Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Temporal jitter produces PI-PB rollover in young normal-hearing listeners Miranda, Terence T. 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-0490.pdf [ 25.85MB ]
Metadata
JSON: 831-1.0089575.json
JSON-LD: 831-1.0089575-ld.json
RDF/XML (Pretty): 831-1.0089575-rdf.xml
RDF/JSON: 831-1.0089575-rdf.json
Turtle: 831-1.0089575-turtle.txt
N-Triples: 831-1.0089575-rdf-ntriples.txt
Original Record: 831-1.0089575-source.json
Full Text
831-1.0089575-fulltext.txt
Citation
831-1.0089575.ris

Full Text

T E M P O R A L JITTER PRODUCES PI-PB R O L L O V E R IN Y O U N G N O R M A L - H E A R I N G LISTENERS by T E R E N C E T. M I R A N D A B.Sc. (Eng.), Queen's University, 1994  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF  M A S T E R OF SCIENCE in THE F A C U L T Y OF MEDICINE (School of Audiology and Speech Sciences)  We accept this thesis as conforming to the required standard  UNIVERSITY OF BRITISH C O L U M B I A October 2000 © Terence T. Miranda, 2000  In  presenting  this  degree at the  thesis  in  University of  partial  fulfilment  of  of  department  this or  thesis for by  his  or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  representatives.  an advanced  Library shall make it  agree that permission for extensive  scholarly purposes may be her  for  It  is  granted  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  0c\.\,  ^000  11  ABSTRACT  Older listeners, even those with audiograms in the normal range, often have more difficulty than younger listeners when listening to speech in background noise. It is thought that one of the reasons for their difficulty may be due to a temporal processing deficit such as a lack of neural synchrony. The present study investigates whether temporal jitter can simulate neural asynchrony in young, normal-hearing listeners. Sixteen young normal-hearing participants were tested with intact and temporally jittered NU-6 word lists in quiet at 40, 55, 65, and (UCL-5) dBHL. Results show significant PI-PB rollover in the jittered but not the intact condition. Results support the claim that temporal jitter simulates neural asynchrony and help explicate the neural basis for PI-PB rollover. Results are consistent with the hypothesis that synchrony coding plays a role in the perception of high-level speech.  T A B L E OF CONTENTS ABSTRACT  ii  T A B L E OF CONTENTS  iii  LIST OF T A B L E S  ix  LIST OF FIGURES  x  ACKNOWLEDGEMENTS  xi  1.0  LITERATURE REVIEW  1  1.1  Introduction.  1  1.2  Speech Perception in the Elderly  3  1.3  Temporal Processing in the Auditory System  6  1.3.1  Neural Synchrony  6  1.3.2  The Involvement of Neural Synchrony in the Coding of Frequency and Intensity  7  1.3.2.1 Frequency Encoding and Neural Synchrony  7  1.3.2.2 Intensity Coding and Neural Synchrony  9  1.3.3  1.4  Measurements of Temporal Processing Ability: Evidence of Neural Asynchrony in the Elderly  9  1.3.3.1 Frequency Discrimination  9  1.3.3.2 Binaural Unmasking  10  1.3.3.3 Monaural Gap Detection  11  1.3.4  The Importance of Temporal Processing in Understanding Spoken Language 13  1.3.5  The Importance of Neural Synchrony in Understanding Speech in Noise  Intensity Transmission through the Auditory System  14 15  1.4.1  Outer Ear  16  1.4.2  Middle Ear  16  1.4.3  Inner Ear  17  1.4.4  Auditory Nerve  18  1.4.4.1 Anatomical Organization  19  1.4.4.2 Spontaneous Activity and Neuronal Thresholds  19  1.4.4.3 Theories of Intensity Coding  20  1.4.4.4 Encoding of Speech by Primary Auditory Neurons.. 20 1.4.4.4.1  Rate-Place Information  21  1.4.4.4.2  Synchrony-Place Information  22  1.4.4.4.3  Rate and Synchrony Coding: The ALSR  22  Synchrony Suppression and the ALSR  23  Limitations of the A L S R and Rate-Place Models in Explaining the Perception of Loud Speech  24  1.4.4.4.4  1.4.4.4.5  1.5  1.6  Word-Recognition Testing  25  1.5.1  Development of Word-Recognition Lists  25  1.5.1.1 Development of the CID W-22 Word Lists  26  1.5.1.2 Development of the NU-6 Word Lists  26  1.5.2  Factors Affecting W R T Scores  27  1.5.3  Performance-Intensity Curves for PB Word Lists  29  PI-PB Rollover  30  1.6.1  32  Physiologic Basis for Rollover  1.6.1.1 Mechanical Basis  32  1.6.1.2 Neural Basis  35  1.6.1.2.1  Evidence from Patients with an Acoustic Neuroma  35  1.6.1.2.2  Evidence from MS Patients  37  1.6.1.2.3  Evidence from the Elderly Population  38  1.6.1.3 Neural-Mechanical Basis  2.0  3.0  39  1.7  Summary  40  1.8  Hypotheses  44  THE PILOT S T U D Y  48  2.1  Purpose  48  2.2  Method  48  2.2.1  Participants  48  2.2.2  Materials  49  2.2.2.1 Jitter Methods  51  2.2.2.2 Calibrating the Sound Level of the Stimuli  54  2.2.3  Apparatus and Physical Setting  54  2.2.4  Procedure  55  2.3  Results  58  2.4  Changes made to the Pilot Procedure to Develop the Experimental Procedure  61  THE MAIN EXPERIMENT  65  3.1  Method  65  3.1.1  65  Participants  4.0  3.1.2  Materials  65  3.1.3  Apparatus and Physical Setting  67  3.1.4  Experimental Design  68  3.1.5  Testing Procedure  70  3.1.6  Scoring Procedure  73  RESULTS OF THE M A I N E X P E R I M E N T  74  4.1  Introduction  74  4.2  Results of the Main Experiment  74  4.2.1  Percent Correct Scores  74  4.2.1.1 Effect of Group  75  4.2.1.2 Effect of Jitter Condition  76  4.2.1.3 Effect of Presentation Level  76  4.2.1.4 Interaction of Jitter Condition and Presentation Level  77  4.2.1.5 Interaction of Group, Presentation Level and Jitter Condition  79  PBmax  80  4.2.2.1 Effect of Group  81  4.2.2.2 Effect of Jitter Condition  81  Rollover  82  4.2.3.1 Effect of Group  82  4.2.3.2 Effect of Jitter Condition  83  Correlations  85  4.2.2  4.2.3  4.2.4 4.3  Word Errors  86  4.4  5.0  Summary of Results  88  DISCUSSION A N D CONCLUSIONS  89  5.1  Review of Hypotheses  89  5.2  Summary of Results  90  5.2.1  5.2.2  5.2.3  Null Hypothesis 1: PI-PB Rollover in the NU6intact Condition  90  Null Hypothesis 2: PI-PB Rollover in the NU6intact versus the NU6jitter Conditions  91  Null Hypothesis 3: PBmax in the NU6intact versus the NU6jitter Conditions  91  5.2.4  Null Hypothesis 4: PI-PB Rollover in the NU6jitter versus the W22jitter Conditions 92  5.2.5  Null Hypothesis 5: Correlations between PBmax and Participant Variables  5.2.6  94  Null Hypothesis 6: Correlations between PI-PB Rollover and Participant Variables  94  5.3  Conclusions and General Discussion  95  5.4  Future Research Directions  97  REFERENCES  99  APPENDIX A: NU-6 Word Lists  107  APPENDIX B: Jitterl4.exe Main Screen  108  A P P E N D I X C: Hearing and Language History Form  109  APPENDIX D: Spectrum of a Jittered 1 kHz Tone  110  APPENDIX E: Informed Consent Form  113  viii  APPENDIX F: W-22 Word Lists  116  APPENDIX G: Loudness Levels  117  APPENDIX H: Raw Results  118  APPENDIX I: Individual Subject Characteristics  145  APPENDIX J: Participant Responses for Most Common NU-6 Word Errors  147  LIST OF T A B L E S  Table 2.1  Pilot study word list presentation order  57  Table 3.1  Schedule for the intact and jittered NU-6 conditions  68  Table 3.2  Schedule for the jittered W-22 condition  69  Table 4.1  Mean percentage scores  78  Table 4.2  Word errors per list, per condition  86  Table 4.3  Most common NU-6 word errors  87  LIST OF FIGURES  Figure 2.1  Rollover (%) for each pilot participant in each jitter condition  58  Figure 2.2  PBmax (%) for each pilot participant in each jitter condition  59  Figure 4.1  Mean percent score at each presentation level at each jitter condition  79  Figure 4.2  Individual PBmax score in each jitter condition  80  Figure 4.3  Mean PBmax score in each jitter condition  81  Figure 4.4  Individual rollover (%) in each jitter condition  83  Figure 4.5  Mean rollover (%) in each jitter condition  84  ACKNOWLEDGEMENTS I would like to thank my thesis supervisor, Kathy Pichora-Fuller for her guidance during this project. I would also like to thank the other members of my thesis committee, Jeff Small and Noelle Lamb, for agreeing to provide their support despite their busy schedules. Thanks go to Carol Jaeger and Trudy Adam for their assistance. A special thank-you to Dennis Phillips for his encouragement at a time when it was needed most. I would like to thank the people living at and involved with Green College; they helped to make these last two years memorable. This research was supported by a grant from the Natural Sciences and Engineering Research Council of Canada.  1  1.0 L I T E R A T U R E REVIEW  1.1  Introduction In general, elderly listeners, even those with audiograms in the normal range,  have more difficulty than younger adults when trying to understand language spoken in background noise (e.g., C H A B A , 1988). One of the possible explanations for the poor performance of these elderly listeners is the existence of a temporal deficit (e.g., neural asynchrony) in their ability to process auditory information (e.g., Pichora-Fuller & Schneider, 1992). In recent studies, high- and low-context SPBSf-R sentences were temporally jittered and used to test young, normal hearing adults in a range of signal-tonoise (S/N) ratios (Pass, 1998; Brown, 2000; Pichora-Fuller, Schneider, Pass & Brown, submitted). The performance, especially in low S/N ratios, of the young normals resembled the performance on intact (unjittered) SPIN-R sentences by older adults with pure-tone audiograms in the normal range. One of the conclusions from these studies was that the external jitter introduced into the stimuli seems to simulate, in young normals, the increased internal temporal jitter hypothesized to exist in the elderly. This thesis tackles the question of whether or not external, temporal jitter simulates internal, neural asynchrony in young, normal hearing adults. The problem is approached by testing the hypothesis that the human auditory system is more dependent on synchrony information at high, versus low, levels of signal intensity in quiet. This first chapter will begin by discussing three general hypotheses that have been proposed to explain the speech perception difficulties of the elderly. The central auditory processing argument, and in particular temporal processing and neural synchrony, will be  2  expanded upon in the following sections. Behavioural evidence that suggests there is a greater degree of neural asynchrony in the older adult's auditory system will be presented. Next, the importance of temporal processing in understanding spoken language and the importance of neural synchrony in understanding speech in noise will be discussed. The literature review up to this point helps to bring out and explain the significance of earlier work. The remainder of the chapter outlines the additional concepts that are important to this thesis in five different sections. The first of these sections includes an explanation of what is known, and what is theorized, about the transmission of intensity information through the auditory system from the outer ear up to, and including, the auditory nerve. Emphasis is placed on the coding of speech in the auditory nerve. The average localized synchronized rate (ALSR), a critical theory that accounts for neural coding of both rate and synchrony information, is explored in depth. The second section includes a description of word-recognition testing, outlining the development of some of the testing materials used in this thesis, and defining the performance-intensity function for phonetically balanced word lists (PI-PB function). The third section explicates the clinical phenomenon of PI-PB rollover. The terms rollover and rollover ratio (Rx) are defined and physiologic explanations for rollover are examined. The possible neural basis of rollover is explored in depth and evidence linking rollover and neural asynchrony is presented. The penultimate section will summarize the important findings in this chapter. Finally, the null and corresponding research hypotheses tested in the present thesis will be stated.  