case of
the 80-dB SL exposure, but not in the case of the 20-dB
SL exposure.

In both the Reger and Lierle 4 13and the Hirsh and
Bilger 514 studies, the first measurement of TTS was made
about 15 sec after termination of exposure. As pointed
out in the Hirsh and Bilger study, the recovery curves
for TTS's produced by various levels cross and recross
one another as time from termination of exposure increases.
By 15 sec following termination of exposure,
recovery curves have already crossed. Thus, the relative
effects of the different levels may depend upon
when in the course of recovery the threshold determination
is made.

Lawrence and Yantis 10 15found that a 60-dB SL, 1kc/sec
tone produced about 2.5 dB more TTS than a 20dB
SL tone of the same frequency when threshold was
measured at 1 kc 6 sec after cessation of exposure. They
found little difference between the effect of 60 and 80
dB SL. Their data were based on both ears of three
subjects.

If measurement is made soon enough following termination
of exposure, TTS might appear more clearly
related to level of exposure. In the present experiment, as
is shown, the technique and equipment permitted a
stable measurement of TTS immediately following termination
of exposure, as well as at various subsequent
times.

Along with the problem introduced by measuring
TTS after a considerable amount of recovery has already
taken place, another problem arises. In a pilot
study by Ward, 1116 it appeared as though recovery from
TTS produced by a tone of 1 kc/sec presented for 1 min
at 20 dB SL was slower if recovery was traced by 3 rain
of continuous testing following termination of the exposure,
than if 2 min of silence were interposed before
threshold testing began. Accordingly, it appears that
testing itself may retard the process of recovery, and
the present experiment was arranged for evaluation of
this possible effect also.

More recently, after the design and performance of
the present experiment, a study by Byers 1217 has been
published that also investigated the effect of postexposure
threshold testing on recovery. Byers used a
90-dB SL, 4-kc/sec tone of 3-min duration to produce
TTS. He found that TTS measured 6 min after exposure
cessation was significantly greater if there had been continuous
tracing during the 6 min than it was if 5 min of
silence were interposed before 1 min of tracing.

I. Method

Plan.

Plan. The effects of three variables on TTS and recovery
were studied. These variables were frequency
and level of exposure tone and duration of recovery (in
silence) before onset of postexposure threshold test tone.
The experiment employed a factorial plan, using frequencies
of 1, 2, and 4 kc/sec, sensation levels of 10, 20,
40, and 60 dB, and recovery periods of 0, 15, 30, and 60
sec. The 12 frequency-by-level combinations were presented
to each subject in a random order generated from
a table, a different order for each subject. Figure 1 shows
the four postexposure conditions. Condition 1 involved
70 sec of continuous tracing. TTS measurements were
made by selecting the value judged to be most representative
during the various 10-sec intervals. In Fig. 1,
the intervals measured are indicated by crosshatching.
152Condition 2 involved 15 sec of recovery (silence)
followed by 55 sec of tracing; condition 3, 30 sec of
recovery followed by 40 sec of tracing; condition 4, 60
sec of recovery followed by 10 sec of tracing. As shown,
the interval of measurement in conditions 2-4 was the
first 10 sec of tracing. Order of postexposure conditions
was assigned according to a systematic Latin square.
Subjects were divided into four groups for this purpose.
This order of recovery periods is listed in Table I. A
subject received all 12 frequency-by-level combinations
with one postexposure condition, then received the 12
combinations in the same order with the next scheduled
postexposure condition. With three frequencies, four
levels, and four recovery periods, each subject received
a total of 48 trials.

Subjects.

Subjects. The subjects were 20 male college students
from 17 to 25 years, with a mean age of 20.1 years.
Nineteen of the subjects showed no hearing loss (HL)
at 1, 2, or 4 kc/sec; one subject showed 3 dB HL at
2 kc/sec and 5 dB HL at 4 kc/sec. All subjects were paid.

Equipment.

Procedure.

Procedure. On the subject's first visit to the laboratory,
an audiogram of his hearing thresholds at 0.5, 1, 2,
3, 4, and 6 kc/sec was made, using a Rudmose automatic
audiometer (model ARJ-3) that had been calibrated
prior to the experiment. The subject was told that he
would be participating in an experiment that involved
listening to tones and tracing his threshold in a manner
similar to that required in making the audiogram.

The subject was then given training in the experimental
situation. The procedure of each trial was as
follows. The subject (1) traced his pre-exposure threshold;
(2) received a 1-min exposure to one of the twelve
stimulus combinations; (3) received one of the four
postexposure conditions described in connection with
Fig. 1. A signal lamp was used to alert the subject that
threshold tracing was required. During pre-exposure
threshold tracing, the lamp was on, and during exposure
and recovery it was off; during postexposure tracing it
was on again. The subject's button that controlled the
reversing of the recording attenuator motor was also
inactivated during exposure and recovery. Both the
signal lamp and the motor-reversing circuit were controlled
by the two timers. The subject was cautioned
against responding to the light itself; he was told that
light onset did not necessarily coincide with the tone
onset. He was also instructed not to attempt to trace his
threshold when the signal lamp was off.

