Content uploaded by Wesley Grantham
Author content
All content in this area was uploaded by Wesley Grantham on Jan 17, 2017
Content may be subject to copyright.
Temporal processing in the aging auditory system
Anne Strouse,a) Daniel H. Ashmead, Ralph N. Ohde, and D. Wesley Grantham
Division of Hearing and Speech Sciences, Vanderbilt University School of Medicine, Nashville,
Tennessee 37232
~Received 29 December 1997; accepted for publication 10 July 1998!
Measures of monaural temporal processing and binaural sensitivity were obtained from 12 young
~mean age526.1 years) and 12 elderly ~mean age570.9 years) adults with clinically normal hearing
~pure-tone thresholds <20 dB HL from 250 to 6000 Hz!. Monaural temporal processing was
measured by gap detection thresholds. Binaural sensitivity was measured by interaural time
difference ~ITD!thresholds. Gap and ITD thresholds were obtained at three sound levels ~4, 8, or
16 dB above individual threshold!. Subjects were also tested on two measures of speech perception,
a masking level difference ~MLD!task, and a syllable identification/discrimination task that
included phonemes varying in voice onset time ~VOT!. Elderly listeners displayed poorer monaural
temporal analysis ~higher gap detection thresholds!and poorer binaural processing ~higher ITD
thresholds!at all sound levels. There were significant interactions between age and sound level,
indicating that the age difference was larger at lower stimulus levels. Gap detection performance
was found to correlate significantly with performance on the ITD task for young, but not elderly
adult listeners. Elderly listeners also performed more poorly than younger listeners on both speech
measures; however, there was no significant correlation between psychoacoustic and speech
measures of temporal processing. Findings suggest that age-related factors other than peripheral
hearing loss contribute to temporal processing deficits of elderly listeners. © 1998 Acoustical
Society of America. @S0001-4966~98!05210-2#
PACS numbers: 43.66.Mk, 43.66.Pn, 43.66.Sr, 43.71.Lz @JWH#
INTRODUCTION
One of the factors identified in psychoacoustic experi-
ments as contributing to poor speech perception is the re-
duced temporal resolving power of the auditory system
~Dreschler and Plomp, 1985; Ginzel et al., 1982; Price and
Simon, 1984; Schneider, 1997; Tyler et al., 1982!. Process-
ing of temporal information may occur via monaural and/or
binaural inputs. Monaural processing refers to what happens
to the signal arriving at a single ear ~although this normally
occurs in parallel for the signals at both ears!. Binaural pro-
cessing refers to the analysis of the differences between sig-
nals arriving at the two ears. They operate at different time
scales ~a few milliseconds for monaural resolution versus
small fractions of a millisecond for binaural resolution!, and
monaural temporal processing is more involved in following
a speech signal whereas binaural processing contributes to
separating the signal from competing sounds. Consequently,
monaural and binaural aspects of temporal resolution may
each contribute uniquely to speech perception.
A. Monaural temporal processing
The most common way of investigating monaural tem-
poral processing is by means of gap detection, defined as the
ability to detect a brief period of silence between two test
signals. Numerous studies have reported that listeners with
hearing loss have larger gap detection thresholds ~Buus and
Florentine, 1985; Glasberg and Moore, 1989; Irwin and
McAuley, 1987; Moore and Glasberg, 1988; Moore et al.,
1989; Tyler et al., 1982!. Because most elderly listeners have
some degree of hearing loss, it is important to determine
whether changes in temporal processing occur independent
of peripheral hearing loss.
Several studies have attempted to control for the con-
founding effect of age-related hearing loss on gap detection
thresholds ~Moore et al., 1992; Schneider et al., 1994; Snell,
1997!. Moore et al. ~1992!measured thresholds for the de-
tection of temporal gaps in sinusoidal signals as a function of
frequency in elderly hearing-impaired subjects and elderly
subjects with ‘‘near-normal’’ hearing ~audiometric thresh-
olds <25 dB HL from 250 to 2000 Hz!. Results were com-
pared to previous data collected from young normally hear-
ing subjects ~Moore et al., 1993!, revealing that elderly
subjects with near-normal hearing had higher gap detection
thresholds than young subjects. Moore et al. ~1992!attrib-
uted this result to the inclusion in the elderly group of some
individuals who had large gap detection thresholds. Never-
theless, when they compared gap thresholds in elderly sub-
jects with near-normal hearing to those with hearing impair-
ment, they found no difference between the two groups.
Schneider et al. ~1994!reached a similar conclusion. In this
study, thresholds for detecting a gap between two Gaussian
modulated 2000-Hz tones were measured in young and eld-
erly listeners with pure-tone thresholds <25 dB HL from
250 to 3000 Hz. Gap detection thresholds were longer and
more variable for elderly listeners than for young listeners.
Snell ~1997!more rigorously controlled for high-frequency
hearing loss in elderly subjects, measuring gap detection
thresholds for noise-burst stimuli in young and elderly listen-
ers with pure-tone thresholds <20 dB HL from 250 to 4000
a!Current address: VA Medical Center, Audiology Service ~126!, Mountain
Home, TN 37684. Electronic mail: anne.strouse@med.va.gov
2385 2385J. Acoust. Soc. Am. 104 (4), October 1998 0001-4966/98/104(4)/2385/15/$15.00 © 1998 Acoustical Society of America
Hz. Again, gap thresholds were significantly larger in elderly
subjects. Thus the studies agree in that all found some eld-
erly individuals who exhibited losses in temporal resolution
that were unrelated to degree of hearing loss. Therefore it is
reasonable to consider factors other than peripheral hearing
loss that could account for age-related differences in monau-
ral temporal resolution.
Whether gap detection ability is important for accurate
perception of speech remains questionable. It is widely ac-
cepted that many speech events that are critical for speech
perception are of short duration ~Dorman et al., 1979; Pickett
and Decker, 1960; Repp et al., 1978!. For example, gap de-
tection ability has been linked to a listener’s ability to pro-
cess time events related to distinctions in speech between
voiced-voiceless cognates and various manners of syllable
transition ~DeFillippo and Snell, 1986!. It has also been dem-
onstrated that in adverse listening situations, segment dura-
tion may be an especially important cue to phonetic identity
~Wardrip-Fruin, 1982!. The relationship between gap detec-
tion and speech perception has been widely studied using
multivariate correlation analyses. Several investigations re-
veal significant correlations between gap detection and
speech recognition ability even when audiometric threshold
is factored out ~Dreschler and Plomp, 1985; Glasberg and
Moore, 1989; Tyler et al., 1982; Tyler and Summerfield,
1980!. Reduced gap detection ability has been associated
with poorer performance on speech in noise tests and dis-
crimination of syllables varying in duration of voice onset
time ~VOT!~Dreshler and Plomp, 1985; Tyler et al., 1982!.
Thus it is conceivable that perceptual problems with short
duration stimuli such as those observed in temporal gap ex-
periments may partially underlie the communication prob-
lems of the elderly. Other studies, however, find no signifi-
cant correlation between these factors ~Divenyi and Haupt,
1997; Festen and Plomp, 1983; Lutman and Clark, 1986; van
Rooij and Plomp, 1990!. Thus the relationship between gap
detection and speech perception has not been firmly estab-
lished.
The perception of temporal differences in speech can be
measured directly by examining categorical perception of
temporal aspects of speech such as duration of formant tran-
sitions, duration of frication, or differences in VOT. VOT,
defined as the temporal interval between the burst or release
of a stop consonant and the onset of periodic vibration of the
vocal cords, is one of the major parameters distinguishing
voiced and voiceless consonants ~Lisker and Abramson,
1970!. Several investigators have used categorical perception
to study temporal factors in speech perception. Some have
reported that neither age nor mild-to-moderate hearing im-
pairment affect a listener’s ability to make phonetic judg-
ments based on temporal stimulus properties ~Dorman et al.,
1985; Tyler et al., 1982!. On the other hand, Godfrey and
Millay ~1978!found that about half of their sample of
hearing-impaired listeners were unable to identify stimuli
which varied in the duration of formant transitions. Price and
Simon ~1984!found that older listeners ~who had good hear-
ing for their age, but higher pure-tone thresholds than
younger listeners!required longer silence durations to report
hearing differences in stop consonants. Age-related difficul-
ties in identification have also been reported for stimuli in
which vowel duration, voice onset time, and fricative noise
duration were manipulated ~Ginzel et al., 1982!.
B. Binaural temporal processing
Older adults with peripheral hearing loss show reduced
performance on such binaural tasks as the masking-level dif-
ference ~Findlay and Schuchman, 1976; Olsen et al., 1976!,
interaural time discrimination ~Herman et al., 1977; Kirikae,
1969; Matzker and Springborn, 1958!, and the precedence
effect ~Cranford et al., 1990!. However, little age-related
change in binaural processing has been reported by other
researchers ~Kelly-Ballweber and Dobie, 1984; Palva and
Jokinen, 1970!.
Much of the psychoacoustic literature directed toward
the effects of aging on binaural processing has focused on
sensitivity to interaural time differences ~ITD!~Herman
et al., 1977; Kirikae, 1969; Matzker and Springborn, 1958!.
These studies have varied widely in method, age groups
tested, and stimuli, however, a consistent finding has been
that younger adults have lower thresholds for ITDs than
older adults. Little attempt was made, however, to control for
the effects of hearing loss on ITD thresholds despite use of
experimental stimuli containing higher-frequency spectra
such as clicks and wideband noise.
Herman et al. ~1977!recognized the possible confound-
ing effects of high-frequency hearing loss on ITD thresholds
and thus attempted to control for such an effect by testing
individuals with normal pure-tone thresholds below 2000
Hz. Although all subjects in this study had normal hearing
below 2000 Hz, older subjects had substantially greater hear-
ing loss than young subjects at higher frequencies. Pure-tone
loss in the elderly group increased from 4.38 dB at 2000 Hz
to 44.38 dB at 4000 Hz, while loss for young subjects in-
creased from 24.0 dB at 2000 Hz to only 1.56 dB at 4000
Hz. Thus it is not clear whether the reported age-related loss
in the ability to lateralize the source of a sound on the basis
of interaural time delay observed in elderly individuals was
caused by aging, per se, or by the high-frequency peripheral
hearing loss. Hearing loss in younger individuals has been
shown to adversely affect binaural abilities ~Durlach et al.,
1981; Hausler et al., 1983!.