3  1.2  Speech Perception in the Elderly Elderly listeners often have more difficulty than younger listeners when trying to  understand speech in everyday listening situations (e.g., Pichora-Fuller, 1997). Three general hypotheses have been proposed in order to explain this finding ( C H A B A , 1988). One line of reasoning is that difficulties are due to the elevated audiometric thresholds, especially at high frequencies, which usually accompany the aging process (e.g., Humes, 1996). Another hypothesis is that speech perception difficulties are due to the cognitive slowing that occurs with age (e.g., Wingfield, 1996). The third hypothesis is that a decline in central auditory processing ability, independent of pure-tone hearing loss and cognitive slowing, is responsible for speech perception problems among the elderly (e.g., Chmiel & Jerger, 1996). Although both the cognitive and central-auditory hypotheses involve higher centers in the auditory pathway, the cognitive hypothesis extends to other sensory modalities (e.g., vision) besides the auditory modality (e.g., Humes, 1996). Elevated pure-tone audiometric thresholds and decreased cognitive ability are undoubtedly responsible, at least to some degree, for age-related problems in understanding spoken language. However, these two explanations, in isolation or combination, can not explain all of the speech perception and language comprehension problems found in the elderly population. Elevation of pure-tone audiometric thresholds can not fully account for the difficulty of the elderly in listening to speech in noise (e.g., Willot, 1991). This is not surprising given the fact that there can be substantial histopathological change in the auditory periphery without a significant elevation in pure-tone thresholds (for a review see Willot, 1991). When matched with young listeners with equivalent audiometric  4  thresholds, older listeners demonstrate poorer performance for language presented at fast speaking rates (e.g., Gordon-Salant & Fitzgibbons, 1995), speech in noise (e.g., Dubno, Dirks & Morgan, 1984) and speech distorted by reverberation (e.g., Nabelek & Robinson, 1982). These findings fit with the common clinical observation that elderly individuals, with little or no clinically significant pure-tone threshold loss, often report that it is effortful to listen to speech in everyday listening situations (CHABA, 1988; PichoraFuller, Schneider & Daneman, 1995). The study by Pichora-Fuller et al., (1995) examined the word identification and recall performance of young and old adults. In the first part of the study, participants had to identify sentence-final words in different signal-to-noise (S/N) ratios. Psychometric functions (percent correct word identification as a function of S/N ratio) were obtained in conditions when the sentence-final words were predictable (high-context) and unpredictable (low-context). The authors used working memory span as a measure of cognitive ability. In the high-context situations, participants could exercise their working memory (cognitive ability) to store information from the beginning of the sentence and use that to help resolve the sentence-final word. In the low-context situations, this was not possible. Thus, it was reasoned that any possible differences in cognitive ability between young and old participants would not affect the results obtained in low-context conditions. The authors expected that the psychometric functions for old adults with near normal audiometric thresholds and for young normal-hearing adults would be identical in low-context conditions if the only age-related differences were cognitive. However, what they found was that the older participants performed worse than the younger participants in the low-context conditions. The authors concluded that there is an age-related  5  deterioration in perception that can be modeled by assuming an increase in the level of internal noise. In the case of significant audiometric loss, perceptual processing would be even more impaired. Pichora-Fuller et al.'s (1995) finding of an age-related deterioration in speech perception not attributable to elevated audiometric thresholds or decreased cognitive ability, and the suggestion that the deterioration can be modeled by assuming an increase in internal noise, is highly supportive of a central auditory processing argument. The central auditory processing argument states that decreased speech perception in the elderly is due to declines in how accurately and efficiently the auditory system analyses the temporal, intensity, and spectral characteristics of the acoustic input, other than declines due to threshold hearing loss or cognitive slowing (see also C H A B A , 1988; Haubert, 1999; Humes, 1996). These three hypotheses concerning speech perception decline in the elderly (elevated pure-tone thresholds, cognitive decline, and reduced central auditory processing ability) are not necessarily independent ( C H A B A , 1988). Depending on the task, all three, or various combinations of the three, may contribute to the difficulties understanding spoken language that are experienced by any particular elderly individual. However, none of the explanations, on its own, can account for all of the speech perception problems in the elderly population. The central auditory processing argument, and, more specifically, temporal processing in the auditory nerve, is fundamental to this thesis and will be explored more fully in the next section.  1.3  Temporal Processing in the Auditory System Temporal processing, or auditory temporal resolution, refers to the auditory  system's ability to detect changes in stimuli over time (Moore, 1989). Given this loose definition and the fact that the acoustic signal is nothing more than changes in amplitude over time, one may ask "what isn't temporal processing?" Taken in its broadest sense, temporal processing is an all encompassing topic. For reviews see Moore (1989), Phillips (1995), Greenberg (1996) and Schneider (1997). A l l of these authors acknowledge the importance of differentiating different types or levels of temporal processing. One important type of temporal processing involves neural synchrony, or periodicity coding.  1.3.1  Neural S ynchron y Neural synchrony refers to the ability of auditory nerve fibres to discharge to one  phase of an incoming stimulus (phase lock). Yost (1994) notes that a measure of neural synchrony (coefficient of synchrony) can be derived from phase-locked post-stimulus time (PST) histograms (also called period histograms). These histograms indicate how well neural patterns of firings conform to the stimulus waveform period. If, in response to a pure tone, the discharge pattern in the period histogram follows the top half of the sine wave (i.e., a phase angle from 0 to 180 degrees), then the measure of neural synchrony is close to perfect (1.0). If the neural discharges are evenly distributed throughout the period of stimulation (i.e., they are flat, covering a phase angle from 0 to 360 degrees), then the coefficient of synchrony is nearly zero.  7  1.3.2  The Involvement of Neural Synchrony in the Coding of Frequency and Intensity The next section discusses the role of neural synchrony in the basic auditory  process of encoding frequency. The following section (Section 1.3.2.2) gives a brief account of the possible role of neural synchrony in encoding the relative intensities of components in complex signals. Because the transmission and encoding of intensity in the auditory system is of crucial importance to this thesis, all of Section 1.4 is devoted to exploring this topic.  1.3.2.1 Frequency Encoding and Neural Synchrony The frequency of a stimulus determines the place along the basilar membrane (BM) at which the traveling wave vibrates with maximum amplitude. Thus, according to place theory, the frequency of the input stimulus is determined by which auditory neuron (identified by its location along the BM) fires with the greatest relative discharge rate (Yost, 1994). The B M is tonotopically organized; there is a spatial representation of frequency according to place. The maintenance of the neural spatial representation of frequency throughout the nuclei of the central auditory pathways is strong evidence in support of place theory. However, place theory alone may be insufficient for explaining all auditory perception. Three examples of where place theory is inadequate include the case of the missing fundamental, perception of speech in high levels of background noise, and perception of loud speech. A l l three of these situations may be explained when neural synchrony in the auditory system is considered. Recordings obtained from auditory nerve discharges show that auditory nerve fibers fire with timing corresponding to the period of the input stimulus when the input  8  frequency is less than 5 kHz. Volley theory states that frequencies less than 5 kHz can be encoded by the periodicity of nerve fibre discharges. The case of the missing fundamental refers to a complex waveform, composed only of harmonics (and not of the fundamental frequency itself), which is perceived at the pitch of the fundamental (e.g., a complex waveform composed of 0.6, 0.9 and 1.2 kHz sine waves is perceived at a pitch equivalent to the pitch of a 0.3 kHz waveform). Since the characteristic place along the B M associated with 0.3 kHz is not stimulated, place theory fails to explain this phenomenon. It should be noted that, although the complex waveform has no spectral energy at the fundamental frequency, the waveform does have a periodicity at the fundamental (e.g., Phillips, 1995). One way of explaining the case of the missing fundamental is through volley theory. That is, the periodicity of the stimulus waveform can be represented in the timing of the spike action potentials it generates within some neural population (e.g., Phillips, 1995). With respect to speech, there may be two situations where place theory alone is insufficient to explain perception. In intense noise, it may be exceedingly difficult to distinguish the portion of the place representation on the B M that is driven by the target signal from that which is driven by the background noise (e.g., Greenberg, 1996, see Section 1.3.5). The second situation is that of perception of loud speech (see Section 1.4.4.4. for discussion). Both of these situations can be explained when a measure of neural synchrony is considered in combination with rate of neural firing and place theory (i.e., the average localized synchronized rate function, see Section 1.4.4.4.3).  1.3.2.2 Intensity Coding and Neural Synchrony In general, intensity is assumed to be encoded by an increase in neural discharge rate within the auditory system (e.g., Yost, 1994). However, animal studies have shown that, when stimulus level is high, the firing rates of neurons saturate and the relative intensity of frequency components of the signal, such as formant information, is lost (e.g., Young & Sachs, 1979). As noted above, the ability to extract formant information for such stimuli can be explained when a measure of neural synchrony is taken into account.  1.3.3  Measurements of Temporal Processing Ability: Evidence of Neural Asynchrony  in the Elderly Auditory temporal resolution can be measured using monaural tasks (e.g., gap detection, frequency discrimination, temporal ordering, speech perception with compressed/expanded speech rates) or binaural tasks (e.g., localization, binaural unmasking). In general, elderly listeners have poorer performance than younger listeners on numerous measures of temporal resolution, even when they do not have significantly elevated pure-tone thresholds in the speech range (e.g., Pichora-Fuller et al., submitted; Schneider, 1997). The next section will explore some of the temporal processing tasks that are suggestive of an increase in temporal asynchrony in the aging auditory system.  