II. Results and Discussion

Effects of Level and Frequency upon Initial TTS

To evaluate the relationship of frequency and level
of exposure stimulus to TTS, the initial measurement
period of postexposure condition 1, shown in Fig. 1, was
considered. As has been explained, the crosshatched
areas show the intervals of postexposure periods used
for threshold measurement. Each interval covered a
10-sec span. The point most representative of the threshold
during that period was used as the threshold
measurement. For the various frequency-by-level combinations,
means and standard deviations for the group
of 20 subjects were calculated, and these are given in
the first column of Table II. Figure 3 is a plot of
these means, which shows how TTS varied as a function
of exposure level with frequency the parameter.

As Fig. 3 shows, TTS was found to be a mono tonic
function of level of exposure stimulus over the range
from 10 to 60 dB SL at each of the three frequencies.
Furthermore, the effect of exposure level is seen to be
substantial. This finding of monotonic relationship appears
to conflict with the results obtained by Hirsh and
Burns and reported by Hirsh and Bilger. 519 The procedures
of the two experiments, however, are not
directly comparable. In the present study, the recording
attenuator changed attenuation at the rate of 8 dB/sec,
permitting a stable measurement usually within 2 sec
after exposure termination. (A stable measurement, as
denned by Hirsh and Bilger, meant that the stimulus
level crossed the subject's threshold from both below
and above the threshold at least once.) Because the
rate of attenuation change used in the experiments reported
by Hirsh and Bilger was 2 dB/sec, stable TTS
measurements were made 15 sec after exposure termination.

In order to examine systematically the apparent increase
in recovery rate with rising exposure levels, an
estimate of recovery rate was developed. For each of the
12 frequency-by-level of exposure combinations, the
measurement from condition 3 (30 sec of silence) was
subtracted from the initial measurement from condition
1. The third measurement sampled from condition 1
was also subtracted from the initial measurement of
that condition. These differences, then, were the amount
of recovery occurring in the first 30 sec. By multiplying
the differences by two, an expression in recovery per
minute was achieved. These expressions appear in
Table III. From them, it is evident that the rate of
recovery does increase monotonically as level of exposure
increases with the continuous-testing condition
and, except for the 1- and 4-kc/sec 40- and 60-dB SL
inversion, the same mono tonic relation holds when the
30 sec of silence is interposed. It can also be observed
that for all 12 exposure combinations the rate of recovery
is greater if the silence is interposed.

The results of this part of the experiment, then,
support the conclusion that TTS varies directly with
sensation level of exposure tone if measured immediately
after cessation of exposure, but TTS from higher levels
tends to recover faster so that more TTS is found after
40 than 60 dB SL when TTS is measured about 15 sec
after exposure cessation if the threshold tracing has been
continuous since the cessation.

Effects of Threshold Testing upon
Rate of Recovery from TTS

In order to answer the question of whether continuous
postexposure threshold testing retards recovery, the
two conditions after 30 sec had elapsed, since exposure
termination, were compared statistically. The group
means and standard deviations for these two sets of
data are shown in the columns headed In and 3 in
Table II. It was fell that this was the most meaningful
single interval because it is after the very fast recovery
of the first 15 sec and before the inevitable drawing
together of the curves as complete recovery is approached
in both conditions. The measurements following
30 sec of continuous tracing, i.e., the initial measurements
from postexposure condition 3, Fig. 1, were
compared with the third measurements from condition
1. In each comparison, the same frequency-by-level-of-exposure
combination was used. The difference in TTS
for each subject, resulting from the two conditions, was
used as an entry in the Wilcoxon matched-pairs signed-ranks
test. 1221 A test was made for each of the 12 exposure
combinations. The research hypothesis was that TTS
would be greater following the 30 sec of continuous
testing than following the 30 sec of silence. At 1 and 2
kc/sec, the null hypothesis could be rejected beyond
the 0.005 level at each of the four exposure levels; at
4 kc/sec, for 10-, 40-, and 60-dB SL exposures, it could
be rejected at or beyond the 0.05 level.

