Session 1pPP: Monday Afternoon, March 15, 1999,
Paper 1pPPa4. Recent advances in models of binaural detection and localization.
Time: 3:00
H. Steven Colburn
(Hearing Res. Ctr. and Biomed. Eng. Dept., Boston Univ.)
There has been a significant increase in the mathematical modeling of
binaural phenomena in the past couple of years by a number of research
groups. Available review chapters [e.g., Colburn, in Auditory Computation
(Springer-Verlag, New York, 1996); Stern and Trahiotis, in Handbook of
Perception (Academic, New York, 1996)] do not include this new material.
Recent advances were stimulated in part by the availability of more
powerful computational tools and in part by recent data that include more
complex stimulus situations such as noise waveforms with interaural time
differences but no interaural intensity differences and vice versa [van de
Par and Kohlrausch, J. Acoust. Soc. Am. (1998)]. This presentation reviews
recent developments in binaural modeling with particular attention given
to
models of binaural detection and sound localization. Both pink-box models,
which incorporate available information from auditory physiology, and
black-box models are considered.
[Work supported by NIDCD Grant No. R01
DC00100.]
Paper 1pPPa10. Modeling the precedence effect: Mechanisms of onset enhancement
in
binaural lateralization models.
Time:5:20
Roberto M. Dizon and H. Steven Colburn
(Hearing Res. Ctr. and Biomed. Eng. Dept., Boston Univ.)
The general behaviors of several cross-correlation-based lateralization
models in response to dynamic impulsive stimuli, specifically those that
elicit the precedence effect, are presented as a means of gaining insight
into the contribution of individual stages of the models to precedencelike
performance. In general, an enhancement of activity at onset relative to
later activity in the model outputs (or equivalently, a temporary
post-onset suppression in activity) is associated with the relative
perceptual emphasis of early portions of stimuli as described by the
precedence effect. Onset enhancement in cross-correlation models has thus
far only been exhibited using dynamic mechanisms such as that used in a
model by Lindemann [J. Acoust. Soc. Am. (1986)]. This work evaluates the
dynamic properties of several cross-correlation models that include
mechanisms of onset enhancement. These mechanisms include that of
Lindemann's model, peripheral mechanisms such as adaptation, and a
physiologically based inhibition mechanism adapted from the IC model of
Cai, Carney, and Colburn [J. Acoust. Soc. Am. (1998)]. In this work,
binaural models are created from various permutations of the mechanisms
described above, and the final and intermediate model outputs in response
to a few sample stimuli are shown and compared.
[Work supported by ONR and NIH (NIDCD R01 DC00100).]
Session 2aPP: Tuesday Morning, March 16, 1999,
Paper 2aPPa8. Monaural, cross-frequency coincidence detection as a
mechanism for decoding perceptual cues provided by the cochlear
amplifier.
Time: 11:00
Michael G. Heinz (Speech and Hearing Sci. Prog., MIT and Hearing Res. Ctr., Boston Univ.) and
Laurel H. Carney (Hearing Res. Ctr., Boston Univ.)
Basilar membrane (BM) response properties associated with the cochlear
amplifier are consistent with broadened BM tuning as level increases
from approximately 20 to 90 dB SPL or higher. Associated with
broadened tuning are compressive BM magnitude responses and
level-dependent phase changes. To evaluate the perceptual significance
of the cochlear amplifier, the associated nonlinear response
properties of the auditory nerve (AN) and physiologically realistic
mechanisms for decoding the perceptual cues they provide were
studied. Most AN fibers' rate and synchrony responses to tones are
saturated above 40 dB SPL; however, nonlinear phase changes with level
are encoded in AN responses up to 90 dB SPL or higher. Broadened
tuning produces an increase in the correlation between adjacent AN
fibers. Monaural coincidence detection is a physiologically realistic
mechanism that can extract nonlinear phase cues across AN fibers with
similar characteristic frequencies over a wide dynamic range. Methods
from statistical decision theory [Colburn, J. Acoust. Soc. Am. 54,
1458--1470 (1973)] were combined with an analytical AN model to
provide a quantitative framework for evaluating the potential of a
coincidence detector to explain human performance in pure-tone
intensity and frequency discrimination.