1.3.3.1 Frequency Discrimination Frequency discrimination tasks determine the smallest frequency difference (difference limeri) that a listener can detect. Konig (1957) found that frequency difference  limens increased linearly between 25 and 55 years of age before becoming markedly worse at older ages. Abel, Krever, and Alberti (1990) found increased frequency difference limens for older listeners with good hearing. Moore and Peters (1992) reported that, for older listeners with normal hearing, frequency difference limens were very large at the low frequencies, as compared to the high frequencies. Schneider (1997) points out that the reverse is found in young participants: their frequency discrimination ability is better at the low frequencies than the high frequencies. Schneider (1997) suggests that, because neural impulses are phase locked to low frequency stimuli (see Section 1.3.2.1), the disproportionate effects of age on discrimination at low frequencies are consistent with a loss of phase-locking or synchrony in the aged ear.  1.3.3.2 Binaural Unmasking Binaural unmasking refers to the ability of the auditory system to discriminate a signal from background noise ("unmask" it) using interaural cues. Binaural unmasking ability can be demonstrated in the laboratory by measuring masking-level differences or MLDs (for reviews, see Grose, 1996; Yost, 1994). The stimuli in M L D tasks usually consist of a masking noise (N) and a target signal (S). In general, diotic configurations occur when identical stimuli are presented to both ears and dichotic refers to different stimuli presented to the two ears. Experiments have shown that the detectability of a signal is better in dichotic versus diotic conditions when the signal is presented in diotic masking noise. This difference in threshold, between the dichotic and diotic signal conditions, expressed in decibels, is called an M L D . The larger the M L D , the greater the binaural unmasking ability.  11  Several configurations of stimuli exist in the M L D task. For example, in the NoS7t condition, the masker is diotic and the signal is presented to one ear is 180 degrees out of phase with the signal presented to the other ear. The M L D in this condition is, on average, about 15 dB at 0.5 kHz in normal listeners (Yost, 1994, pg. 180). This example shows that the intact auditory system is capable of comparing binaural inputs such that the masking effects of a noise in one ear are partially cancelled by the identical noise in the opposite ear (Schneider, 1997). Other conditions that show a positive M L D in normal-hearing listeners include MTCSO, MoSm (diotic masker, signal to one ear only), and MjtSm. Elderly listeners have much smaller M L D s than do younger listeners (e.g., Grose, Poth & Peters, 1994; Pichora-Fuller & Schneider, 1991; Pichora-Fuller & Schneider, 1992). Because binaural processing is very sensitive to interaural time differences, any loss of auditory temporal resolution would be expected to reduce the size of the M L D (Durlach, 1972). Based on the pattern of age-related differences in MLDs, Pichora-Fuller and Schneider (1992) argue that there is a decrease in neural synchrony in the aging auditory system.  1.3.3.3 Gap Detection (Monaural) The smallest silent interval between two sounds (markers) that a listener can detect is determined using the gap detection task. Studies by Schneider, Pichora-Fuller, Kowalchuk, and Lamb (1994) and Snell (1997) have shown that older listeners, when matched to younger listeners with respect to audiometric hearing thresholds, have poorer gap detection thresholds than their younger counterparts. Schneider and Hamstra (1999)  12  have shown that there are significant age effects on gap detection thresholds for short marker durations (e.g., 2.5 ms) but not for long marker durations (e.g., 500 ms). In explaining their results, these authors propose that in older adults, the response to signal onsets recovers more slowly from neural adaptation and that this obscures gap detection, particularly when using short gap markers. That is, as the duration of the markers surrounding the gap is shortened, less time is available for neural recovery and therefore fewer neurons are available to mark the onset of the marker following the gap. With respect to the perception of speech, gap duration is a cue for perception of stop consonants (e.g., Dorman, Marton, Hannley, & Lindhol, 1985). It can also be used for discriminating between a fricative and an affricate (Dorman, Raphael & Liberman, 1979). Older adults, who were poorer than younger adults at detecting gaps embedded in markers of short duration, also had greater difficulty identifying words when a gap served to differentiate them, particularly for fast speech (Haubert, 1999). Older adults' reduced ability to recover from neural adaptation, as suggested by their gap detection results, may be associated with increased neural asynchrony in the aging auditory system. For example, the frequency of the markers used in Haubert's study (2 kHz) was well within the frequency range in which auditory neurons are able to code periodicity (see Section 1.3.2.1). If these neurons are less able to recover from neural adaptation, it follows that they will also be less able (individually or as a group) to fire in accordance with the periodicities of a low-frequency stimulus.  13  1.3.4 The Importance of Temporal Processing in Understanding Spoken Language The temporal characteristics of speech can be divided into two groups (Shannon, Zeng, Kamath, Wygonski & Ekelid, 1995; Viemeister & Plack, 1993): 1) the outer, temporal envelope of the speech signal which changes slowly over time; 2) the inner, fine structure of the signal which changes rapidly over time. Shannon et al., (1995) showed the importance of the temporal envelope and called into question the importance of temporal fine structure in speech perception. In one of their conditions, they divided the speech signal into three spectral regions (0-0.5 kHz, 0.5-1.5 kHz, and 1.5-4.0 kHz) and determined the amplitude envelope in each region. The amplitude envelope for each speech region was then used to modulate bandlimited noise associated with each spectral region. The three amplitude modulated noises were then added together to produce a signal in which fine structure was destroyed but whose temporal envelopes in the three regions were retained. Results showed that speech recognition for sentences processed in this way exceeded 90%. Although a number of studies (e.g., Drullman, 1995a; Turner, Souza & Forget, 1995) have illustrated the importance of the temporal envelope cues in speech perception in quiet, relatively few studies have examined the ways that fine-structure temporal cues are used in speech perception, even in noisy conditions where envelope cues may be reduced (Drullman, 1995b, Pichora-Fuller, Schneider, Pass & Brown, submitted). In recent studies to investigate the role of fine structure in speech perception in noise, temporal jitter was graded and used to alter SPIN-R sentences which were presented in high and low S/N ratios (Pass, 1998; Pichora-Fuller et al., submitted). The intent of this temporal jitter was to disrupt the effectiveness of periodicity coding (phase  14  locking) by the primary auditory neurons to the fundamental frequency and harmonic structure of the speech signal. That is, an attempt was made to introduce the effects of neural asynchrony in the auditory system by creating a less periodic acoustic stimulus. The degree of temporal jitter, the S/N ratio, and sentence context all had significant effects on participants'abilities to recognize sentence-final words. Intelligibility was worse for higher degrees of jitter. The effect of jitter on intelligibility was more pronounced for low-context sentences and at lower S/N ratios. Additionally, like older listeners who report that their conversational partners mumble, experimental participants noted that, in the most jittered condition, the speech had a mumbled quality. One of the observations made based on the results of this study was that performance on the jittered SPIN-R test by young normals resembled performance on the unjittered SPIN-R test by older listeners with pure-tone audiograms in the normal range. One of the possible explanations for the poor performance of these elderly listeners is the existence of a temporal processing deficit (e.g., neural asynchrony). Thus, it seems that the temporally jittered stimuli in Pass' study successfully simulated the effects presumed to arise from age-related neural asynchrony. This was later confirmed by Brown (2000).  1.3.5  The Importance of Neural Synchrony in Understanding Speech in Noise The effects of both reverberation and noise can be conceptualized as smoothing  out the temporal fluctuations in the waveform's envelope (Houtgast & Steeneken, 1973). Because of the evidence (e.g., Shannon et al., 1995) showing the importance of the envelope in speech perception, it is reasonable to expect that the problems of the elderly in background noise/reverberation may be attributable, to some extent, to the effect these  15  conditions have on the waveform's envelope. However, the importance of the ability of the elderly to process the fine structure of speech in the presence of background noise (e.g. through neural synchrony) should not be ignored when trying to understanding their difficulty perceiving speech in these conditions. Intense noise containing energy overlapping with frequency regions important for speech can damage the auditory representation of the speech spectrum (Greenberg, 1996). In rate-place models (see Section 1.4.4.4.1), there is no systematic way of separating the neural activity associated with the signal from that associated with the noise (Greenberg, 1996). Models that take a measure of temporal synchrony into account (see Section 1.4.4.4.2 and 1.4.4.4.3) have an advantage because temporal (e.g., fine structure or signal periodicity) information is more robust in the presence of noise than firing rate information (Greenberg, 1996). Thus, there is reason to believe the recent findings (e.g., Pichora-Fuller et al., submitted) which suggest the importance of the role of fine structure in speech perception in noise.  1.4  Intensity Transmission through the Auditory System The purpose of this section is to summarize current knowledge and proposed  theories on how the properties of sound are analyzed and transmitted in the auditory pathway up to, and including, the auditory nerve. Emphasis will be placed on intensity coding. However, as will be seen, the coding of intensity often depends on frequency. The contributions of the outer, middle, and inner ear, and the auditory nerve to this coding will be described.  16  1.4.1  Outer Ear When sound is produced, waves of air molecules are propagated towards the ear.  Sound travels through the cavities of the outer ear towards the eardrum, setting it into vibration. Resonances in the outer ear amplify sound frequencies; this pattern of amplification can be described by a transfer function (e.g., Yost, 1994). Although this transfer function is unique to each individual, for the average adult there is a 10-20 dB amplitude gain in the 1.5 to 7.0 kHz frequency range, with a peak at 3.4 kHz (Yost, 1994). This transfer function also depends on the azimuth and elevation of the sound source.  1.4.2  Middle Ear Vibration of the eardrum sets the middle ear ossicles (malleus, incus and stapes)  into motion. The ossicular chain provides an intensity boost of up to 33 dB, depending on frequency. The greatest intensity gain is in the frequency range from 1-3 kHz. The middle ear ossicles are suspended in the tympanum of the middle ear cavity by axial ligaments. The tendon of the tensor tympani muscle is attached to the manubrium of the malleus, and the tendon of the stapedius muscle is attached to the head of the stapes. When the ear is excited by a loud sound, these muscles contract, dampening the transmission of pressure through the ossicular chain. Yost (1994) reports that the reduction in transmission is approximately 10 to 30 dB for loud sounds and is frequency-dependent, having more effect for frequencies below 2 kHz. For the stapedial reflex, attenuation is thought to be between 10 and 15 dB for the frequencies below 0.1 kHz and 0-6 dB for the frequencies above 0.1 kHz (Borg, 1968; Wormald, Rogers & Gatehouse, 1995).  