To the extent that there are age-related changes in bin-
aural processing, older people might be expected to have
problems in situations where binaural hearing is useful, such
as sound localization and comprehending speech in noisy
settings. Although age-related changes in speech perception
have not been tightly linked to binaural processing deficits,
older adults tend to do worse than younger adults at various
dichotic listening tasks ~Kelly-Ballweber and Dobie, 1984;
Martin and Cranford, 1991; Roush, 1985!.
A common measure of binaural processing for speech is
the masking level difference ~MLD!. The MLD is a phenom-
enon resulting in improved binaural hearing sensitivity when
a phase reversal is imposed on either a primary signal or its
masking noise. Several theories have been proposed to ac-
count for this phenomenon, but all are based on binaural
analysis of interaural differences in time and intensity for
dichotically presented stimuli ~Moore, 1982!. Several studies
2386 2386J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
have indicated that elderly subjects have reduced MLDs
compared to younger subjects ~Findlay and Schuchman,
1976; Tillman et al., 1973; Warren et al., 1978!. In a recent
investigation, Grose et al. ~1994!compared MLDs for
speech in a group of elderly listeners with normal hearing
through 2000 Hz to those obtained from a group of young
normal-hearing listeners. Results again showed that elderly
subjects performed more poorly than the young listeners. Al-
though these findings may be due, in part, to age differences,
peripheral hearing loss in the older subjects may have ac-
counted for their smaller MLDs compared with those of the
young normal-hearing groups, since it has been established
that presence of peripheral hearing loss significantly reduces
the MLD ~Jerger et al., 1984!. Nonetheless, several studies
have examined the MLD, accounting for peripheral effects,
and have found statistically significant differences between
young and elderly subjects, suggesting further decline in bin-
aural processing with advancing age ~Olsen et al., 1976;
Pichora-Fuller and Schneider, 1991!.
To determine whether losses in temporal resolution are
attributed to age-related factors other than sensorineural
hearing loss, temporal resolution must be measured in an
elderly population with good hearing sensitivity. Findings of
previous studies may have been confounded by age-related
hearing loss since high-frequency audiometric thresholds
were not included in selection criteria of older subjects with
normal hearing. For the present investigation, we were in a
unique position in that elderly subjects available for study
had audiometric thresholds <20 dB HL from 250 to 6000
Hz. By more rigorously controlling for degree of peripheral
hearing loss over a wider range of audiometric frequencies,
any changes in temporal processing ability could be more
strongly attributed to factors associated with aging. Thus for
the present sample, if temporal processing is normal as long
as hearing is normal, regardless of age, this would mean that
reported decreases in temporal processing ability in older
listeners are primarily the result of reduced hearing sensitiv-
ity. On the other hand, if temporal processing is ‘‘abnormal’’
in older subjects even when hearing is better than 20 dB HL,
this would imply that there exist age-induced alterations in
auditory structures and/or processes which are not detected
by conventional pure-tone measures.
I. METHOD
A. Subjects
Two groups of subjects were tested, 12 normally hearing
young adults, aged 20–30 years (mean526.1, range 22–30!
and 12 normally hearing elderly adults aged 65–75 years
~mean570.9, range 66–75!. The groups were matched for
gender ~ten female and two male subjects!and hearing sen-
sitivity. Normal hearing was defined as 20 dB HL or better
pure-tone thresholds for the frequencies 250–6000 Hz bilat-
erally. At 8000 Hz, close matching of young and elderly
subjects could not be achieved despite extensive audiometric
screening of potential subjects. Regarding the interpretation
of ‘‘normal hearing,’’ some researchers have used published
age-relative norms, whereas others have selected subjects
who meet criteria established for young adults. We used the
latter approach, based on evidence that reduced temporal
processing ability is associated with hearing impairment, in-
dependent of age ~Florentine and Buus, 1984; Moore and
Glasberg, 1988; Moore et al., 1989!. The present study in-
cluded older subjects with pure-tone thresholds as close as
possible to those of a control group of young subjects, to
help ensure that any resulting age differences would not be
attributable to peripheral hearing loss, although this means
that the older subjects had better hearing than is normal for
their age. The mean pure-tone audiometric thresholds ~in dB
HL re: ANSI, 1969!for each group are shown in Table I.
Hearing thresholds for all subjects were symmetrical ~inter-
aural differences <10 dB at each frequency!, with no con-
ductive component. There was no significant difference in
pure-tone thresholds between groups at any frequency. In an
analysis of variance there were no significant main effects or
interactions involving age @F(1,22)53.72; p50.07#, ear
@F(1,22)50.93; p50.34#, or tone frequency @F(1,22)
50.30; p50.143#.
B. Procedures
Four tasks were presented to each subject. Monaural
temporal processing was measured using a gap detection
paradigm. Binaural sensitivity was measured by interaural
time difference ~ITD!thresholds. Data were collected from
the two groups at three different presentation levels ~4, 8,
and 16 dB!, defined with respect to each individual’s thresh-
old for detecting the test stimulus. Low presentation levels
were chosen based on evidence that the temporal resolution
of the auditory system has been found to worsen somewhat
at low sound levels for all types of stimuli ~Gregory, 1974!.
Thresholds for the detection of a temporal gap in a noise
stimulus increase when the level per critical band is less than
40 to 50 dB above the absolute threshold ~Buus and Floren-
tine, 1985; Fitzgibbons, 1983; Shailer and Moore, 1983!.
Further evidence is consistent with a worsening of temporal
resolution at low sound levels. For example, the rate of re-
covery from forward masking decreases as the sensation
level of the masker is reduced ~Jesteadt et al., 1982; Moore
and Glasberg, 1983!. The ability to detect amplitude modu-
lation at various modulation rates worsens at low sound lev-
els, again indicating reduced temporal resolution at low lev-
TABLE I. Mean audiometric thresholds ~dB HL!and standard deviations
~dB!for young and elderly listeners.
Frequency ~kHz!
0.25 0.5 12346
Young subjects (N512)
Mean—Right ear 7.9 7.5 7.9 8.3 10.4 11.2 13.8
~4.9!~4.5!~3.9!~4.4!~3.9!~4.3!~4.3!
Mean—Left ear 7.1 5.0 5.8 8.8 9.6 12.9 13.3
~4.9!~5.2!~4.2!~2.3!~2.6!~2.6!~3.9!
Elderly subjects (N512)
Mean—Right ear 9.6 10.8 10.0 10.4 12.5 14.1 15.8
~2.6!~3.6!~6.0!~5.8!~3.9!~3.6!~5.1!
Mean—Left ear 7.1 5.8 7.1 10.4 11.6 14.6 16.3
~3.9!~3.6!~3.3!~3.9!~4.4!~3.9!~4.8!
2387 2387J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
els ~Bacon and Viemeister, 1985!. Although no published
data are available, researchers have examined age differences
in sensitivity to interaural time differences as a function of
sound level ~Ashmead et al., unpublished data!. They tested
young and older adults, obtaining ITD thresholds at each of
three overall sound levels ~8, 16, and 30 dB SL relative to
individual detection thresholds!. Findings revealed that the
difference between age groups was greater at low stimulus
levels.
Each subject was also evaluated on two measures of
speech perception. One of these was the masking level dif-
ference ~MLD!, which examined the ability to utilize binau-
ral cues including interaural temporal differences to recog-
nize speech. The second test was a syllable identification
task that included phonemes varying in voice onset time
~VOT!. All testing took place in a double-walled sound
booth using TDH49P headphones, where subjects were
seated in front of a computer. The computer monitor pro-
vided visual feedback during testing and subject responses
were entered on the computer keyboard ~with the exception
of the responses for the MLD task, which were repeated
aloud!. The order of presentation for the four tasks was ran-
domized across subjects. Subjects were tested in three ses-
sions, each lasting 1–1.5 h, and were paid for their partici-
pation.
C. Gap detection
1. Stimuli
Stimuli for the gap detection task were computer-
generated 1000-Hz sinusoidal signals presented in a continu-
ous background noise with a spectral notch at the frequency
of the sinusoid. Signals were generated using a Dell XPSP90
laboratory computer and Tucker-Davis Technologies ~TDT!
Psychoacoustic System, with 16-bit digital-to-analog ~D/A!
converters, a 10-kHz sampling rate ~low-pass filtered at 4
kHz to prevent aliasing!, and a rise–fall time of 5 ms. The
overall duration of each observation interval was 200 ms. To
determine the duration of the signal on either side of the gap,
the gap duration was subtracted from 200 ms and the result
divided by two. The duration of the signal preceding and
following the gap was then rounded to the nearest ms. The
gap started with the signal at a positive-going zero crossing
and ended with the phase needed for a given gap duration.
The 1000-Hz signal was presented in a continuous back-
ground noise, with a sharp notch at the signal frequency,
designed to mask the spectral splatter associated with the
abrupt gating of the gap. The noise masker was a 65 dB SPL
~22-dB spectrum level!Gaussian noise with a notch arith-
metically centered at the frequency of the test signal ~fc!; the
width of the notch was 0.4fc at the 3-dB down points. The
depth of the notch was 37 dB. The noise was recorded on
digital audio tape ~DAT!, and played back through a Pana-
sonic SV3700 DAT deck that led to a programmable attenu-
ator. The signal and noise were combined in a weighted sig-
nal mixer and passed through a headphone buffer to the
TDH49P headphones.