[Work supported by NIH Grant No. T32DC00038 and NSF Grant
No. 9601215.]
Paper 2aPPb1. Near-field localization in echoic rooms.
Time: 12:00
Scott Santarelli, Norbert Kopco, Barbara G. Shinn-Cunningham (Dept. of Cognit. and Neural Systems, Boston Univ.), and
Douglas Brungart (Armstrong Lab., Wright-Patterson AFB, OH)
The current study examines the ability of subjects to indicate the distance and direction of sources within 1 m of the head in a medium-sized, echoic classroom. Two conditions were tested in the same group of subjects. In the first, subjects were seated in the center of the room, relatively far from any hard reflective surfaces. In the second condition, subjects were located at the same position in the room, but a 12x4 enamel-covered wallboard was positioned next to the listener to create an additional artificial wall approximately 6 in. from the left ear of the listeners. The initial hypothesis was that previous localization results from tests in anechoic space would be nearly indistinguishable from the results in the first condition, since for sources near the head, the direct-to-reverberant energy ratio in this first condition would be very large. However, it was believed that the addition of a single, short-latency echo might bias some localization judgments, particularly judgments of distance and elevation. Instead the results indicate that localization accuracy and variability are comparable for the two echoic conditions, but that both measures of localization ability are worse in echoic conditions than in anechoic conditions.
Session 3aPP: Wednesday Morning, March 17, 1999,
Paper 3aPPa4. Physiological studies and neural mechanisms of echo suppression in
the inferior colliculus of the cat.
Time: 9:00
Ruth Y. Litovsky (Hearing Res. Center and Dept. of Biomed. Eng., Boston Univ.),
Bertrand Delgutte (Eaton-Peabody Lab., Massachusetts Eye and Ear Infirmary), and
Tom C. T. Yin (Dept. of Neurophysiology, Univ. Wisconsin, Madison, WI)
Neurons in the inferior colliculus (IC) are sensitive to stimulus
direction
in free field and to individual directional cues, such as interaural
disparities in time (ITD) and level (ILD). Our single-unit recordings in
the IC of anesthetized cats explored neural mechanisms in the IC that
might
mediate echo suppression, using free-field and virtual space (VS) stimuli.
Click pairs were used to simulate a sound source (lead) followed by a
single reflection (lag) at delays of 1--100 ms. In both free field and VS,
correlates of psychophysical echo suppression were found in the responses
of most IC neurons. In many neurons, the suppression was directionally
dependent in a way that could be linearly predicted from the response to
the lead, with the most effective lead locations giving maximum
suppression. Some neurons showed a decoupling between echo suppression and
the directional response to the lead, suggesting that excitation and
suppression are mediated by different neural mechanisms. Further
decoupling
was achieved by selectively holding constant some directional cues such as
ITD and ILD in the VS stimuli. Results suggest that neural suppression in
the IC depends on specific directional cues and is therefore a likely
correlate of echo suppression in spatial hearing.
[Work supported by NIH.]
Paper 3aPPa6. Binaural hearing by listeners with hearing impairments.
Time: 9:40
Monica L. Hawley, Ruth Y. Litovsky, and H. Steven Colburn
(Hearing Res. Ctr. and Biomed. Eng., Boston Univ.)