17  1.4.3  Inner Ear Vibrations are passed on to the cochlea via vibration of the stapes at the oval  window. Von Bekesy (1960) was the first to demonstrate the mechanics of the passive (dead) cochlea. He showed that the passive cochlea codes frequency according to place. A travelling wave, or surface acoustic wave, is set up along the basilar membrane (BM) in the cochlea. The amplitude of vibration varies with location depending on the frequency and intensity of the incoming stimulus. The amplitude of the travelling wave grows as it moves along the B M and peaks at the characteristic frequency (CF) location. For high frequencies, this occurs near the base of the B M and for low frequencies, peak amplitude is achieved near the apex. Beyond the C F location, the traveling wave dies out rapidly. The passive cochlea acts like a low-pass filter; any specific place on the B M responds not only to its C F but also to lower frequencies and to a much lesser extent, higher frequencies. The higher the intensity of the input, the greater the spread of activation on the basilar membrane. The active (living) cochlea acts more like a bandpass filter than a lowpass filter. In vitro and in vivo experiments show that when the cochlea is in good physiologic condition, the B M is very sensitively and sharply tuned at the C F place (for reviews, see Dallos, 1996; Geisler, 1998). In the active cochlea, the travelling wave response is composed of a broadly tuned component (passive mechanics) and a sharply tuned component (active mechanics). It is possible to explain active cochlear mechanics only if it is assumed that there is an active source of mechanical energy that amplifies the travelling wave in a limited region on its basal slope. It is proposed that in this active region, prior to the point of resonance, outer  18  hair cells (OHCs) feed mechanical energy into the travelling wave, making it grow. This in turn produces more stimulation of the OHCs leading to a wave that grows more and more steeply. Soon the wave passes the active region and its amplitude drops quickly. Another feature of the active cochlea is that the B M responses grow nonlinearly with intensity for stimulus frequencies around the CF; the active component decreases with increasing amplitude. At high intensities, active processes saturate and B M displacements become linear, as in the passive cochlea. Thus, in the active cochlea, B M displacements at the C F place are non-linearly amplified in response to low-intensity stimuli but respond linearly to high-intensity stimuli. Mechanical deformation of the B M produces an electrical potential in the hair cell whose magnitude is proportional to the amount of stimulation (e.g., Yost, 1994). The hair cell then becomes depolarized and a neurotransmitter is released which initiates a graded electrical potential. This potential is propagated to the dendrites of the spiral ganglion cells where generation of a neural action potential occurs. The action potential is an all or none response and does not vary with the level of the stimulus, i.e., it is not a graded response (e.g.,Yost, 1994). Due to mechanical coupling to the B M , hair cell activation is directly influenced by the non-linear processes of the B M (e.g., Patuzzi, 1996).  1.4.4  Auditory Nerve This section outlines the organization of fibres in the auditory nerve and goes into  details on spontaneous activity and the neuronal thresholds of primary afferent fibres. Theories of intensity coding in the auditory nerve are reviewed and then explanations of how the primary afferent fibres encode speech are investigated. The A L S R is defined, its  19  ability to account for the encoding of loud speech is explained using the concept of synchrony suppression, and the limitations of the A L S R as a theory are noted.  1.4.4.1 Anatomical Organization The auditory nerve consists of both afferent (to the brain) and efferent (from the brain) nerve fibres. There are two types of afferent fibres. Type I (radial or primary) fibres, which comprise ~ 90% of all afferents, each innervate one (or a maximum of 2) IHCs. There is a "many-to-one" relationship; there are many (8-10) Type I fibres to each IHC. Type II (outer spiral or secondary) fibres innervate OHCs in a "one-to-many" relationship; each secondary fibre innervates many OHCs. Efferent fibres to the cochlea come from the olivocochlear system and can be either crossed (coming from the opposite side of the brain) or uncrossed (coming from the same side of the brain). The medial olivocochlear system sends efferents to the OHCs and the lateral olivocochlear system sends efferents to the IHCs (for a review, see Yost, 1994).  1.4.4.2 Spontaneous Activity and Neuronal Thresholds Auditory neurons exhibit spontaneous activity. That is, neurons fire in the absence of any acoustic stimulus. Spontaneous activity in single auditory neurons may range from a few spikes per minute to more than 100 spikes per second. Using their level of spontaneous activity, neurons are classified as low-, medium- or high-spontaneous rate fibres. The neuronal threshold, the minimum dB that causes an increase in firing rate above the spontaneous rate, is lowest for high-rate neurons and highest for low-rate neurons. A neuron's firing rate will increase above its threshold up to a certain level, at  20 which firing rate plateaus. This range, from threshold to plateau, is called the neuronal dynamic range and is 20-50 dB for all neurons. By combining the neuronal dynamic range of low-, medium- and high-spontaneous rate fibres, the 140 dB intensity range to which humans are sensitive could be covered (for a review, see Yost, 1994).  1.4.4.3 Theories of Intensity Coding Intensity is assumed to be encoded by an increase in discharge rate by auditory nerves (Yost, 1994). However, the maximum neuronal dynamic range for an individual auditory nerve fibre is 50 dB and this is not enough to cover the 140 dB range of humans. This has been called the dynamic range problem and two models have been proposed in order to address it (Delgutte, 1996). The spread of excitation model states that greater spread of activation along the B M , at higher levels of input, signals higher intensities. The tuning curve of an auditory neuron typically shows a steep high-frequency side and a shallow sloping low-frequency "tail". These tuning curves evidence the ability of auditory neurons to respond to intense stimulation at frequencies that are lower than their own CF. The place model proposes that intensity is encoded for specific place along the B M by combining the neuronal dynamic range of low-, medium-, and high-spontaneous rate neurons.  1.4.4.4 Encoding of Speech by Primary Auditory Neurons For practical, ethical reasons, there is almost no direct information about how the primary auditory neurons of humans encode speech; essentially all knowledge on this subject is obtained from laboratory mammals (Geisler, 1998).  21  1.4.4.4.1 Rate-Place Information Young and Sachs (1979) have shown that, in a cat, the population response (the average rate response of a large number of auditory nerve fibers) to a complex sound (e.g. [e]) well represents the spectrum of that sound if the stimulus level is low. However, when the stimulus level is high, the firing rates of the neurons saturate and the formant information is lost. Based on the original work done by Young and Sachs (1979), it may be difficult to envision how vowels are encoded by place-rate information at moderate to high intensities. However, there is evidence to suggest that a small proportion (15%) of auditory nerve fibres (ANFs) with spontaneous rates less than 0.5 spikes per second (i.e., low spontaneous rate neurons) may encode the spectral envelope on the basis of rateplace information, even at the highest stimulus levels, due to an extended neuronal dynamic range (Sachs, Winslow, & Blackburn, 1988; Blackburn & Sachs, 1990; Greenberg, 1996; Geisler, 1998). Greenberg (1996) notes that the case for a rate-place code for vocalic stimuli is equivocal at the level of the auditory nerve. He notes that at moderate-to-high intensity levels, the majority (85%) of A N F activity becomes saturated and the spectral envelope representation provided by the low-spontaneous rate neurons is less than ideal. Greenberg (1996) notes that rate-place representation may be enhanced in the cochlear nucleus and higher auditory stations relative to that observed in the auditory nerve.  22 1.4.4.4.2 Synchrony-Place Information In addition to influencing the average firing rate of ANFs, vowels also affect the temporal patterns of those firings (Geisler, 1998). Data from Miller and Sachs (1983) and Shamma (1985) illustrate the massed responses of primary ANFs across a range of CFs to the vowel in the syllable /da/. Through phase-locking, the glottal pulse was strongly represented in both high-CF and low-CF neurons, and the whole ensemble of primary ANFs was also shown to carry information about the fundamental periodicity (Geisler, 1998). Furthermore, it was shown that ANFs with CFs from 0.5-1.0 kHz synchronized primarily to the vowels first formant (~ 0.6 kHz) and ANFs around the second formant (~1.3 kHz) discharged in synchrony with F2. This shows that, because humans can usually identify a vowel on the basis of its first two formants (Lieberman, 1977), neural synchrony alone may provide sufficient information for perceptual recognition of vowels (Geisler, 1998). Additionally, when the coefficient of synchrony (see Section 1.3.1) is determined as a function of stimulus level, the resulting synchrony-level function shows an increase in synchrony over a much larger intensity range than the 50 dB neuronal dynamic range of an individual A N F ' s firing rate.  1.4.4.4.3 Rate and Synchrony Coding: The A L S R Rate-place and synchrony-place information need not be considered independently. Young and Sachs (1979) were some of the first authors to propose a model that combines the information from both sources. The average localized synchronized rate (ALSR) is a measure that combines the rate at which a neuron discharges and its synchrony. The A L S R is a computation for estimating the magnitude  23  of neural response in a given frequency channel based on the product of firing rate and temporal correlation with a predefined frequency band (Greenberg, 1996). Geisler (1998) notes that the temporal encoding of vowels, when combined with spatial encoding, provides a robust transmission of information regarding formant frequencies. The A L S R representations of A N F responses to the vowels [i], [a] and [e] clearly show the spectral peaks associated with the lower three formants, i.e., F l , F2 and F3 (Handel, 1989). Because neurons continue to respond in synchrony to a sound's periodicities at high stimulus levels where the rate has saturated, using synchrony information along with rate information preserves the neural code for speech at high intensities (Yost, 1994).  