2. Procedure
Gap detection thresholds were measured using an adap-
tive two-interval forced choice ~2IFC!procedure. On each
trial, two 200-ms signals were presented, separated by 1000
ms. One of the two sounds was the ‘‘signal’’ ~containing the
gap!; the other sound was the ‘‘standard’’ ~no gap!. The
order, either signal-standard or standard-signal, was chosen
randomly with equal probability. Following stimulus presen-
tation, the subject chose the interval that contained the signal
by pressing the appropriate button on the computer key-
board. Gap duration on the first trial was 50 ms so that the
stimulus gap was easily detectable. On subsequent trials, gap
duration was increased by a factor of &, then rounded to the
nearest ms after each incorrect response and decreased by a
factor of &after every two successive correct responses.
This two-down, one-up algorithm estimates the 71% point on
the psychometric function ~Levitt, 1971!. In order to ensure
that subjects remained attentive to the appropriate cue, gap
duration was increased to the starting level of 50 ms follow-
ing four incorrect responses and remained at 50 ms until the
subject responded correctly. Following a correct response,
testing was resumed at the gap duration being evaluated be-
fore the jump to 50 ms occurred. Trials measured at 50 ms
were not considered in gap threshold calculations. Testing
continued until ten reversals occurred and gap threshold was
estimated as the geometric mean of the gap durations at the
last eight reversals.
Before testing began, a practice run was administered at
a level of 30 dB relative to the masked threshold for a
1000-Hz continuous tone1to ensure that subjects were famil-
iar with the task. Following the practice run, each subject
contributed three gap threshold estimates at levels of 4, 8,
and 16 dB relative to masked threshold.
Stimuli were presented monaurally to the ear with the
least pure-tone threshold loss. If there was no difference be-
tween ears, the test ear was chosen randomly. The order of
presentation level was randomized across subjects. The mean
of the three gap threshold estimates obtained at each inten-
sity level was taken as the gap threshold value.
D. Interaural time differences „ITD…
1. Stimuli
Stimuli for the ITD task were 400-ms-long trains of 40
rectangular clicks, each click lasting 50
m
s with an interclick
interval of 10 ms. Signals were generated using the same
equipment as for the gap detection task, but with a 200-kHz
sampling rate. Interaural time differences were implemented
by delivering the signal to one ear by an integer multiple of
the sampling period ~5
m
s!.
2. Procedure
Thresholds for interaural time differences were mea-
sured using an adaptive 2IFC procedure. On each trial, two
successive 400-ms click trains were presented. The first click
train was presented with a 0-
m
s ITD, thus presented to both
ears simultaneously. For the second click train in the series,
the desired ITD was created by offsetting the clicks to the
two channels in 5-
m
s intervals. Thus the auditory impression
2388 2388J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
was of a sound image that occurred first at midline and then
to the side ~left or right, depending on the trial!. The ear
receiving the delayed click train was chosen randomly with
equal probability. Following stimulus presentation, the sub-
ject indicated whether the sound image was perceived to the
left or to the right by pressing the appropriate button on the
computer keyboard. The ITD on the first trial was 100
m
s.
On subsequent trials, step size was 20
m
s through the first
reversal and 5
m
s thereafter. In order to encourage subjects to
remain attentive to the appropriate cue, ITD duration was
increased to a level of 400
m
s following four incorrect re-
sponses and remained at 400
m
s until the subject responded
correctly. Following a correct response, testing was resumed
at the ITD being evaluated before the increase occurred. Tri-
als measured at 400
m
s were not considered in ITD threshold
calculations. Testing continued until ten reversals occurred
and ITD threshold was estimated as the geometric mean of
the ITD from the last eight reversals.
Initially, a practice run was administered at a level of 30
dB relative to the binaural click threshold.2During testing,
each subject contributed three ITD threshold estimates at
levels of 4, 8, and 16 dB relative to binaural click detection
threshold. The order of the presentation level was random-
ized across subjects. The mean of the three ITD threshold
estimates obtained at each intensity level was taken as the
ITD threshold value.
E. Voice onset time „VOT…
1. Stimuli
VOT was examined by creating a sound continuum
which varied in the duration of VOT in small steps. A con-
tinuum of consonant–vowel ~CV!syllables ranging from /ba/
to /pa/ was created using the Computerized Speech Research
Environment ~CSRE!cascade/parallel synthesis program
modeled after Klatt ~1980!~AVAAZ Innovations, Inc.,
1994!and a laboratory computer. Acoustically, the stimuli
differed only in VOT. VOT ranged from 0 to 60 ms in 10-ms
steps, creating a continuum of seven stimuli. A /ba/ syllable
with a duration of 300 ms was synthesized at a sampling rate
of 20 kHz. Fundamental frequency (F0) contour began at
103 Hz and rose to 125 Hz over the duration of the syllable.
During a 40-ms transition period, the three lowest formant
onset frequencies moved until they reached appropriate
steady-state values ~Blumstein and Stevens, 1980!. For gen-
eration of the six additional members of the continuum, VOT
variations were accomplished by altering the onset of voic-
ing and the duration of aspiration in 10-ms steps following
the initial burst.
2. Procedure
Listeners identified and discriminated the /ba/–/pa/ con-
tinuum. Stimuli for both tasks were presented to the subject’s
better ear at an individually determined most comfortable
loudness level. Subjects listened to conversational speech
through headphones and adjusted the level of the speech us-
ing the computer keyboard to their preferred listening level.
Because young and elderly subjects had similar hearing sen-
sitivity, presentation levels did not vary significantly be-
tween groups @F(1,23)50.362; p.0.05#. Presentation lev-
els ranged from 68 to 86 dB SPL.
For the identification task, subjects initially responded to
the series of seven stimuli representing the ordered con-
tinuum from /ba/ to /pa/. This set was then repeated to famil-
iarize the listener with the range of stimuli involved. Follow-
ing familiarization, stimuli were presented randomly. A set
of ten stimulus blocks ~70 trials!was used to generate iden-
tification functions. Data were collected using a single-
interval forced-choice paradigm. After presentation of each
stimulus, subjects identified the initial consonant of each syl-
lable as either /b/ or /p/.
The discrimination task involved one-step and two-step
presentations of a pair of stimuli in an AX ~same–different!
format. The stimulus pairs contained either two identical
stimuli ~i.e., ‘‘catch’’ trials!, or two stimuli which differed in
VOT. In the one-step condition, pairs of CV stimuli that
differed by 10-ms VOT were presented, resulting in six pairs
of experimental trials. Pairs were each presented 20 times, in
random order, with 500-ms interstimulus intervals. Inter-
mixed with the 120 experimental trials were 120 catch trials.
The 240 stimuli were arranged in 20 blocks, each containing
12 randomized stimulus pairs. In the two-step condition,
pairs of experimental stimuli that differed by 20-ms VOT
were presented, resulting in five pairs which were each pre-
sented 20 times. Intermixed with the 100 experimental trials
were 100 catch trials. The 200 stimuli were arranged in 20
blocks, each containing ten randomized stimulus pairs. After
presentation of each stimulus pair in both one- and two-step
tasks, the subject indicated whether the stimuli were the
‘‘same’’ or ‘‘different.’’
F. Masking level difference „MLD…
1. Stimuli
The MLD was measured using speech as the stimulus
and was determined by presenting a continuous speech noise
binaurally and in-phase (N0), then determining the speech
reception threshold when the binaural speech signal was pre-
sented interaurally in-phase (S0) and interaurally out-of-
phase 180° (S
p
). The threshold difference ~in decibels!be-
tween the two masking conditions (S0N0minus S
p
N0)
defined the MLD.
Both the speech signal and masker were presented
through a standard two-channel clinical audiometer ~GSI-10!
which has a network allowing phase reversal of either the
noise or test signal. The stimuli for the speech MLD were the
36 spondaic words of the CID W-1 list. The spondaic words,
and their associated calibration tone, were presented using a
compact disc recording ~Department of Veterans Affairs,
1991!played on a JVC XL-V161 compact disc player routed
through the audiometer. The recording consisted of two ran-
domizations of the CID W-1 word list for a total of 72 spon-
daic words separated by 4-s interstimulus intervals. The
masker for the speech MLD was a broadband noise with
equal energy per Hz from 250 to 1000 Hz with 12 dB/oct
rolloff from 1000 to 6000 Hz. During the experiment, the
masker was presented continuously at an overall level of 65
dB SPL. Presentation levels of both the speech signal and
2389 2389J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
masker were verified using sound level measurements
~Larson-Davis Laboratories, model 800B!.
2. Procedure
To familiarize listeners with the test materials, a written
list of the CID W-1 spondaic words was provided for sub-
jects to read before testing began. Each listener was then
presented with a randomization of the stimuli at a comfort-
able listening level in the absence of masking noise. To ob-
tain the MLD, threshold was determined for spondee words
in the diotic (S0N0) and dichotic (S
p
N0) conditions in the
presence of a continuous masking noise. To determine
threshold, a one-up, one-down adaptive procedure was used.
The first word of each test sequence was presented at 85 dB
SPL, representing a level that was 20 dB greater than the
level of the continuous noise masker. The subject reported
aloud the perceived word. Following each correct response,
the level of the speech was attenuated in 5-dB steps until an
incorrect response was recorded. Thereafter, the level of the
signal was decreased in 2-dB steps following a correct re-
sponse and increased in 2-dB steps following an incorrect
response. Testing continued until ten reversals occurred, and
the mean of the final eight reversal levels was taken as an
estimate of threshold. Once thresholds for both diotic and
dichotic conditions were obtained, the dB difference between
the two conditions was recorded as the speech MLD.
II. RESULTS AND DISCUSSION
Results of the present study indicated significant age dif-
ferences on measures of gap detection, ITD thresholds, slope
of the VOT identification function, discrimination of VOT
cues, and MLD thresholds. Thus elderly subjects performed
more poorly than their younger counterparts on these tasks,
even though all subjects had normal-hearing sensitivity.
A. Gap detection
Figure 1 shows mean gap detection thresholds obtained
at the 4-, 8-, and 16-dB sound levels ~relative to individual
masked thresholds!. Data were examined using an analysis
of variance ~ANOVA!with sound level as the within-
subjects factor and age group as the between-subjects factor.