Figure 6 shows plots of the mean recovery curves for
the 20 subjects following each of the 12 exposure combinations.
The upper curve for each conbination is a
plot of the means obtained during the continuous-testing
condition, postexposure condition 1. The lower
curve is a plot of the means obtained during the initial
measurements of postexposure conditions 2 through 4.
It is apparent that the recovery curves drawn from
initial measurements made after the various periods of
silence are distinctly different from recovery curves
drawn in the usual manner, that is, recovery curves
made up of time samples from continuous testing. The
pairs of curves in Fig. 6 support the hypothesis that
postexposure threshold testing retards the recovery process,
and that this is the case for all levels, frequencies,
and postexposure times within the range of values
employed in the experiments. All data on frequencies
and levels were pooled, and the resultant pair of recovery
curves is shown in Fig. 7.

From Figs. 6 and 7, it appears that resting the ear
for 15 sec permits an amount of recovery from TTS
that is roughly equal to the recovery after 60 sec when
continuous testing is practiced. It also appears that the
major difference in rate of recovery between the two
postexposure conditions occurs during the first 15 sec.
It may be that this is the period during which the
effect of testing is relatively greatest. However, the
pilot study by Ward 1122 used much longer silent recovery
periods. He found that TTS, measured 3 min after
exposure termination, was less when 2 min of silence
were interposed. Also, Byers 1223 tested for a short period
immediately following exposure cessation to determine
initial TTS and still found that 5 min of silence produced
more recovery than 5 min of testing.

If the post exposure threshold testing in the present
study had involved a stimulus at pre-exposure level, i.e.,
the normal resting-threshold level, it seems probable
that the tone would not have had a retarding effect on
recovery. Lierle and Reger, 15 25in a study of the effect of
tone at threshold intensity, found that a subject with
normal hearing showed little or no change in sensitivity
over a test period of 20 min when tracing with a 2-kc/sec
tone. (However, subjects with recruitment and eighth-nerve
tumors showed a definite decrement in sensitivity.)

There are other differences between condition 1,
continuous testing, and conditions 2 through 4, that
interpose silence before testing begins. In condition 1,
the subject is involved in only one activity following
exposure, that of tracing. In conditions 2 through 4,
however, the subject is involved in two activities: first
resting, then tracing. The subject's orientation may be
different in condition 1 from what it is in conditions 2
through 4. Immediately following exposure (when the
subject begins tracing in condition 1), the subject's
criterion of threshold loudness may be altered by a
preservation of the sensation associated with the relatively
loud exposure. He may also experience tinnitus
following the higher-level exposures, which he tended
to confuse with the threshold stimulus. If the subject
responded to the tinnitus rather than the threshold
stimulus, his threshold would appear erroneously low,
but, if the subject did not respond until the threshold
stimulus is well above his threshold because he is unsure
of what he hears, or if the stimulus is masked by the
tinnitus, his threshold will appear to be erroneously
high.

Another possible result of the difference in activities
associated with condition 1 and conditions 2 through
4 may be a change in threshold criteria because of the
length of time during which tracing is required. This
may be especially the case for the two extreme conditions,
one involving 70 sec of tracing and the other involving
only 10 sec of tracing. Although the subjects
were continually instructed about the criteria to be used
during threshold tracing, they may have reinstructed
themselves. Ten seconds of tracing is an easy task, the
breath can be held, and the subjects may have been
more exacting in performing it; but 70 sec of tracing is
more tiring, the breath cannot be held, and they may
have been less exacting in performing it, using a higher-than-threshold
stimulus value because it is easier to.
trace.

Whatever the cause, the results of this part of
the study clearly shows that continuous postexposure
threshold testing retards measured recovery from TTS's
produced by tones within the frequency and level range
employed in this experiment.

III. Summary

Mean TTS measurements made on 20 subjects within
5 sec after termination of a 60-sec exposure interval
were found to vary directly with level as exposure levels
ranging from 10 to 60 dB SL were presented, exposure
and test frequencies both being at 1, 2, or 4 kc/sec.
Rate of recovery from TTS so induced increased directly
with SL (and with initial TTS) when the measurement
of rate was made from continuous postexposure tracing,
and the same relation was usually found when intervals
of silence were interposed between exposure stimulus
and testing. At 40 and 60 dB SL, TTS increased directly
with frequency from 1 to 4 kc/sec when measured
immediately after cessation of exposure, but at 10 and
20 dB no evidence of such relationship was found. Continuous
postexposure threshold testing consistently and
significantly retarded recovery from TTS. This retardation
averaged about 3.5 dB when measured 30 sec after
exposure.

Acknowledgments

We wish to acknowledge certain perceptive suggestions
by W. Dixon Ward that led to part of the research.
We are grateful also for the encouragement and support
afforded by Aram Glorig, and are appreciative of the
aid of Ronald Gallo and Hervey Stern, who assisted
with the conduct of parts of the experiment. The work
was carried out, in part, under grant No. OH 00053
from the National Institutes of Health, Public Health
Service, U. S. Department of Health, Education, and
Welfare.157