M.Kathleen Pichora-Fuller -
kpf@audiospeech.ubc.ca
Bruce A. Schneider, Hollis Pass
School of Audiology and Speech Sciences, University of British Columbia
Vancouver, BC, Canada, V6T 1Z3
Popular version of paper 5aPPb
Presented Friday morning,
March 19, 1999
ASA/EAA/DAGA ‘99 Meeting, Berlin, Germany
Age-related hearing loss, presbycusis, begins in middle age and increases into old age. When people are in their forties they may begin to notice that it is more difficult to listen when in a crowd. By the time people are in their seventies, about half have a clinically significant hearing loss that would warrant wearing a hearing aid because without a hearing aid quiet sounds can no longer be heard. In many ways, presbycusis is like hearing loss resulting from other causes, such as exposure to high levels of industrial noise. However, presbycusis also seems to differ from other kinds of hearing loss in an important way. The major difference of interest has been the common observation that older adults have more difficulty than younger adults when they are listening in a noisy background, such as at a cocktail party or in a cafeteria. Loss of ability to hear quiet sounds does not seem to explain this type of problem since the differences in noise are observed even when younger and older adults have similar ability to hear quiet sounds. Even older adults who do not have a clinically significant hearing loss for quiet sounds that would warrant wearing a hearing aid nevertheless experience problems when they are in noisy situations. Furthermore, making sounds louder does not necessarily help. This is of serious concern because most everyday listening in the real world takes place in noisy backgrounds and difficulty hearing in these environments can reduce participation in social activities. Over the last decade, auditory scientists have focused more and more research effort to discovering what might explain this problem of older listeners.
It seems that the problem of listening in noise must involve a hearing ability that plays a special role when we listen to sounds such as speech that are moderately loud. People usually have no difficulty knowing that someone is speaking but older listeners complain that speech becomes unclear or jumbled -- they can hear but don't understand. In searching for the explanation for loss of clarity for sounds that are loud enough, researchers have begun to study how the auditory system codes the timing of sounds.
In the present study, we set out to mimic how we think the aging auditory system loses the ability to precisely code the timing of sounds. Our hypothesis was that an important aspect of presbycusis is loss of synchrony in the firing of neurons in the auditory system. We attempted to simulate such a loss of synchrony by recording sentences and then disrupting the time pattern of the sentences by jittering them. How did we jitter the sentences? We used a computer to record the sentences. The computer sound files store points of information about the soundwave 20,000 times per second. For the unjittered sentences, the sound was played out with no change in the timing of information points as they were originally recorded. We created the temporally jittered sentences by delaying the play-out timing of the stored information points. For each point of information, we could vary the delay in play-out time by small or large amounts and we could also vary how often we reset the amount of delay in the play-out of the information points. This is captured by a mathematical model where if x(t) is the input signal, the internal representation of this signal in the auditory system, y(t), is assumed to be a time-delayed version of the input with the time delay, d, varying over time as y(t) = x[t-d(t)]. The first factor contributing to the jitter is the range of delays that might occur over time, modeled as the standard deviation (RMS) of a band-limited noise. The second factor is the rate at which delays change, modeled as the frequency bandpass (BP) of a band-limited noise.
Intact and three different kinds of jittered sentences from a standard test were presented in two different levels of background noise and sentence-final word recognition was measured in young listeners with good hearing to see if they would demonstrate the same patterns of difficulty hearing speech in noise that we know older adults have. Word recognition for the intact sentences (soundfile 1) was essentially the same as for jittered sentences (soundfile 2) when the second factor, the frequency with which we changed the amount of delay, dominated the jitter (RMS = .05 ms, BP = 500 Hz). Word recognition was reduced slightly (10%) when the first factor (soundfile 3), the range of delays used, dominated (RMS = .25 ms, BP = 100 Hz). Word recognition (soundfile 4) was reduced greatly (50%) when both factors contributed to the jitter (RMS = .25 ms, BP = 500 Hz). The similarity between these results and the results found previously for older listeners is consistent with the idea that temporal asynchrony may be an important feature in presbycusis.