Results indicated significant effects of age group @F(1,22)
523.57; p,0.0001#, sound level @F(2,44)527.13, p
,0.0001#, and the interaction of age group and sound level
@F(2,44)514.05, p,0.0001#. As shown in Fig. 1, the differ-
ence in performance between young and elderly adults was
especially large at very low sound levels. Performance was
significantly different between age groups at all three sound
levels ~4 dB: @F(1,22)514.49, p,0.001#; 8 dB: @F(1,22)
56.71, p,0.05#;16dB:@F(1,22)521.45, p,0.0001#!.
To further assess the relationship between gap detection
performance and signal level, a linear regression line was fit
for each individual subject ~linear fits were used as they
proved as good as linear plus quadratic fits during regression
analysis!, and slopes were averaged across subjects within
each group. The mean slope values based on individual best-
fit linear regression lines were significantly @F(1,22)
515.33, p,0.001#different for young adults ~20.44!as
compared to those for the elderly group ~22.45!.
Present findings are in agreement with earlier studies
using sinusoidal stimuli which reported larger gap detection
thresholds in elderly listeners with minimal hearing loss
~Moore et al., 1992; Schneider et al., 1994!. In fact, gap de-
tection thresholds for young and elderly adult subjects mea-
sured at the highest presentation level in the present study
~2.8 and 6.7 ms for young and elderly subjects, respectively!
are quite similar to those reported by Schneider et al. ~1994!
(young53.8 ms; elderly56.2 ms) using suprathreshold
stimuli. Collectively, findings of Moore et al. ~1992!,
Schneider et al. ~1994!, and the present investigation are
similar in that each identified some elderly individuals who
exhibited deficits in temporal processing that were unrelated
to hearing loss. Snell ~1997!also observed larger thresholds
for detection of gaps using noise-burst stimuli for elderly
subjects with audiometric thresholds similar to those ob-
tained for the present study.
Previous studies reported temporal processing ability as
measured by the ability to detect gaps in sinusoids at rela-
tively high sound levels. Moore et al. ~1992!obtained gap
detection thresholds at 25, 40, 55, 70, or 80 dB SPL and
subjects were tested only at levels for which the signal was
clearly audible. Schneider et al. ~1994!evaluated three
young control subjects at 10, 20, 40, and 60 dB SL, but
elderly listeners were evaluated at a single high-intensity sig-
nal level. Thus the only information reported regarding per-
formance of elderly subjects on gap detection tasks was mea-
sured well above audiometric threshold.
Schneider et al. ~1994!did not find an effect of level on
gap detection thresholds over most of the range of levels
tested in their three young control subjects ~with the excep-
tion of 10 dB SL!. Therefore they concluded that it was not
likely that the poorer gap detection performance of the older
subjects was due to the effect of sensation level. Their use of
Gaussian-enveloped tones minimized the likelihood that
spectral differences contributed to the ability to detect tem-
poral gaps. They showed that spectral differences within
notched noise were much larger for gaps in continuous noise
~used in the present study!than for gaps between two
FIG. 1. Mean gap detection thresholds for young and elderly adult listeners
at the 4-, 8-, and 16-dB sound levels presented relative to individual masked
thresholds. Error bars indicate 6standard error.
2390 2390J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
Gaussian-enveloped tones. Thus it is possible that the in-
crease in gap detection thresholds at lower presentation lev-
els may have been influenced by spectral differences. Since
all subjects in the present study had hearing within normal
limits, however, we would expect the contribution of spectral
cues, if any, to have a similar effect on gap threshold for all
listeners regardless of age. Thus the large differences be-
tween groups at low sensation levels cannot be entirely ex-
plained by spectral cues. It is possible that the effect of level
is only apparent closer to audiometric threshold, as even two
of three young control subjects in the Schneider et al. ~1994!
study had elevated gap detection thresholds at the 10-dB
sensation level ~compared to the 20-, 40-, and 60-dB levels
in the same study!. Since stimulus levels in the present study
were defined with respect to individual thresholds, and since
the groups had comparable pure-tone thresholds, the age dif-
ferences cannot be attributed to overall hearing sensitivity.
Rather, results suggest that there is a general tendency for
decreased performance on the gap detection task at low
sound levels.
B. Voice onset time
Figure 2 shows mean identification functions for the
/ba/–/pa/ syllable series. All listeners in both groups identi-
fied two clear phoneme categories. Phoneme intercept values
~phonetic boundaries!along the identification functions were
determined for each subject by performing a linear regres-
sion across z-transformed identification scores for stimuli on
the continuum. The mean phonetic boundary for the young
adult group was 27 ms, which is in agreement with previous
findings for normal-hearing young adults ~Lisker and
Abramson, 1970; Tyler et al., 1982!. For the elderly adult
group, the mean phonetic boundary was 32 ms. ANOVA on
the data revealed that there was no significant difference in
mean phonetic boundary between young and elderly listeners
@F(1,22)52.41, p.0.05#, indicating that the two groups
used a similar criterion for identifying VOT. The mean slope
of the boundary was also determined for each subject using
linear regression. The mean slope value was 21.41 for the
young adult group and 20.93 for the elderly adult group.
This identification slope for the younger group was signifi-
cantly steeper than that of the elderly group @F(1,22)
54.30, p,0.05#.
In Fig. 3, the open symbols show mean one-step and
two-step discrimination functions for the /ba/–/pa/ con-
tinuum from each subject group. Predicted discrimination
functions ~shown by filled symbols!were derived from the
identification responses using the procedures described in
Pollack and Pisoni ~1971!, which are based on the assump-
tion that subjects rely on phonetic labels in the discrimina-
tion task ~Liberman et al., 1967!. The peak of the predicted
curve depends on the position and slope of the phoneme
boundary in the identification function. Predicted scores
were calculated for each subject using the following formula:
Pc~a,b!50.5@~12Pa1!21~12Pb2!2#1Pa1Pb2,
where Pc(a,b)5predicted percent correct for any pair of
stimuli along a continuum. Pa15probability that the first
stimulus in the pair is identified as a member of a given
phoneme category. Pb25probability that the first stimulus in
the pair is identified as a member of another phoneme cat-
egory. The phonetic model is based on the assumption that
sounds are coded as one phoneme or another. Thus this
model indicates the classical categorical perception phenom-
enon. That is, sounds that are of different categories should
be discriminable whereas sounds in the same category
should not be discriminable. When both stimuli involved in a
comparison come from one phoneme category, the predicted
percent correct discrimination scores are near chance ~50%!.
When the pair of stimuli come from two different phoneme
categories, the predicted score is greater than chance, with
the magnitude of the score dependent on consistency of the
identification performance.
Figure 3 ~top!displays mean one-step discrimination
scores for young and elderly adult listeners. Data were ex-
amined using an ANOVA with score type ~obtained versus
predicted!and stimulus pair as within-subjects factors and
age group as the between-subjects factor. Results revealed a
significant effect of age group @F(1,22)523.1, p,0.0001#,
stimulus pair @F(5,110512.76, p,0.0001#, and a significant
interaction between score type and stimulus pair @F(5,110
FIG. 2. Mean identification functions including standard error values for
young and elderly listeners for the seven-step /"Ä/–/!Ä/ continuum.
2391 2391J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
53.8, p,0.01#. There was a significant interaction between
age group and score type @F(1,22)517.15, p,0.0001#, in-
dicating that the relationship between obtained and predicted
scores differed between groups. All listeners exhibited im-
proved discrimination performance at or near the phoneme
boundary. An obvious difference is that obtained scores were
higher than predicted scores for young adult subjects but not
for the elderly subjects. As observed by others ~Stevens
et al., 1969; Sharf et al., 1988!, actual discrimination perfor-
mance for the young adult group exceeded that predicted by
identification data on the one-step discrimination function.
For elderly listeners, discrimination performance was poorer
than predicted by identification data. To interpret the inter-
action effect of age group3score type, the simple main effect
of age group at each of the two conditions was assessed.
Comparisons revealed a significant difference between
groups for both predicted and obtained scores at the p
,0.01 level.
Figure 3 ~bottom!shows two-step discrimination scores
for young and elderly adult listeners. There was a significant
effect of age group @F(1,22)516.12, p,0.001#, stimulus
pair @F(4,88)524.36; p,0.0001#, and a significant interac-
tion between score type and stimulus pair @F(4,88)52.73,
p,0.05#. The interaction between age group and score type
was also significant @F(1,22)512.94, p,0.01#. All listeners
exhibited improved discrimination performance at or near
the phoneme boundary. Actual discrimination performance
for the young group exceeded that predicted by identification
data for all stimulus pairs. For the elderly group, although
performance increased near the phoneme boundary, overall
discrimination ability was lower than predicted values for all
stimulus pairs. Post hoc comparisons using simple effects
testing revealed a significant difference between groups for
both predicted and obtained scores (p,0.01).
Results of the present study indicate that elderly listen-
ers have reduced sensitivity to differences in VOT as com-
pared to younger listeners. Identification performance re-
vealed a more gradually sloping phonetic boundary,
indicating that elderly subjects were less able to clearly dis-
tinguish phoneme boundary categories. Discrimination data
revealed that actual performance on both one-step and two-
step discrimination tasks followed a similar pattern as pre-
dicted scores for both groups; the percent correct discrimina-
tion scores increased at areas near the phoneme boundary.
Actual percentage scores of elderly listeners, however, were
lower than what was predicted based on identification data,
again indicating decreased ability to distinguish the temporal
cue of VOT. The fact that young and old listeners performed
differently as a function of predicted and observed discrimi-
nation indicates that the categorical model performed differ-
ently for these populations. It is likely that the difference in
the populations relates to their ability to use auditory cues.
Young adults appear to utilize temporal auditory cues within
sound categories as evidenced by higher obtained than pre-
dicted scores. Older adults are limited in this auditory ability
as reflected in poorer obtained than predicted scores across
stimulus pairs. This may also indicate that the younger sub-
jects might have been using acoustic as well as phonetic
information in forming their discrimination judgments. In the
future, one way to assess the auditory component would be
to examine predicted functions using the dual-coding model
that includes both an auditory and phonetic component
~Fujisaki and Kawashima, 1970!.