1.4.4.4.4 Synchrony Suppression and the A L S R Greenberg (1996) notes that the mechanism underlying the A L S R representation is synchrony suppression. At low intensities, neural temporal synchrony to a low frequency is generally restricted to a tonotopic region around the signal's CF. However, with increasing intensity, there is an orderly basal recruitment towards ANFs with high CFs; these higher CF fibres begin to respond in synchrony to the lower frequency stimulus. This phenomenon is accounted for by the spread of excitation model of intensity coding (see section 1.4.4.3). At higher SPLs, as noted by Young and Sachs (1979), the average-rate response saturates, degrading the rate-place representation of the formant pattern in the spectrum of the signal. The activity of fibres with CFs near the spectral peaks remain phase locked to the formant frequencies. However, fibres whose CFs lie between the spectral peaks suppress their synchrony to energy in their own spectral  24  r e g i o n . Instead, these fibres b e c o m e s y n c h r o n i z e d to the f o r m a n t f r e q u e n c i e s , m o s t t y p i c a l l y F I . B e c a u s e the a m p l i t u d e o f F I i s t y p i c a l l y 2 0 to 4 0 d B a b o v e that o f the h a r m o n i c s b e t w e e n the f o r m a n t s , w h e n the s i g n a l i n t e n s i t y b e c o m e s great e n o u g h , f o r m a n t e n e r g y contributes to the a c t i v a t i o n i n n e i g h b o u r i n g f r e q u e n c y c h a n n e l s , e s p e c i a l l y h i g h e r f r e q u e n c i e s . T h i s is a c o n s e q u e n c e o f the r e l a t i v e b r o a d t u n i n g o f l o w f r e q u e n c y f i b r e s a n d the s h a l l o w s l o p i n g l o w - f r e q u e n c y tails o f their t u n i n g c u r v e s . T h e e n d r e s u l t i s t h a t , at h i g h i n t e n s i t i e s , s y n c h r o n y s u p p r e s s i o n r e d u c e s t h e a m o u n t o f n e u r a l phase l o c k i n g to f r e q u e n c i e s other than the f o r m a n t s . In c o m p u t i n g the A L S R , the a m p l i t u d e o f e a c h f r e q u e n c y is g i v e n b y the m a g n i t u d e o f the n e u r a l r e s p o n s e s y n c h r o n i z e d to that c o m p o n e n t b y n e r v e fibres w h o s e C F s lie w i t h i n a quarter o f an o c t a v e ( G r e e n b e r g , 1996). A l t h o u g h fibres w h i c h are m o r e t h a n 1/4 o f a n o c t a v e a w a y f r o m a f o r m a n t m a y b e p h a s e l o c k e d t o it, t h e y a r e n o t i n c l u d e d i n the A L S R c a l c u l a t i o n o f that f r e q u e n c y . T h e y c o n t r i b u t e to the  ALSR  r e p r e s e n t a t i o n o f the i n p u t s p e c t r u m i n a n i n d i r e c t w a y i n that t h e y s u p p r e s s the  ALSR  response to frequencies near their o w n C F .  1.4.4.4.5 L i m i t a t i o n s o f the A L S R a n d R a t e - P l a c e m o d e l s i n E x p l a i n i n g the P e r c e p t i o n o f L o u d Speech T h e A L S R a s s u m e s there is a c e n t r a l m e c h a n i s m that c o m b i n e s rate a n d s y n c h r o n y i n f o r m a t i o n appropriate to the C F s o f the p r o j e c t i n g fibres (e.g., G r e e n b e r g , 1996). H o w e v e r , p h y s i o l o g i c a l e v i d e n c e f o r s u c h a m e c h a n i s m is l a c k i n g ( G r e e n b e r g , 1996).  With respect to the low-spontaneous rate fibre, rate-place model, one would expect that intelligibility should decline as intensity increases above 40 dB SPL. Above this level, the rate-place representation of the vocalic spectrum for most ANFs becomes much less well defined, and only the low-spontaneous rate fibre population can continue to encode spectral information. However, speech intelligibility is somewhat better at higher intensity levels (for a review, see Greenberg, 1996).  1.5  Word-Recognition Testing The most common way that audiologists describe suprathreshold hearing is with  word-recognition measures (Stach, 1998). Standard clinical word-recognition testing (also referred to as speech discrimination, word discrimination and PB-word testing) is an assessment of a person's ability to identify and repeat monosyllabic words presented at chosen suprathreshold levels (Stach, 1998). The word recognition score (WRT) is routinely administered in the basic audiologic evaluation.  1.5.1  Development of Word-Recognition Lists Original attempts at word-recognition testing (Egan, 1948) involved compiling  lists of words that were phonetically balanced (PB), i.e., words that contain all the phonetic elements of connected English discourse in their normal proportion to one another (Martin, 1994). The following sections describe the development of the two groups of PB word lists used in this thesis: the Central Institute for the Deaf (CID) W-22 word lists and the Northwestern University Auditory Test Number 6 (NU-6) word lists.  26  1.5.1.1 Development of the CID W-22 Word lists The P A L PB-50 word lists are a recording of eight lists of 50 words each. The lists were selected from a larger list of words created at the Harvard Psycho-Acoustic Laboratories (Davis, Morrical, & Harrison, 1949, as cited in Penrod, 1994). The original Harvard lists were devised to meet the criteria of being monosyllabic, of equal average difficulty, having a range of difficulty and phonetic composition for each list, and being representative of English speech (Penrod, 1994). However, based on clinical use, a number of problems were reported with the P A L PB-50 word lists (Eldert & Davis, 1951, as cited in Penrod, 1994). Problems included low reliability, unfamiliar vocabulary, and between-list differences in difficulty. The CID W-22 word lists were created to rectify some of these problems. The CID W-22 word lists consist of 4 lists of 50 words each; 120 of the words were selected from the original P A L PB-50 lists and 80 words were new. Campbell (1965), using listeners with sensory-neural or mixed hearing loss, argued that the CID W-22 lists were not all equivalent with respect to word difficulty. However, Elpern (1960) and Ross and Huntington (1962) concluded that the differences among W22 lists were minimal and insignificant clinically. Martin and Forbis (1978) reported that the W-22 word lists are widely used in clinical practice.  1.5.1.2 Development of the NU-6 Word lists Lehiste and Peterson (1959) noted that phonetics concerns the physiological and acoustical properties of speech and that there are extensive coarticulatory effects in speech sequences, e.g., a vowel will have different phonetic properties depending on the plosive which precedes it. Because speech articulations are rarely physically identical and  27  are not particularly similar unless in identical linguistic contexts, the authors note that it is not possible to have lists of words that are phonetically balanced. They define intelligibility as a property of speech communication requiring meaning and note that only units with meaning have normal linguistic distributional properties. They acknowledge that phonemes don't have intelligibility, only larger linguistic units (e.g., morphemes and words) do. They claimed that it is possible to have lists of words, which are in certain respects phonemically balanced by having each initial consonant, each vowel and each final consonant appear with the same frequency of occurrence within each list. They began to develop phonemically balanced word lists using consonantvowel-consonant (CVC) words. These words were drawn from lists (Thorndike & Lorge, 1944, as cited in Penrod, 1994) which included all C V C words appearing at least once per million words. This provided a pool of 1263 monosyllabic words from which list words were selected. Tillman, Carhart and Wilber (1963) used 95 words from Lehiste and Peterson's original lists and added 5 additional C V C words in developing two new 50-item lists. These were recorded and designated as Northwestern University Auditory Test number 4 (NU-4). Tillman and Carhart (1966) expanded this to four 50-item lists with 185 of the words coming from the 1263 word parent list and 15 from other sources. These lists are known as Northwestern University Auditory Test number 6 (NU-6) and are used broadly in research and clinical applications (Penrod, 1994).  1.5.2 Factors affecting W R T scores There are numerous factors affecting WRT scores including physical factors (e.g., presentation level, frequency composition, S/N ratio), linguistic factors (e.g., listener's  2 8  dialect and native language, familiarity of words to the listener), and test administration variables (e.g., response mode, scoring, stimulus materials, live voice vs. recorded presentation). The effects of frequency composition, use of a carrier phrase, half- versus full-list presentation, and response mode will be briefly discussed in this section. The effects of presentation level will be explored in Section 1.5.3. Using high- and low-pass filters, French and Steinberg (1947, as cited in Penrod, 1994) demonstrated the importance of high frequencies for correct identification of C V C syllables. When frequencies above 1.0 kHz were passed, 90% of the syllables were recognized correctly and when only the frequencies below 1.0 kHz were passed, correct identification dropped to 27%. Hirsh, Reynolds, and Joseph (1954, as cited in Penrod, 1994) found similar results using filtered CID W-22 stimuli. Although the high frequency consonants contain the least acoustic energy, they provide the major contribution to intelligibility (Fletcher, 1953). Fletcher and Steinberg (1930, as cited in Penrod, 1994) reported that identification of C V C syllables was higher when using an introductory sentence. However, Martin, Hawkins, and Bailey (1962) found no difference in performance when the carrier was omitted. Other authors have reported the opposite (e.g., Gelfand, 1975). Lynn and Brotman (1981) indicate that the carrier phrase "You will say..." contains perceptual cues that may assist identification of some initial sounds of test words. Martin (1997) recommends that, when using monosyllables and a conventional presentation, a carrier phrase should be used to prepare the patient for the stimulus word. It has become common practice for many audiologists to use only half of a 50item speech discrimination list (Penrod, 1994). Although investigations of NU-6 half-list  29 testing have been carried out (e.g., Schumaier & Rintelmann, 1974; Beattie, Svihovec, & Edgerton, 1978), no consensus exists regarding the clinical use of half-list testing (Penrod, 1994). There are two concerns when using half-lists: validity and reliability (Penrod, 1994). If scores are high, there is reasonable expectation that the results are likely valid. Thornton and Raffin (1978) note that reliability increases as the number of test items on a list increases. Furthermore, reliability is greatest for scores that are very high or very low. Accordingly, it would be advisable to use full-lists, especially when scores are neither very high nor very low. With either talkback or written responses, auditor error may affect W R T scores (Merrell & Atkinson, 1965). Nelson and Chaiklin (1970) found that examiner experience, monitoring level, distortion with the talkback system, and patient articulation were all contributing factors to auditor error. Northern and Hattler (1974) have advocated the use of written responses as a means of eliminating auditor error.  1.5.3  Performance-intensity curves for PB word lists The degree of understanding of speech materials is related to the signal intensity  and varies with respect to the type of speech signal (i.e., monosyllabic, polysyllabic, words, sentences, etc.) (Penrod, 1994). Typically, as intensity increases, speech understanding also increases. This relationship is often shown on a graph, called the performance-intensity (PI) function or the articulation-gain function. The steepness of the PI function will vary depending on the particular stimulus material used. The PI function for PB word lists (PI-PB function) shows that, for young normal-hearing adults, the  30  maximum PB score (PBmax) is obtained about 35 to 40 dB above the speech reception threshold (SRT) (Martin, 1997). To create a complete PI function, it is necessary to present words up to a level that is sufficiently loud, but not uncomfortable for the participant. The uncomfortable loudness level (UCL) is the sound pressure level (dBHL) at which speech becomes uncomfortably loud (Martin, 1997). The U C L is also called the threshold of discomfort (TD), the tolerance level, and the loudness discomfort level (LDL) (Martin, 1997). Since the U C L is a subjective measure, it must be determined for each individual participant. There are many factors that must be considered when determining the U C L . These include instructions to the patient, type of signal used and what is accepted as a response e.g., facial expression versus verbal response (Stach, 1998). Martin (1997) notes that, for normal-hearing listeners, the U C L often extends to the upper limit of the speech audiometer (100-110 dB HL).  