The ability of listeners with hearing impairments to understand speech
in the presence of competing speech sources is often compromised;
however, the extent to which poor binaural hearing contributes to this
compromise is not known. These studies are aimed at relating
listeners' abilities to utilize binaural information in speech-based
and non-speech-based tasks. In one set of measurements, speech
intelligibility and localization of speech material were tested in the
presence of one to three competing sources, placed at various spatial
configurations. A second set of measurements tested the sensitivity to
binaural information (i.e., MLD, ITD, and ILD
discrimination). Listeners who obtained maximal binaural benefit in
the speech intelligibility tasks generally showed good performance in
all other tasks. Listeners who showed normal ITD/ILD discrimination
could not always exploit that ability in more complicated tasks (MLD
and speech-based tasks). Data were compared with the prediction of a
model [Zurek, in Acoustical Factors Affecting Hearing Aid Performance
(Allyn and Bacon, Boston, 1993)] which separates benefit of the
``better monaural ear'' and binaural interaction. Prediction was best
when individuals MLDs were used to predict the binaural interaction
term. Knowledge of the individual's binaural abilities may help
clinicians choose an appropriate rehabilitative strategy.
[Work supported by NIH.]
Session 4aSC: Thursday Morning, March 18, 1999,
Paper 4aSCb10. Rapid changes in speech parameters with a change in auditory
stimulation from a cochlear implant.
Time:
Joseph Perkell (M.I.T.-R.L.E. Cambridge, MA),
Lekisha Jackson (Anderson Consulting, Houston, TX)
Majid Zandipour,
Jennell Vick, and Harlan Lane (M.I.T.-R.L.E., Cambridge, MA), and
Melanie Matthies (Hearing Res. Cen., Boston U. and M.I.T.-R.L.E., Cambridge, MA)
To investigate the speed of speech production changes in response to a
change in hearing, the speech processors of two cochlear implant users
were
switched on and off a number of times in a single experimental session for
each subject. The subjects repeated short utterances (``a said,'' ``a
shed,'' ``a sad,'' and ``a shad'') many times in semi-random order. The
transitions between hearing (on) and nonhearing (off) states were timed to
occur between utterances, and the number of utterances between transitions
was varied to minimize subject anticipation of the change. Using
transition
times as line-up points, values of median and skewness of sibilant spectra
and vowel F1, F2, duration, and SPL were each averaged over repetitions of
each utterance and compared across the on--off and off--on transitions.
Vowel SPL and duration had changed by the first utterance following the
transition, indicating that the subjects were using subtle aspects of the
processor output to detect its state even in the absence of overt sound
input. Changes in spectral median and skewness and F1 and F2 were less
immediate, more varied, and differed between the two subjects. The
paradigm
is currently being refined to further investigate this issue.
[Work supported by NIH.]
Session 5aPP: Friday Morning, March 19, 1999,
Paper 5aPPa11. Effects of the spatial distribution of competing sounds on speech
reception thresholds.
Time: 11:40
John F. Culling, Monica L. Hawley, and Ruth Y. Litovsky
(Hearing Res. Ctr. and Dept. of Biomed. Eng.,
Boston Univ., Boston, MA)
The experiment tested whether listeners can suppress competing sound from
several different locations simultaneously. Speech reception thresholds
were measured adaptively in anechoic virtual auditory spaces with a target
voice directly ahead and competing sounds in four spatial configurations.
Three concurrent competing sounds were distributed on the horizontal plane
among consecutively numbered positions at 30-degree intervals, position 4
being directly ahead, 1 far left, and 7 far right. The competitor
configurations were 444, 367, 567, and 777. Four types of competing sound
were used: speech (same voice), reversed speech, speech-shaped noise, and
speech-modulated, speech-shaped noise. The patterns of results across
spatial configurations were similar for each type of competitor, although
binaural advantage (777 vs 444) was 2--3 dB greater for speech competitors
than for noise. Speech gave thresholds 3 dB higher than reversed speech
throughout, suggesting linguistic interference effects. Most notably,
however, 567 and 777 always gave similar thresholds (7--10 dB lower than
444). The multiple source locations of 567 produce a sound field with
relatively low interaural coherence, but binaural unmasking experiments
indicate that masking release demands maskers of high within-channel
interaural coherence. Multiple competing sources on both sides (367) were
more difficult to suppress.
[Work supported by MRC and NIH.]