FIG. 3. Mean discrimination scores for the one- and two-step discrimination
conditions for young and elderly adult listeners.
2392 2392J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
C. Interaural time differences
Figure 4 shows mean ITD thresholds for young and eld-
erly adults at the 4-, 8-, and 16-dB sound levels. As in the
gap experiment, the ITD thresholds of younger adults were
lower than those of older adults at all presentation levels.
Data were examined using an ANOVA with sound level as
the within-subjects factor and age group as the between-
subjects factor. Results indicated significant effects of age
group @F(1,22)519.06, p,0.0001#, sound level @F(2,44)
597.9, p,0.0001#, and the interaction of age group and
sound level @F(2,44)59.52, p,0.0001#. Performance was
significantly different between groups at the 16 dB
@F(1,22)59.18, p,0.0001#,8dB@F(1,22)516.98, p
,0.0001#,and4dB@F(1,22)522.59, p,0.0001#sound
levels. The mean slope values based on individual best-fit
linear regression lines were significantly @F(1,22)524.22,
p,0.01#different for young adults ~23.55!as compared to
those for the elderly group ~26.78!. Results for young adult
listeners are in agreement with previous findings reporting
increased thresholds for detecting interaural phase differ-
ences at low sensation levels ~Zwislocki and Feldman,
1956!.
Previous studies have examined the effects of aging on
binaural hearing by measuring sensitivity to ITDs ~Herman
et al., 1977; Kirikae, 1969; Matzker and Springborn, 1958!,
however, little attempt was made in these studies to ad-
equately control for the effects of hearing loss on ITD thresh-
olds despite the use of experimental stimuli containing
higher-frequency spectra such as clicks and wideband noise.
The present data for elderly subjects with normal-hearing
sensitivity demonstrate a clear age-related loss in the ability
to lateralize the source of a sound on the basis of an interau-
ral time delay. Similar to findings of Herman et al. ~1977!,
data revealed that older listeners required approximately
twice the interaural time delay as young listeners for the
same level of performance. The interaction of age and sound
level indicated that a decrease in stimulus level had a more
deleterious effect on ITD discrimination in the elderly group
than in the younger group.
D. Masking level difference
The mean MLD was 7.0 dB for the young listeners,
which is consistent with previous MLD studies using similar
stimuli ~Grose et al., 1994; Wilson et al., 1982!. The elderly
group had a mean MLD of 4.9 dB. This age difference was
significant, as shown by a ttest @t(22)55.31, p,0.0001#.
Individual thresholds for the S0N0and S
p
N0conditions
are displayed in the two panels of Fig. 5 for young and
elderly listeners. Pichora-Fuller and Schneider ~1991, 1992!
using pure-tone stimuli, and Grose et al. ~1994!using speech
stimuli, found that elderly listeners with normal hearing sen-
FIG. 4. Mean ITD thresholds for young and elderly adult listeners at the 4-,
8-, and 16-dB sound levels presented relative to individual detection thresh-
olds. Error bars indicate 6standard error.
FIG. 5. Individual and group mean speech thresholds for S0N0and S
p
N0for
young and elderly listeners. Individual listeners are ordered by magnitude of
masking level difference, within each age group.
2393 2393J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
sitivity from 250 to 2000 Hz, exhibited an elevation in S
p
N0
thresholds but not S0N0thresholds as compared to younger
listeners. Present data show differences between the young
and elderly groups for both S0N0and S
p
N0conditions. This
was assessed using an ANOVA with signal condition (S0N0
vs S
p
N0) as the within-subjects factor and age group as the
between-subjects factor. Results revealed significant effects
of age @F(1,22)529.66, p,0.0001#and signal condition
@F(1,22)5892.67, p,0.0001#, as well as a significant inter-
action @F(1,22)528.19, p,0.0001#. Unlike past studies,
S0N0thresholds were more variable in the young group as
compared to the elderly group. Nonetheless, intersubject
variability of both groups was relatively small (S0N052.0
and 1.1 standard deviations; S
p
N051.9 and 1.5 standard de-
viations, for young and elderly groups, respectively!.Toin-
terpret the interaction effect, the simple main effect of age
group at each of the two conditions was assessed. In contrast
to previously reported findings, analysis revealed that the
two groups differed significantly for both the S0N0condition
at the p,0.0001 level and for the S
p
N0condition at the p
,0.001 level. Although both were statistically significant,
differences in threshold between groups were greater in the
S
p
N0~4.6 dB!than in the S0N0~2.5 dB!condition.
Previous research has reported age-related deficits in
binaural processing reflected by reduced MLDs in elderly
listeners ~Grose et al., 1994; Pichora-Fuller and Schneider,
1991; Tillman et al., 1973; Warren et al., 1978!. Findings
may have been due, in part, to age differences, although
peripheral hearing loss in the older subjects may have ac-
counted for their smaller MLDs since the presence of periph-
eral hearing loss significantly reduces the MLD ~Jerger et al.,
1984!. Present data support that binaural release from mask-
ing declines significantly as a function of advancing age,
independent of peripheral hearing loss.
E. Correlations among measures
There was no significant correlation between the experi-
mental measures and individual pure-tone thresholds or age
for either subject group. Pearson rcorrelation coefficients
were calculated to examine the relationships among experi-
mental measures. The correlation analysis used the following
variables: GAP~4!, GAP~8!, GAP~16!, GAP~slope!, ITD~4!,
ITD~8!, ITD~16!, ITD~slope!~representing mean gap and
ITD thresholds at each presentation level and mean slope
values for both tasks!, MLD, VOT~slope!, VOT1 and VOT2
~representing the slope of the identification function and per-
cent identification scores at the phoneme boundary for one-
and two-step discrimination tasks!. For the following discus-
sion we refer to the gap detection and ITD thresholds as
‘‘psychoacoustic measures’’ and the VOT scores and MLDs
as ‘‘speech measures.’’ The correlation matrices for young
and elderly adult subjects are presented in Tables II and III,
respectively.
For young adult subjects, the slope measures of ITD and
gap functions were strongly correlated. A significant rela-
tionship was found between GAP~4!measures across all ITD
variables. The same finding was observed between GAP~8!
measures and all ITD variables. No significant correlations
were found for the GAP~16!condition and ITD variables.
ITD~16!scores had lower correlations with the gap tasks
than the ITD~4!and ITD~8!scores. Taken together, these
patterns suggest that when trying to identify individual dif-
ferences across tasks, it may be best to challenge the auditory
system by working at low sound levels. When only supra-
threshold levels are used, individual differences may be dif-
ficult to find. For elderly subjects, ITD thresholds were nega-
tively and significantly correlated with gap detection thresh-
olds at lower sound levels, however, this relationship was
only of moderate statistical significance. There was no sig-
TABLE II. Correlation coefficients between psychoacoustic and speech measures of temporal processing for young adult subjects.
2394 2394J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
nificant correlation between the slopes of ITD and gap func-
tions for the elderly adult group. Overall, correlations be-
tween gap and ITD thresholds did not show a strong or
consistent relationship for elderly subjects.
The relationship between speech measures was also ex-
amined. There were no significant correlations between
MLD and VOT measures for either age group, indicating
that performance on the MLD task was independent of per-
formance on the VOT tasks.
Finally, there were no significant correlations observed
between psychoacoustic and speech measures for young
adult listeners. For elderly adult listeners, ITD thresholds
obtained at 16 dB were moderately correlated with perfor-
mance on the one-step discrimination task (r520.602). No
other significant correlations were observed, however, sug-
gesting an incidental relationship.
III. GENERAL DISCUSSION
A. Aging and monaural temporal processing
In the limited amount of research available on monaural
temporal processing ability in elderly listeners, results have
been mixed, and interpretation has been complicated by the
presence of peripheral hearing loss in the elderly samples.
Present data for normally hearing elderly listeners demon-
strate a clear deficit in the ability of elderly subjects to pro-
cess temporal information, as measured by detection of tem-
poral gaps. Results are in agreement with recent studies by
Moore et al. ~1992!, Schneider et al. ~1994!, and Snell
~1997!who reported larger gap detection thresholds in eld-
erly listeners with minimal hearing loss. Additional studies
measuring gap discrimination thresholds and duration dis-
crimination thresholds support that there is an age-related
deficit in monaural temporal processing that is independent
of hearing loss ~Abel et al., 1990; Fitzgibbons and Gordon-
Salant, 1994; Trainor and Trehub, 1989!.
A similar result was observed for monaural temporal
processing of speech stimuli, as measured by sensitivity to
changes in VOT. Elderly subjects were less able to clearly
distinguish phoneme categories and were less accurate at dis-
criminating speech stimuli which differed in VOT duration.
Findings suggest that elderly listeners may be at a functional
disadvantage in the perception of temporal changes in the
acoustic waveform that comprise everyday conversational
speech.
The question of whether deficits in temporal processing
ability contribute to speech perception difficulties is an area
of considerable controversy. Reduced temporal processing
has been linked to some speech perception errors made by
hearing-impaired listeners ~Erber, 1972; Price and Simon,
1984; Tyler et al., 1982!, and similar errors have been dem-
onstrated by normal-hearing listeners as a consequence of
changes in temporal relations among signal components
~Klatt, 1975; Schouten, 1980!. In relation to gap detection,
several investigations have revealed significant correlations
between gap detection thresholds and speech recognition
ability, even when audiometric threshold is factored out
~Dreschler and Plomp, 1985; Glasberg and Moore, 1989;
Tyler et al., 1982!. Taken together, these results suggest that
poorer speech perception might be related, in part, to abnor-
mal temporal processing. Despite this evidence, auditory
temporal processing measures have not consistently proven
to be strong predictors of speech perception performance in
elderly listeners ~Festen and Plomp, 1983; Lutman and
Clark, 1986; vanRooij and Plomp, 1990!. Results of the
present investigation support this latter finding.