Section 1.6  PI-PB Rollover  It has been shown (e.g., Schuknecht & Woellner, 1955, as cited in Jerger & Jerger, 1971) that PBmax is often disproportionately poor in patients with retrocochlear (i.e. 8th nerve) disorder. Jerger and Jerger (1971) first demonstrated that PI-PB rollover was of clinical value for differential diagnosis of cochlear versus retrocochlear disorders. They showed that, in patients with retrocochlear disorder, as the speech intensity increased, the PB score reached a maximum and then "rolled over" and became substantially poorer than PBmax as the intensity was further increased. In normals, the PB score stays at PB max as intensity is increased (e.g., Stach, 1998).  31  Rollover is defined as: Rollover = (PBmax - PBmin) where: PBmax is the highest score in the PI function and PBmin is the lowest score at an intensity greater than PBmax. Dirks, Kamm, and Bower (1977) and Jerger and Jerger (1971) showed that, although providing some differentiation between cochlear and retrocochlear groups, there is a substantial amount of overlap in rollover scores between the two groups. This is partly because rollover is necessarily restricted if the PBmax score is low. Therefore, in order to minimize the possible biasing effect of the PBmax score, they divided rollover by the PBmax score to obtain the rollover ratio (Rx): Rollover Ratio (Rx) = (PBmax-PBmin)/PBmax Overlap between cochlear and retrocochlear groups is "almost completely eliminated" if a rollover ratio of 0.45 is used as the criterion for differentiating cochlear (ratio <0.45) from retrocochlear (ratio >0.45) pathology (Dirks et al., 1977; Jerger & Jerger, 1971). Relatively low rollover ratios are exhibited by those with normal hearing or conductive pathology. The calculation of rollover ratio does not take into account the findings of Egan (1948) and Thornton and Raffin (1978) which show that WRT scores are more variable for mid-range scores than extreme scores. For example, when a listener scores 92% on a 50 word test, the 95% confidence interval is from 81% to 98% but for a score of 68%, the confidence interval is larger, from 54% to 80%. Thus, the Rx will be higher for, say, a 20% rollover from 70% to 50% (Rx=0.2/0.7=0.29) than for a rollover from 100% to 80% (Rx=0.2/1.00=0.2), but the higher Rx will not reflect the greater range of uncertainty  32  associated with it. Therefore, Rx alone is less than ideal for evaluating differences in rollover. When comparing rollover scores, a measure of uncertainty associated with the scores should be taken into account.  1.6.1  Physiologic Basis for PI-PB rollover Review of the literature reveals three physiologic explanations for PI-PB rollover.  Rollover has been explained by mechanical effects (caused by absent acoustic reflexes), neural damage (caused by eighth cranial nerve tumors), and combined neuro-mechanical reasons. The following sections explore these hypotheses.  1.6.1.1 Mechanical Basis An absent acoustic reflex may be one of the factors contributing to rollover. Wormald, Rogers, & Gatehouse (1995) note that contraction of the stapedius muscle attenuates the lower frequencies much more than the higher frequencies. As stated earlier, attenuation is thought to be between 10 and 15 dB for the frequencies below 0.1 kHz and 0-6 dB for the frequencies above 0.1 kHz. Thus, in cases where the stapedius muscle is paralysed, there would be decreased attenuation of the lower frequencies. At higher sound levels, this would result in an upwards spread of masking (i.e., masking of the higher frequencies) as explained by the shallow-dipping, low-frequency tails of primary auditory neurons (see Section 1.4.4.3). Wormald et al., (1995), argue that because more of the key information necessary for understanding speech is in the higher frequencies (above 1 kHz) than in the lower ones (e.g., French & Steinberg, 1947), this upwards spread of masking may lead to decreased speech discrimination.  33  McCandless and Goering (1974) studied twenty patients who had undergone unilateral stapedectomy and three patients suffering from unilateral Bell's palsy. One of the tests included speech discrimination scores at comfort levels and at levels ascending up to 100 dB H L . A l l listeners tested scored better than 90% at comfortable listening levels bilaterally. At high presentation levels, in no cases did discrimination drop by more than 20% in the normal ear. However, in 54% of cases, scores dropped below 60% in the operated ear. In 39% of the cases, rollover was greater than 50%. The authors note that in every case where the discrimination score dropped 20% or more, the abrupt change was at 80 to 90 dB H L . This intensity is within the normal range where the acoustic reflex begins to act (e.g., Stach, 1998). Of the three Bell's palsy patients studied, two showed greater than 50% rollover. The authors identified the absent stapedius reflex, and resulting lack of dampening effect at high intensity levels, as one possible reason for the rollover. However, they could not explain why rollover did not occur with greater consistency among all the stapedectomized patients. McCandless and Schumacher (1979) studied 58 patients with idiopathic facial paralysis. Patients were assigned to one of three groups. Group 1 (N=26) patients had facial paralysis, normal hearing and absent acoustic reflexes ipsilateral to the paralysis. Group 2 (N=20) patients had facial paralysis, mild-to-moderate high-frequency sensorineural hearing loss, and absent ipsilateral reflexes. Group 3 (N=12) had facial paralysis, normal hearing, and intact acoustic reflexes bilaterally. For Group 1, mean rollover at 100 dB H L was 50% in the affected ear and 8% in the unaffected ear. For Group 2, mean rollover was 41% in the affected ear and 12% in the unaffected ear at 100 dB H L . For Group 3, the majority (75%) showed rollover less than 20% in the affected  34  ear and, for the whole group, there was no statistically significant difference in rollover found between the ears. The fact that rollover occurred without exception in patients with paralyzed stapedius muscles (and not in those with intact acoustic reflexes) strongly suggested to the authors that the source of distortion was mechanic rather than neural. Furthermore, site of lesion tests failed to demonstrate any evidence of eighth nerve dysfunction. Hannley and Jerger (1985) used P A L PB-50 word lists to determine whether patterns of phoneme identification error differed among listeners with cochlear and retrocochlear disorder. They found that vowel errors were more prevalent in the retrocochlear group and varied directly with increasing stimulus presentation level, i.e., there was rollover with respect to vowel identification. However, consonant errors were not different between the two groups and there was no level-dependent effect for consonants. These authors found that the presence of rollover was directly related to the status of the acoustic reflex. Of the 15 people in their retrocochlear group, 10 had absent reflexes (at equipment output limits) whereas 5 had measurable reflexes at one or more frequencies from 0.25 to 4 kHz. Each of the 10 participants without reflexes showed rollover greater than 20%, while the 5 with intact or partially intact reflexes maintained PBmax at higher presentation levels. Taken together, these studies suggest that rollover can be associated with the absence, or partial absence, of acoustic reflexes. As noted by Wormald et al., (1995), the suspected mechanism whereby absent acoustic reflexes lead to rollover is the upwards spread of masking. As stated earlier, the attenuation by the stapedius muscle is thought to be 15 dB below 0.1 kHz and minimal at higher frequencies. The key information  35  necessary for speech perception is thought to be contained in the frequencies above 1 kHz. Nevertheless, it is questionable as to whether an increase of 15 dB below 0.1 kHz can provide enough masking energy to the frequencies above 1 kHz such that intelligibility is reduced at high levels in the ear with absent acoustic reflexes.  1.6.1.2 Neural Basis The theory of the A L S R (see Section 1.4.4.4.3) stresses the importance of neural synchrony in encoding the formants of high-intensity speech. At low intensities, the auditory nerve can encode speech without relying on neural synchrony. Thus, neural asynchrony could be a cause of PI-PB rollover. There exists a fairly substantial amount of clinical evidence to support this idea. The following sections explore pertinent evidence from three distinct populations: those with an acoustic neuroma, multiple sclerosis (MS), and the elderly.  1.6.1.2.1 Evidence from Patients with an Acoustic Neuroma Acoustic neuromas, or vestibular schwannomas, are benign, encapsulated tumors arising from the Schwann cells of the superior branch of the vestibular nerve in 2/3 of acoustic neuroma patients (Jacobson, Jacobson, Ramadan & Hyde, 1994). Acoustic neuromas can also originate from the acoustic portion of C N VIII or the inferior branch of the vestibular nerve. The tumors tend to encroach on and displace neurovascular structures. Small tumors originating in the internal auditory canal (IAC) are usually asymptomatic until they increase in size and apply pressure against the originating nerve fibre, adjacent nerves and vascular supply.  36  A neural basis for rollover is implied by several authors who have reported that a high rollover ratio is associated with the presence of an acoustic neuroma (e.g., Bess, Josey, & Humes, 1979; Dirks et al., 1977; Jerger & Jerger, 1971; Meyer & Mishler, 1985). Furthermore, rollover has been found in neuroma patients whose reflexes are still intact (e.g., Hannley & Jerger, 1981; see Section 1.6.1.3). One of the major applications of auditory evoked potentials (AEPs) is the detection and localization of disorders affecting the auditory nerve and other, more rostral, stations in the auditory pathway (Jacobson, 1994). The auditory brainstem response (ABR) is a sensitive indicator of the integrity of the VHIth nerve and auditory brain stem function; taking audiometric thresholds into account, if the A B R is abnormal, there is a very high likelihood of retrocochlear disorder (Stach, 1998). Jacobson et al., (1994), note that mechanisms of damage to the auditory nerve identified by the A B R trace include compression, ischemia, hemmorrhage, and demyelination/degeneration. Any or all of these conditions can coexist. What all of these mechanisms have in common is the disruption of the myelin sheath that coats the auditory nerve. Myelin is a protein that is needed for fast conduction of nerve impulses. When the myelin sheath becomes damaged, leakage of current could ensue, resulting in conduction failure or a slowing of conduction and a corresponding temporal dispersion of electrical activity (Jacobson et al., 1994). The authors note that if the neural response is not synchronized with the presentation of a suitable auditory stimulus (e.g., if the electrical response to the stimulus is temporally dispersed), A E P waveform components can be depressed, delayed or abolished. The authors go on further to say that the A E P test is a test of neural synchronization, e.g., the A B R is not a test of whether afferent information reaches the  37  auditory nerve but rather it is a test of whether the information passes through in a synchronized manner. However, Jacobson (1994) cautions that neural synchronization is a necessary but not sufficient condition for recording activity at a remote electrode. There exists clinical evidence that associates rollover with acoustic neuromas. Furthermore, neural asynchrony is identified as the mechanism by which acoustic neuromas lead to abnormal ABRs. Therefore, it is likely that rollover is associated with neural asynchrony.  