As Dorman et al. ~1985!have suggested based on pre-
vious research, observed differences between young and eld-
erly listeners on psychoacoustic measures, although statisti-
cally significant, may be small when compared to the
magnitude of the differences in acoustic segment duration
which signal phonetic contrasts in normal speech. In the case
TABLE III. Correlation coefficients between psychoacoustic and speech measures of temporal processing for elderly adult subjects.
2395 2395J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
of gap detection, there is no doubt that silent intervals in
continuous speech can have linguistic importance. For ex-
ample, introduction of a period of silence between the @s#and
@l#in the word ‘‘slit’’ results in the perception of the word
‘‘split’’ ~Bastian et al., 1961; Marcus, 1978!. Relative dura-
tion of the silent interval is also important. One of the cues to
voicing in intervocalic stops ~e.g., rapid–rabid!is the dura-
tion of closure for the stop consonant ~Lisker, 1957!. The
question, therefore, may not be whether periods of silence
are linguistically important but, rather, whether gap detection
thresholds are so elevated in elderly listeners that these silent
periods are poorly perceived. For example, in speech, when a
silent interval signals the presence of a stop consonant in a
cluster, that interval is on the order of 80 ms or longer, while
the absence of a stop is signaled by intervals of less than 20
ms ~Dorman et al., 1985!. The distinction between ‘‘slit’’
and ‘‘split’’ is evoked with a pause of at least 30–45 ms.
Relative to the present data, the poorest gap detection thresh-
old measured at the loudest presentation level among elderly
listeners was approximately 15 ms, and most elderly listeners
exhibited even better resolution at this sound level. Therefore
linguistically relevant silent periods in speech should be eas-
ily perceived by these listeners if the signal-to-noise ~S/N!
ratio is adequate. However, at near-threshold levels, gap de-
tection thresholds for elderly listeners ranged from 6 to 60
ms ~as compared to the 4–15-ms range for young adult lis-
teners!, suggesting that the magnitude of the loss in some
elderly listeners may be such that speech perception is ad-
versely affected.
It may also be the case that the failure to find a signifi-
cant correlation indicates that the mechanisms underlying de-
tection of temporal gaps are somehow different from mecha-
nisms underlying the ability to distinguish temporal
characteristics of speech. Although a complete discussion is
beyond the scope of this paper, there is evidence that speech
stimuli are perceived and processed in a different way from
nonspeech stimuli. The evidence derives from studies of cat-
egorical perception, the phenomenon that speech sounds can
be discriminated only when they are identified as being lin-
guistically different ~Liberman et al., 1967!; from studies of
cerebral asymmetry, which indicate that certain parts of the
brain are specialized for dealing with speech ~Broadbent and
Gregory, 1964; Kimura, 1964!; and from the speech–
nonspeech dichotomy, which shows that when a listener is
presented with speechlike sounds there is a perceptual di-
chotomy in that the sounds are either perceived as speech or
they are not ~House et al., 1962; Stevens and House, 1972!.
Thus it appears that the perception of speech sounds may be
fundamentally different from that of nonspeech stimuli,
which could account for the lack of significant correlation
observed between measures of monaural temporal process-
ing in the present study. Finally, significant correlations
might have been found if a similar presentation level was
used for both psychoacoustic and speech perception mea-
sures or if different tasks had been used.
Despite the lack of observed correlations and potential
differences in the underlying processes involved in temporal
processing of speech and nonspeech stimuli, results of the
present investigation still suggest an age-related decline in
monaural temporal processing of elderly listeners for both
speech and nonspeech stimuli, even when peripheral hearing
sensitivity is considered clinically normal. Importantly, these
deficits are likely to be exaggerated in listeners with hearing
loss.
B. Aging and binaural processing
The present data demonstrate a clear age-related loss in
the ability to lateralize the source of a sound on the basis of
an interaural time delay, replicating results first reported by
Matzker and Springborn in the late 1950s. Similar to findings
of Herman et al. ~1977!, the present data revealed that older
listeners required approximately twice the interaural time de-
lay as young listeners for the same level of performance.
To examine how decreased sensitivity to interaural time
differences might contribute to speech perception difficulties,
binaural processing for speech was measured using the
MLD. Because binaural processing is acutely sensitive to
interaural time differences, any loss of temporal resolution in
the nervous system would be expected to reduce the size of
the MLD ~Durlach, 1972!. It was found that elderly listeners
performed more poorly than young listeners, primarily for
the S
p
N0condition, contributing to a 2-dB age effect on the
MLD. This supports ITD findings that the elderly have a
reduced ability to process interaural time cues, even when
they have normal hearing. The MLD difference, while not
large numerically, is important because it represents another
way in which the effects of age probably reduce hearing
efficiency in complex listening situations. It has been dem-
onstrated that a loss of 1 dB S/N ratio may result in as much
as a 20% reduction in speech intelligibility ~Plomp, 1986!.
Thus a loss of 2 dB S/N ratio could result in considerable
difficulty in everyday listening. Furthermore, it should be
remembered that while the experimental group in this study
was composed of elderly individuals, they did not manifest
clinical symptoms of hearing impairment. The deficits which
they showed in binaural signal processing are likely to be
exaggerated in listeners with significant hearing loss.
Since ITD and MLD thresholds both measure aspects of
binaural processing of auditory stimuli, it was hypothesized
that performance on the ITD task may be related to the abil-
ity to use directional cues to perceive speech in noise as
measured by the MLD. Correlations between mean ITD
threshold and mean MLD scores for subjects in the present
study, however, did not reach statistical significance. This
lack of significant correlation between the two measures is
consistent with findings of Warren et al. ~1978!using similar
binaural measures. The failure to find a significant correla-
tion is perhaps not surprising since the mechanisms involved
in binaural signal analysis are complex. As previously sug-
gested, results indicate that the binaural mechanisms under-
lying discrimination of interaural time differences are likely
different from those of binaural processing for speech.
Again, significant correlations might have been found if a
similar presentation level was used for both psychoacoustic
and speech perception measures or if different tasks had been
used.
Although no significant correlations between measures
of binaural processing were found, present findings do indi-
2396 2396J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
cate an age-related decline in binaural processing of temporal
information in both speech and nonspeech stimuli. The re-
duced ability of elderly listeners to utilize interaural time
cues has implications for tasks involving lateralization or
localization. More importantly, the introduction of binaural,
dichotic time cues has been shown to significantly improve
the detection and perception of speech signals, especially in
a background of noise. Although some of this ability arises
from monaural cues such as frequency, timing, or syntax of
the source, binaural cues improve the range of situations in
which effective communication is possible. If the aging au-
ditory system fails to preserve these cues, older adults may
be at a functional disadvantage in the perception of speech in
any situation where surrounding auditory space is noisy.
C. Relationships between psychoacoustic measures
Previous research has shown that temporal resolution
worsens at low sound levels ~Ashmead et al., unpublished
data; Buus and Florentine, 1985; Shailer and Moore, 1983;
Viemeister, 1979!. Present results revealed a significant in-
teraction between age and sound level, indicating that at a
given stimulus level older adults performed more poorly than
younger adults, and this age difference increased as the over-
all stimulus level decreased. This interaction between age
and signal level was found consistently across gap detection
and ITD tasks. Thus the temporal resolution of elderly lis-
teners was more adversely affected by low stimulus level,
even when level was specified relative to individual thresh-
olds. The contribution of spectral cues to the ability to detect
temporal gaps as reported by Schneider et al. ~1994!, how-
ever, cannot be ruled out.
Correlations between the slopes of ITD and gap detec-
tion functions showed a highly significant positive relation-
ship for young adult subjects, suggesting that changes in per-
formance as a function of signal level were comparable
across monaural and binaural types of processing. For eld-
erly subjects, however, there was no significant correlation
between the slope of the ITD function and that of the gap
detection function. This result indicates that although there
was a significant interaction of age and signal level across
tasks, changes in performance as a function of signal level
for elderly listeners were not consistent across measures.
Thus the relationship between ITD and gap detection thresh-
olds was quite different for young versus elderly subjects,
indicating processing difficulties of the aging auditory sys-
tem at low sound levels that are not attributable to a single,
simple factor such as temporal resolution loss in the auditory
periphery. Rather, the two types of temporal processing mea-
sured in this study may be affected differently by the aging
process. It is likely that central factors play a role as well,
especially with regard to ITD discrimination ~vonWedel
et al., 1991; Yin and Chan, 1988!. It also may be the case
that binaural performance on the ITD task may have been
affected by differences in monaural performance between
ears on the gap detection task. Results for gap detection in
the present study were obtained for one ear of each subject.
Schneider et al. ~1994!found that in older listeners, there
were sometimes large differences in gap detection thresholds
within the same subject between the two ears. Thus the fact
that a significant correlation was found between gap detec-
tion and ITD performance in young but not older subjects
may be because there are only small interaural differences in
the gap detection thresholds of young listeners but larger
differences may occur in older listeners.
IV. CONCLUSION
The results of this study support an age-related decrease
in temporal processing ability. Elderly listeners had higher
thresholds versus younger listeners on gap detection and ITD
tasks, were less able to benefit from binaural release from
masking, and were less accurate in discriminating changes in
voice onset time. Importantly, these findings were observed
in elderly listeners with no clinical symptoms of hearing
loss. Findings support the following conclusions:
~1!Elderly persons with excellent hearing sensitivity
nonetheless perform worse than younger persons on mea-
sures of temporal resolution and speech perception. It should
be noted that although strict criteria were used to select sub-
jects for the present study, future studies could apply more
stringent criteria to ensure that elderly subjects have auditory
peripheries that match those of young subjects.
~2!On measures of monaural and binaural processing of
temporal information, elderly persons are more adversely af-
fected than younger persons by having to listen at very low
sound levels, even when those sound levels are defined with
respect to individual thresholds.
~3!For younger persons, monaural performance ~of one
ear!and binaural performance were related when testing was
performed at low sound levels ~suggesting that the auditory
system must be challenged in order to see individual differ-
ences!, but for elderly persons there appeared to be no rela-
tion between measures of temporal resolution ~perhaps sug-
gesting unique effects of aging on specific auditory abilities!.
~4!There is no evidence that individual differences in
speech perception can be accounted for by temporal resolu-
tion abilities. This result is based on a limited number of
subjects and a limited range of tasks.
ACKNOWLEDGMENTS
This work was supported by a Dissertation Enhancement
Award through the Graduate School at Vanderbilt Univer-
sity. The authors express appreciation to Gene Bratt and Le-
slie Smith for constructive comments on an earlier version of
this paper, Xuefeng Yang, for assistance in computer pro-
gramming, and the Audiology staff at the VA Medical Cen-
ter, Mountain Home, for comments and suggestions.
1To obtain masked thresholds for the gap detection task, each subject was
presented with a 75 dB SPL continuous 1000-Hz pure tone. Using the
computer keyboard, the subject manually attenuated the level of the tone
until it was just barely audible in the noise background ~descending series!.
A second threshold was measured using an ascending series, where the tone
was initially presented at 0 dB SPL and the subject again adjusted the level
of the tone until it was judged as barely audible. A third threshold was
measured with the initial amount of attenuation adjusted to match the indi-
vidual subject’s threshold based on the first two threshold approximations,
and the subject manually adjusted the level of the tone for a third time to a
level just barely audible in the noise background. The mean of the three
estimates was considered threshold. Masked tone thresholds ranged from
43 to 51 (mean547.1) dB SPL for the young adult group and from 43 to
2397 2397J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
50 (mean547.3) dB SPL for the elderly group. Masked gap detection
thresholds for the 1000-Hz tone did not differ significantly between groups
@F(1,23)50.073, p.0.05#.
2To obtain binaural click detection thresholds for the ITD task, each subject
was presented with a 70 dB SPL continuous click train. A method of
adjustment was used to determine individual thresholds, identical to that
used in the gap detection task. Click thresholds ranged from 14 to 26
(mean520.3) dB SPL for young adult listeners and from 18 to 28 (mean
523.3) dB SPL for the elderly group. Click thresholds did not differ sig-
nificantly between groups @F(1,23)52.30, p.0.05#.
Abel, S. M., Krever, E. M., and Alberti, P. W. ~1990!. ‘‘Auditory detection,
discrimination and speech processing in aging, noise sensitive and
hearing-impaired listeners,’’ Scand. Audiol. 19, 43–54.
ANSI ~1969!. ANSI S3.6-1969 ~R 1973!, ‘‘Specifications for Audiometers’’
~American National Standards Institute, New York!.
Ashmead, D. H., Tetzeli, M., Chandler, D., and Thomas, A. ~unpublished
data!. ‘‘Aging and sound level effects in interaural time discrimination:
Efficiency of auditory processing.’’
AVAAZ Innovations, Inc. ~1994!.Computerized Speech Research Environ-
ment ~Ontario, Canada!.
Bacon, S. P., and Viemeister, N. F. ~1985!. ‘‘Temporal modulation transfer
functions in normal-hearing and hearing-impaired listeners,’’ Audiology
24, 117–134.
Bastian, J., Eimas, P. D., and Liberman, A. M. ~1961!. ‘‘Identification and
discrimination of a phonetic contrast induced by a silent interval,’’ J.
Acoust. Soc. Am. 33, 842.
Blumstein, S. E., and Stevens, K. N. ~1980!. ‘‘Perceptual invariance and
onset spectra for stop consonants in different vowel environments,’’ J.
Acoust. Soc. Am. 67, 647–662.
Broadbent, D. E., and Gregory, M. ~1964!. ‘‘Accuracy of recognition for
speech presented to the right and left ears,’’ J. Exp. Psychol. 16, 359–360.
Buus, S., and Florentine, M. ~1985!. ‘‘Gap detection in normal and impaired
listeners: The effect of level and frequency,’’ in Time Resolution in Audi-
tory Systems, edited by A. Michelson ~Springer-Verlag, Berlin!.
Cranford, J. L., Boose, M., and Moore, C. A. ~1990!. ‘‘Effects of aging on
the precedence effect in sound localization,’’ J. Speech Hear. Res. 33,
654–659.
DeFillippo, C. L., and Snell, K. ~1986!. ‘‘Detection of temporal gaps in
low-frequency narrow-band signals by normal-hearing and hearing-
impaired listeners,’’ J. Acoust. Soc. Am. 80, 1354–1358.
Department of Veterans Affairs ~1991!.Speech Recognition and Identifica-
tion Materials (Disc 1.1.) ~Auditory Research Laboratory, VA Medical
Center, Long Beach, CA!.
Divenyi, P. L., and Haupt, K. M. ~1997!. ‘‘Audiological correlates of speech
understanding deficits in elderly listeners with mild-to-moderate hearing
loss. III. Factor representation,’’ Ear Hear. 18, 189–201.
Dorman, M. F., Marton, K., Hannley, M. T., and Lindholm, J. M. ~1985!.
‘‘Phonetic identification by elderly normal and hearing-impaired listen-
ers,’’ J. Acoust. Soc. Am. 77, 664–670.
Dorman, M. F., Raphael, L. J., and Liberman, A. M. ~1979!. ‘‘Some experi-
ments on the sound of silence in phonetic perception,’’ J. Acoust. Soc.
Am. 65, 1518–1532.
Dreschler, W. A., and Plomp, R. ~1985!. ‘‘Relations between psychophysi-
cal data and speech perception for hearing impaired subjects. II,’’ J.
Acoust. Soc. Am. 78, 1261–1270.
Durlach, N. I. ~1972!. ‘‘Binaural signal detection: Equalization and cancel-
lation theory,’’ in Foundations of Modern Auditory Theory, edited by J. V.
Tobias ~Academic, New York!.
Durlach, N. I., Thompson, C. L., and Colburn, H. S. ~1981!. ‘‘Binaural
interaction in impaired listeners: A review of past research,’’ Audiology
20, 181–211.
Erber, N. P. ~1972!. ‘‘Speech-envelope cues as an acoustic aid to lipreading
for profoundly deaf children,’’ J. Acoust. Soc. Am. 51, 1224–1227.
Festen, J. M., and Plomp, R. ~1983!. ‘‘Relations between auditory functions
in impaired hearing,’’ J. Acoust. Soc. Am. 73, 652–662.
Findlay, R. C., and Schuchman, G. I. ~1976!. ‘‘Masking level difference for
speech: Effects of ear dominance and age,’’ Audiology 15, 232–241.
Fitzgibbons, P. J. ~1983!. ‘‘Temporal gap detection in noise as a function of
frequency, bandwidth, and level,’’ J. Acoust. Soc. Am. 74, 67–72.
Fitzgibbons, P. J., and Gordon-Salant, S. ~1994!. ‘‘Age effects on measures
of auditory duration discrimination,’’ J. Speech Hear. Res. 37, 662–670.
Florentine, M., and Buus, S. ~1984!. ‘‘Temporal gap detection in senso-
rineural and simulated hearing impairments,’’ J. Speech Hear. Res. 27,
449–455.
Fujisaki, S., and Kawashima, K. ~1970!. ‘‘Some experiments on speech
perception and a model for the perceptual mechanism,’’ Annual Report of
the Engineering Research Institute, Faculty of Engineering, University of
Tokyo, Tokyo, Vol. 29, pp. 207–214.
Ginzel, A., Pedersen, C. B., Spliid, P. E., and Andersen, E. ~1982!. ‘‘The
role of temporal factors in auditory perception of consonants and vowels,’’
Scand. Audiol. 11, 93–100.
Glasberg, B. R., and Moore, B. C. J. ~1989!. ‘‘Psychoacoustic abilities of
subjects with unilateral and bilateral cochlear impairments and their rela-
tionship to the ability to understand speech,’’ Scand. Audiol. 32, 1–25.
Godfrey, J., and Millay, K. ~1978!. ‘‘Perception and rapid spectral change in
speech by listeners with mild and moderate sensorineural hearing loss,’’ J.
Am. Audiol. Soc. 3, 200–208.
Gregory, R. L. ~1974!. ‘‘Increase in neurological noise as a factor in aging,’’
in Concepts and Mechanisms of Perception, edited by R. L. Gregory
~Duckworth, London!.
Grose, J. H., Poth, E. A., and Peters, R. W. ~1994!. ‘‘Masking level differ-
ences for tones and speech in elderly listeners with relatively normal au-
diograms,’’ J. Speech Hear. Res. 37, 422–428.
Hausler, R., Colburn, S., and Marr, E. ~1983!. ‘‘Sound localization in sub-
jects with impaired hearing,’’ Acta Oto-Laryngol. Suppl. 400, 6–62.
Herman, G. E., Warren, L. R., and Wagener, J. W. ~1977!. ‘‘Auditory lat-
eralization: Age differences in sensitivity to dichotic time and amplitude
cues,’’ J. Gerontol. 32, 187–191.
House, A. S., Stevens, K. N., Sandel, T. T., and Arnold, J. B. ~1962!. ‘‘On
the learning of speechlike vocabularies,’’ J. Verb. Learn. Verb. Behav. 1,
133–143.
Irwin, R. J., and McAuley, S. F. ~1987!. ‘‘Relations among temporal acuity,
hearing loss, and the perception of speech distorted by noise and rever-
beration,’’ J. Acoust. Soc. Am. 81, 1557–1565.
Jerger, J., Brown, D., and Smith, S. ~1984!. ‘‘Effect of peripheral hearing
loss in the masking level difference,’’ Arch. Otolaryngol. 110, 290–296.
Jesteadt, W., Bacon, S. P., and Lehman, J. R. ~1982!. ‘‘Forward masking as
a function of frequency, masker level, and signal delay,’’ J. Acoust. Soc.
Am. 71, 950–962.
Kelly-Ballweber, D., and Dobie, R. A. ~1984!. ‘‘Binaural interaction mea-
sured behaviorally and electrophysiologically in young and old adults,’’
Audiology 23, 181–194.
Kimura, D. ~1964!. ‘‘Left-right differences in the perception of melodies,’’
J. Exp. Psychol. 16, 355–358.
Kirikae, I. ~1969!. ‘‘Auditory function in advanced age with reference to
histological changes in the central auditory system,’’ Intern. Audiol. 8,
221–230.