1.6.1.2.2 Evidence from MS patients Jacobson (1994) notes that any disorder that disrupts neural synchrony can affect A E P waveform components. Multiple Sclerosis (MS) is one such example. M S is a primary, degenerative disease characterized by demyelinating lesions scattered throughout the nervous system (Rappaport, Gulliver, Phillips, Van Dorpe, Maxner, & Bhan, 1994). Because the number and loci of the lesions vary considerably between patients, the sequelae vary considerably as well (Stach, Delgado-Vilches, & SmithFarach, 1990). Rappaport et al., (1994) note that, in terms of hearing, the demyelinating nature of MS essentially results in a disruption of neural timing. These authors studied several MS patients using a speech in noise paradigm and found them to be selectively impaired, confirming a temporal processing deficit (Rappaport et al., 1994). The incidence of A B R abnormalities in M S range from 19 to 93% (e.g., Stockard & Rossiter, 1977). The unpredictability of the demyelinating lesions of M S is thought to be associated with the unpredictability of the normality of the A B R . PI-PB rollover has also been inconsistently found in MS patients (Stach et al., 1990, Rappaport et al., 1994). It  38  can be inferred, from the discussion in Section 1.6.1.2.1 and the findings by Rappaport et al. (1994), that the mechanism by which M S produces an abnormal A B R is neural asynchrony. Thus, the inconsistent finding of rollover in the MS population provides additional clinical evidence, albeit weak, suggesting an association between rollover and neural asynchrony.  1.6.1.2.3 Evidence from the Elderly Population Several authors (e.g., Dirks et al., 1977; Gang, 1976; Jerger & Jerger, 1971; Shirinian & Arnst, 1980) have found that pronounced rollover, though rare in the general clinical population, is more common among aged listeners. For example, of the 741 ears (normal hearing and conductive, mixed, and sensorineural losses) studied by Jerger and Jerger (1971), 85% had a rollover ratio less than 0.20, 13% had a ratio between 0.20 and 0.39 and rollover exceeding 0.39 was observed in only 2% of the ears evaluated. Only nine listeners had very high rollover ratios (between 0.50 and 0.59). Eight of the nine were more than 50 years old, seven were more than 60 years old and six were more than 70 years old. They concluded that pronounced rollover was rarely seen in the general clinical population, but when observed was more common among aged listeners. Dirks et al. (1977) reported that in an evaluation of 102 patients, 3 elderly patients had rollover indices greater than 0.45 despite radiological studies contraindicating the presence of a space-occupying lesion. In explaining the results, the authors suggested that the same elderly listeners who exhibited rollover may also have neural presbycusis. Only Shirinian and Arnst (1980) controlled for the presence or absence of the acoustic reflex in their rollover study. The authors found the PI-PB functions for 66 aged  39  listeners (108 ears). A l l listeners had acoustic reflexes present at 0.5 kHz and 1.0 kHz. They found that 79% of the ears had rollover ratios below 0.2, 17% of the ears showed rollover between 0.20 and 0.39, and approximately 4% of the ears had rollover greater than 0.40. These results showed more rollover than the findings reported by Jerger and Jerger (1971) for a general clinical population and the authors concluded that the frequency and magnitude of the rollover phenomenon was greater in the aged population. Radiological evaluations contraindicated the existence of space occupying lesions in all 4 of the elderly patients who showed rollover greater than 0.40. By ruling out any contributions to rollover by absent acoustic reflexes or acoustic neuromas, the authors interpreted the results of the study as being representative of rollover's sensitivity to the retrocochlear effects of age on the auditory system. There exists substantial evidence to suggest that neural asynchrony is more common in the older adult population, even in those with audiometric thresholds in the normal range (see Section 1.3.3). Furthermore, the work by Shirinian and Arnst (1980) suggests that there is a greater degree of rollover in the older adult population than in the general clinical population. Thus, this evidence points to an association between rollover and neural asynchrony in the aged population.  1.6.1.3 Neuro-mechanical Basis Hannley & Jerger (1981) studied the relationship between the acoustic reflex and high-level speech intelligibility in 52 patients with confirmed acoustic tumors. A l l reflexes (contralateral, 0.5 to 4 kHz) were intact for one group of participants, in another they were absent, and in the third group, participants had reflexes present at some  40 frequencies. The authors found that the patients with absent reflexes had significantly greater rollover than did those whose reflexes could be elicited at one or more frequencies. The average rollover was less than 20% for the group with intact reflexes and greater than 50% for the group with entirely absent reflexes. The authors suggested a neural-mechanical interaction as a basis for rollover in patients with retrocochlear disorder and then proposed that whatever speech abnormalities are attributable to a neural basis are exacerbated by failure of the normal function of the acoustic reflex. However, Hannley and Jerger (1981) did not document the size of the acoustic tumors of the patients in their study. Thus, it is possible, even likely, that those patients with absent acoustic reflexes also had larger/more advanced tumors than the patients with present or only partially absent reflexes. Therefore, the possibility remains that the high degree of rollover in the retrocochlear patients with absent reflexes may not be indicative of mechanical effects, but rather neural effects due to increased tumor size and a resulting increased demyelination of the auditory nerve.  1.7  Summary Older people, even those with normal audiograms, have an especially hard time  understanding speech in noisy conditions. Such clinical observations can not be fully explained by cognitive slowing factors and/or losses in pure tone sensitivity with age. The central auditory processing argument states that the elderly have problems due to deficits in auditory processing ability, other than pure-tone hearing loss or general cognitive slowing. Specifically, it is believed that many of the problems in the elderly are due to temporal processing deficits. Temporal processing, or auditory temporal  41  resolution, refers to the auditory system's ability to detect changes in stimuli over time. A n important temporal processing ability of the auditory system is neural synchrony. Neural synchrony refers to the ability of auditory nerve fibres to discharge (phase lock) to one phase of an incoming frequency component of a stimulus. It is generally agreed, based on neurophysiologic evidence, that neural synchrony is important in encoding low frequencies (i.e., less than 5 kHz). There is less evidence to support the role of synchrony in encoding intensity patterns such as formants. In general, neural firing rate encodes intensity in the auditory system. However, recordings from the auditory nerve fibres of cats, in response to speech stimuli, show that firing rate (for high spontaneous rate neurons) saturates at high intensities. The theory behind the A L S R proposes that neural synchrony, in combination with neural firing rate, is able to sufficiently encode formant information at high intensities. However, currently, there is no physiologic evidence to support the A L S R . Temporal processing ability (measured using behavioural tasks such as gap detection, frequency discrimination and binaural unmasking) is poorer for older versus younger adults, even when audiometric thresholds are taken into account. Some of the results in the older population are consistent with, and/or indicative of, a decrease in neural synchrony in the aging auditory system. The temporal characteristics of speech can be split into two levels: envelope and fine structure characteristics. The conditions identified as problematic for the elderly (e.g., background noise) act to smooth the temporal fluctuations in the waveform envelope and several studies have shown the importance of the envelope to speech intelligibility. It is therefore reasonable to expect that the problems of the elderly in  42 background noise may be attributable, to some extent, to the effect these conditions have on the waveform's envelope. However, background noise can also contain energy that overlaps with frequency regions important for speech. If speech is encoded by rate-place mechanisms only, noise can damage the auditory representation of the speech spectrum. Theories (i.e., the ALSR) that take temporal fine structure information (i.e. signal periodicity) into account are more robust in the presence of noise than rate-place theories by themselves. Thus, the A L S R provides a theoretical account for the recent findings that suggest that the role of fine structure may also be important in perceiving speech in noise. In these experiments, the performance by young normal-hearing participants using temporally jittered SPIN-R sentences resembled that of the elderly (with audiograms in the normal range) using intact stimuli (Pass, 1998; Brown, 2000; Pichora-Fuller et al., submitted). Because this elderly population is thought to suffer from neural asynchrony, it was concluded that temporally jittered stimuli could be simulating neural asynchrony in young normals. The present thesis takes a new angle to look at the question of whether temporally jittered stimuli simulate neural asynchrony in young normals. If young normal-hearing participants, with intact acoustic reflexes ipsilateral to the test ear, show significant PI-PB rollover when presented with temporally jittered stimuli, then jitter also simulates neural asynchrony like that found in retrocochlear pathology. This is a valid claim because, of the possible physiologic explanations available to explain PI-PB rollover, only the neural explanation remains if acoustic reflexes are accounted for. The backing for the claim has support in both basic auditory neuroscience and observed clinical phenomena. The A L S R suggests that neural synchrony is necessary to encode speech at high intensity levels.  43  Clinically, pronounced rollover is often found in patients with acoustic neuromas, even those with intact acoustic reflexes. Acoustic neuroma patients also often have abnormal ABRs, a finding that has been linked to decreases in neural synchrony. Thus, there is evidence that associates PI-PB rollover with neural asynchrony. Direct clinical evidence linking rollover with asynchrony would be the documentation of patients with PI-PB rollover, intact acoustic reflexes, and an abnormal A B R due solely to the presence of a surgically confirmed acoustic neuroma. Although this direct evidence was not available in the literature, it is quite possible that these cases do exist, yet have not been documented as such. Further evidence associating PI-PB rollover with neural asynchrony exists in the pattern of findings for both the MS and elderly populations. The claim of the present thesis would be refuted if the young participants of this study also showed PI-PB rollover when presented with unmodified stimuli or if jitter was of no consequence to PIPB rollover. The main research question of the present thesis is: when presented with temporally jittered stimuli, will young normal-hearing participants, with intact acoustic reflexes, show significantly greater rollover than when presented with intact stimuli? A positive finding would support the claim that temporal jitter simulates neural asynchrony based on: 1) the clinical association of rollover and neural asynchrony and, 2) the A L S R ' s theoretical prediction of rollover in cases of neural asynchrony.  44  If the finding is negative, the claim is not necessarily refuted. Other theories/mechanisms have been proposed to explain the coding of intense speech, e.g., intense speech is sufficiently coded by the firing rate of low spontaneous rate neurons (e.g., Geisler, 1998).  