Klatt, D. H. ~1975!. ‘‘Voice onset time, frication, and aspiration in word-
initial consonant clusters,’’ J. Speech Hear. Res. 18, 686–706.
Klatt, D. H. ~1980!. ‘‘Software for a cascade/parallel formant synthesizer,’’
J. Acoust. Soc. Am. 67, 971–995.
Levitt, H. ~1971!. ‘‘Transformed up-down methods in psychoacoustics,’’ J.
Acoust. Soc. Am. 49, 467–477.
Liberman, A. M., Cooper, F. S., Shankweiler, D. P., and Studdert-Kennedy,
M. ~1967!. ‘‘Perception of the speech code,’’ Psychol. Rev. 74, 431–461.
Lisker, L. ~1957!. ‘‘Closure duration and the inter-vocalic voiced-voiceless
distinction in English,’’ Language 33, 42–49.
Lisker, L., and Abramson, A. S. ~1970!. ‘‘The voicing dimension: Some
experiments in comparative phonetics.’’ in Proceedings of the VI Interna-
tional Congress of Phonetic Sciences ~Academia, Prague!.
Lutman, M. E., and Clark, J. ~1986!. ‘‘Speech identification under simulated
hearing-aid frequency response characteristics in relation to sensitivity,
frequency resolution, and temporal resolution,’’ J. Acoust. Soc. Am. 89,
1030–1040.
Marcus, S. M. ~1978!. ‘‘Distinguishing ‘slit’ and ‘split’-an invariant timing
cue in speech perception,’’ Percept. Psychophys. 23, 58–60.
Martin, D. R., and Cranford, J. L. ~1991!. ‘‘Age-related changes in binaural
processing. II. Behavioral findings,’’ Am. J. Otolaryngol. 12, 365–369.
Matzker, V. J., and Springborn, E. ~1958!. ‘‘Richtungshoren und leben-
salter,’’ Zeitschrift fur Laryngologie, Rhinologie, Otologie und Ihre Gren-
zgebiete 37, 739–745.
Moore, B. C. J. ~1982!.Introduction to the Psychology of Hearing ~Aca-
demic, London!.
Moore, B. C. J., and Glasberg, B. R. ~1983!. ‘‘Growth of forward masking
for sinusoidal and noise maskers as a function of signal delay: Implication
for suppressions in noise,’’ J. Acoust. Soc. Am. 73, 1249–1259.
2398 2398J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
Moore, B. C. J., and Glasberg, B. R. ~1988!. ‘‘Gap detection with sinusoids
and noise in normal, impaired, and electrically stimulated ears,’’ J.
Acoust. Soc. Am. 83, 1093–1101.
Moore, B. C. J., Glasberg, B. R., Donaldson, E., McPherson, T., and Plack,
C. J. ~1989!. ‘‘Detection of temporal gaps in sinusoids by normally hear-
ing and hearing-impaired subjects,’’ J. Acoust. Soc. Am. 85, 1266–1275.
Moore, B. C. J., Peters, R. W., and Glasberg, B. R. ~1992!. ‘‘Detection of
temporal gaps in sinusoids by elderly subjects with and without hearing
loss,’’ J. Acoust. Soc. Am. 92, 1923–1932.
Moore, B. C. J., Peters, R. W., and Glasberg, B. R. ~1993!. ‘‘Detection of
temporal gaps in sinusoids: Effects of frequency and level,’’ J. Acoust.
Soc. Am. 93, 1563–1570.
Olsen, W. O., Noffsinger, D., and Carhart, R. ~1976!. ‘‘Masking level dif-
ferences encountered in clinical populations,’’ Audiology 15, 287–301.
Palva, A., and Jokinen, K. ~1970!. ‘‘Presbyacusis,’’ Acta Oto-Laryngol. 70,
232–241.
Pichora-Fuller, M. K., and Schneider, B. A. ~1991!. ‘‘Masking level differ-
ences in the elderly: A comparison of antiphasic and time-delay dichotic
conditions,’’ J. Speech Hear. Res. 34, 1410–1422.
Pichora-Fuller, M. K., and Schneider, B. A. ~1992!. ‘‘The effect of interau-
ral delay of masker on masking-level differences in young and old
adults,’’ J. Acoust. Soc. Am. 91, 2129–2135.
Pickett, J. M., and Decker, L. R. ~1960!. ‘‘Time factors in the perception of
a double consonant,’’ Lang. Speech 3, 11–17.
Plomp, R. ~1986!. ‘‘A signal-to-noise ratio model for the speech reception
threshold of the hearing-impaired,’’ J. Speech Hear. Res. 29, 146–154.
Pollack, I., and Pisoni, D. B. ~1971!. ‘‘On the comparison between identi-
fication and discrimination tests in speech perception,’’ Psychonom. Sci.
24, 299–300.
Price, P. J., and Simon, H. J. ~1984!. ‘‘Perception of temporal differences in
speech by normal-hearing adults: Effects of age and intensity,’’ J. Acoust.
Soc. Am. 72, 405–410.
Repp, B., Liberman, A. M., Eccardt, T., and Pesetsky, D. ~1978!. ‘‘Percep-
tual integration and differentiation of acoustic cues for stop, fricative, and
affricate manner,’’ J. Exp. Psychol. Hum. Percept. Perf. 4, 621–637.
Roush, J. ~1985!. ‘‘Aging and binaural auditory processing,’’ Sem. Hear. 6,
135–146.
Schneider, B. A. ~1997!. ‘‘Psychoacoustics and aging: Implications for ev-
eryday listening,’’ J. Speech Lang. Path. Audiol. 21, 111–124.
Schneider, B. A., Pichora-Fuller, M. K., Kowalchuk, D., and Lamb, M.
~1994!. ‘‘Gap detection and the precedence effect in young and old
adults,’’ J. Acoust. Soc. Am. 95, 980–991.
Schouten, M. E. H. ~1980!. ‘‘The case against the speech mode of percep-
tion,’’ Acta Psychol. 44, 71–98.
Shailer, M. J., and Moore, B. C. J. ~1983!. ‘‘Gap detection as a function of
frequency, bandwidth, and level,’’ J. Acoust. Soc. Am. 74, 467–473.
Sharf, D. J., Ohde, R. N., and Lehman, M. E. ~1988!. ‘‘Relationship be-
tween the discrimination or /w-r/ and /t-d/ continua and the identification
of distorted /r/,’’ J. Speech Hear. Res. 31, 193–206.
Snell, K. B. ~1997!. ‘‘Age-related changes in temporal gap detection,’’ J.
Acoust. Soc. Am. 101, 2214–2220.
Stevens, K. N., and House, A. S. ~1972!. ‘‘Speech perception,’’ in Founda-
tion of Modern Auditory Theory, Volume 2, edited by J. V. Tobias ~Aca-
demic, New York!.
Stevens, K. N., Liberman, A. M., Studdert-Kennedy, M., and Ohman, S. E.
G. ~1969!. ‘‘Crosslanguage study of vowel perception,’’ Lang. Speech 12,
1–23.
Tillman, T. A., Carhart, R., and Nicholls, S. ~1973!. ‘‘Release from multiple
maskers in elderly persons,’’ J. Speech Hear. Res. 16, 152–160.
Trainor, L. J., and Trehub, S. E. ~1989!. ‘‘Aging and auditory temporal
sequencing: Ordering the elements of repeating tone patterns,’’ Percept.
Psychophys. 45, 417–426.
Tyler, R. S., and Summerfield, Q. ~1980!. ‘‘Psychoacoustical and phonetic
measures of temporal processing in normal and hearing-impaired listen-
ers,’’ in Psychophysical, Physiological, and Behavioural Studies in Hear-
ing, edited by G. van den Brink and F. A. Bilsen ~Delft U.P., Delft!.
Tyler, R. S., Summerfield, Q., Wood, E. J., and Fernandes, M. A. ~1982!.
‘‘Psychoacoustic and phonetic temporal processing in normal and hearing-
impaired listeners,’’ J. Acoust. Soc. Am. 72, 740–752.
vanRooij, J. C. G. M., and Plomp, R. ~1990!. ‘‘Auditive and cognitive
factors in speech perception by elderly listeners. II. Multivariate analy-
ses,’’ J. Acoust. Soc. Am. 88, 2611–2624.
vonWedel, H., vonWedel, U., and Streppel, M. ~1991!. ‘‘Monaural and bin-
aural time resolution ability in the aged,’’ Acta Oto-Laryngol. 476, 161–
166.
Viemeister, N. F. ~1979!. ‘‘Temporal modulation transfer functions based on
modulation thresholds,’’ J. Acoust. Soc. Am. 66, 1364–1380.
Wardrip-Fruin, C. ~1982!.Final Stop Voicing: Vowel Duration the Primary
Cue in Noise, Meetings of the American Speech and Hearing Association,
November.
Warren, L. R., Wagener, J. W., and Herman, G. E. ~1978!. ‘‘Binaural analy-
sis in the aging auditory system,’’ J. Gerontol. 33, 731–736.
Wilson, R. H., Hopkins, J. L., Mance, C. M., and Novak, R. E. ~1982!.
‘‘Detection and recognition masking-level differences for the individual
CID W-1 spondaic words,’’ J. Speech Hear. Res. 25, 235–242.
Yin, T. C. T., and Chan, J. C. K. ~1988!. ‘‘Neural mechanisms underlying
interaural time sensitivity to tones and noise,’’ in Auditory Functioning:
Neurobiological Bases of Hearing, edited by G. M. Edelman, W. E. Gall,
and W. M. Cowan ~Wiley, New York!.
Zwislocki, J., and Feldman, R. S. ~1956!. ‘‘Just noticeable differences in
dichotic phase,’’ J. Acoust. Soc. Am. 28, 860–864.
2399 2399J. Acoust. Soc. Am., Vol. 104, No. 4, October 1998 Strouse
et al.
: Temporal processing in aging