1.8  Hypotheses This study will test the following null and research hypotheses:  Null Hypothesis 1 PI-PB rollover will not be greater than zero in the intact (unjittered) condition for NU-6 lists presented in quiet. This hypothesis is based on the finding that, in normalhearing individuals, WRT score stays at PBmax as intensity is increased (e.g., Stach, 1998).  Accompanying Research Hypothesis PI-PB rollover will be greater than zero in the intact (unjittered) condition for NU-6 lists presented in quiet.  Null Hypothesis 2 Rollover in the NU6jitter condition will not be greater than rollover in the NU6intact condition for NU-6 lists presented in quiet.  45 Accompanying Research Hypothesis There will be greater PI-PB rollover in the jitter condition than in the intact (unjittered) condition for NU-6 lists presented in quiet. This is the main research hypothesis of this thesis and is based on the fact that if acoustic reflexes are normal, only a neural basis (e.g., neural asynchrony) is left to explain the existence of PI-PB rollover. Acceptance of this hypothesis would support the claim that temporal jitter simulates neural asynchrony in young normal-hearing listeners.  Null Hypothesis 3 There will be no difference in PBmax between the intact and jittered condition for NU-6 lists presented in quiet.  Accompanying Research Hypothesis There will be a difference in PBmax between the intact and jittered NU-6 condition. This hypothesis is based on the expectation that jitter will disrupt word intelligibility because of reductions in periodicity in the stimulus.  Null Hypothesis 4 There will be no difference in PI-PB rollover between the within-band NU6jitter and between-band W22jitter conditions.  46 Accompanying Research Hypothesis There will be a difference in PI-PB rollover between the NU6jitter and W22jitter conditions. This hypothesis is based on the fact that the word lists between the two conditions are different and the fact that the W22jitter condition contains both withinband and between-band jitter (see Sections 2.2.2.1, 2.4, and 3.1.2).  Null Hypothesis 5 There will not be a correlation between PBmax and sex, age, years of education, left-ear or right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds, or handedness.  Accompanying Research Hypothesis There will be a correlation between PBmax and sex, age, years of education, leftear or right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds, or handedness.  Null Hypothesis 6 There will not be a correlation between PI-PB rollover and sex, age, years of education, left-ear or right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds or handedness.  47  Accompanying Research Hypothesis There will be a correlation between PI-PB rollover and sex, age, years of education, left-ear or right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds or handedness.  48  2.0 PILOT S T U D Y  2.1 Purpose A pilot study was conducted in order to guide decisions regarding the main experiment. Specifically, the pilot study was used to determine appropriate jitter parameters, refine procedural details, and gauge participant reactions regarding comfort and difficulty of the experimental task and conditions. Although limited in the number of participants, the pilot study provided much valuable information and led to changes in the protocol for the main experiment.  2.2 Method The pilot study procedure involved standard clinical testing of auditory function (pure-tone air-conduction audiometry and immittance measures) and obtaining PI-PB functions for both unjittered NU-6 word discrimination lists and for within-band temporally jittered word lists. Two degrees of jitter were employed to help decide which would be the most effective jitter parameters to be used in the main experiment.  2.2.1 Participants There were four participants in the pilot study, two males (MV, ES) and two females (TA, ZK). Ages ranged between 25 and 34 years. M V was tested in the right ear and TA, ES and Z K were tested in the left. Both M V and T A spoke English as their first language. ES (Danish) and Z K (Turkish) both spoke English as their second language.  A l l four pilot participants had pure-tone thresholds and ipsilateral acoustic reflexes within normal limits except for T A who had a mild 6 kHz notch bilaterally and elevated/absent ipsilateral reflex thresholds. Measurements of UCLs ranged between 80 and 90 dBHL. For the purposes of this experiment, the criterion for normal hearing was pure-tone thresholds < 20 dBHL at each of the following test frequencies: 0.25, 0.5, 1, 2, 4 and 8 kHz (e.g., Martin, 1997; Stach, 1998). Normal acoustic reflex thresholds were taken to be between 70 and 100 dB H L (e.g., Stach, 1998).  2.2.2 Materials The four, 50-item, NU-6 word lists were used to obtain the PI-PB functions in this pilot. See Appendix A for the items of each list. NU-6 word lists (form A) were purchased on compact disc from Auditec of St. Louis. The recorded voice is male, with a mid-western American (Oklahoma) dialect. Auditec wrote their four previously recorded NU-6 words lists to a C D from cassette at a sampling rate of 44.1 kHz in stereo (two separate channels). The two channels are identical. Only one of the two channels was saved. Each NU-6 word list was saved as a *.SND file (16 bit, mono, signed) using the C D Extraction feature on the demonstration version of the Goldwave (Craig, 2000) digital audio editor (www.goldwave.com). Files were saved on a Dell Dimension XPS T700r PC at a sampling rate of 20 kHz to save disk space and to reduce the computation time required to jitter the stimuli. A sampling rate of 20 kHz preserves signal components up to 10 kHz, such that no speech information was lost when the sampling rate was reduced.  50  Each of the four *.SND files had 50 phrases; each phrase included a carrier phrase (Say the word...) and the target word. In preparation for jittering, each phrase had to be excised and saved as a separate file using the Goldwave software. It was reasoned that the target word should be jittered and not the carrier phrase because it was thought that a jittered carrier phrase might confuse the participant. In the original NU-6 recordings, there is almost no silent interval between the carrier phrase "Say the word...." and the monosyllabic target stimulus. The last word of the carrier phrase is often co-articulated with the target. This presented a problem in that, for the jitter program used (see section 2.2.2.1), it is necessary to specify a time interval in which to apply the jitter. Thus, a silent interval of approximately 1 second was inserted at the most appropriate place between the carrier phrase and the target word. This was done to ensure that the target word was always in the interval between 2.0 and 3.5 seconds after the start of each file. Jitter was applied to this interval only. However, when inserting the silent interval between the carrier and the target in preparation for jittering, a clean cut was not always possible. For example, the unjittered target word "laud" sounded like "blaud" and "hole" sounded like "ole". Prior to testing the pilot participants, unjittered words were presented, in experimental conditions (i.e., at appropriate levels, with masking, etc.) to volunteers. An immediate observation by the volunteers was that some of the words sounded strange; they sounded abruptly cut off at the end of the carrier and beginning of the target word. It was determined that the problem was due to the co-articulation issue described above. A decision was made to go over the words again, paying extremely close attention to the locations of the cuts, and to modify the cuts (and improve perception of the target word) where possible. After this was done, the procedure for the preparation of the stimuli was  51  repeated. Following these modifications, the same volunteers noted a definite improvement in most of the unjittered words but a few words still sounded degraded. Nonetheless, these newly modified stimuli were jittered and used for the pilot study. After all the individual files were jittered, the phrases were concatenated into files of 25 phrases each. Periods of silence were inserted between each phrase to ensure a silent interval of approximately 4 seconds between each complete phrase. This was done to allow the listener time to repeat and write down the target word. Each of the concatenated files was then resampled at a rate of 44.1 kHz and saved as a * . W A V file in preparation for writing back to CD. Files were written to C D using the Adaptec Easy C D creator, Version 4.02 (www.adaptec.com).  2.2.2.1 Jitter Methods When a sound is digitized, the sampling rate determines the number of data points (amplitude values), per time interval, that describe the sound. For example, if a sound is sampled at 20 kHz there are 20,000 amplitude values per second, each separated from the next by a 0.05 ms. interval. When a digital signal is played, the data (amplitude over time) between data points is interpolated and smoothed. Thus, the temporal envelope of the digital signal is based on the amplitude and time coordinates of the soundfile. When a digitized signal is temporally jittered, amplitude values are shifted from their original temporal positions. Bruce Schneider (1997) wrote the original Jitter computer program, which is a program that temporally jitters a given digital signal. Carol Jaeger (2000) made modifications to the original program to produce Jitterl4.exe (see Appendix B) and this version of the Jitter program was used in the present study.  52  The Jitterl4.exe program's method of temporal jittering is modelled on the Gaussian distribution of amplitude in low-pass (LP), band-limited white noise. L P bandlimited white noise has a distribution of amplitudes that range between a maximum (positive) value and a corresponding minimum (negative) value (e.g., +20 dBSPL and -20 dBSPL), with a mean and a standard deviation. The frequency bandwidth of the noise impacts the rate at which amplitude changes occur. The higher the upper cut-off frequency of the L P band-limited noise, the more rapid the changes in amplitude. For each data point in the digitized sound file, the Jitterl4.exe program selects a delay value by referring to the distribution of a noise, determining the amplitude value at the corresponding point in time and then converting this amplitude into an associated delay value. The delay value determines the position (in time) in the original, digitized file whose amplitude value is to be substituted in for the data point that is being jittered. It is thought that in naturally-occurring, internal jitter, such as that hypothesized to occur in the elderly, the size of each delay value and the number of times this value changes varies randomly over time (Durlach, 1972; Pichora-Fuller & Schneider, 1992). In the Jitter 14.exe program there are essentially two parameters, input by the user, which control the degree of jitter applied to the input signal: standard deviation (SD) and bandwidth (BW). The SD is entered into the program as a number of sample points, e.g., 10. When multiplied by the time interval between two data points (which depends on the sampling rate), the SD can be expressed in msec. The larger the SD, the greater the size of the delay values that can be selected from the distribution. For example, if the sampling rate is 20 kHz and the user enters SD=10, the SD of the delay value is ±0.50 msec. (10x0.05msec. = 0.50 msec). Note that, when the program selects an amplitude value  53  from the L P white noise, the delay value will be positive or negative depending on whether the amplitude value is positive or negative. If the selected amplitude value happens to be zero, then the delay value is zero and the original amplitude value is kept for the data point that is being jittered. The SD parameter value determines the range within which delay values can vary. The other parameter governing the degree of temporal jitter is B W . B W represents the upper cut-off frequency of the LP band-limited noise. The higher the B W value, the more rapid the changes in amplitude of the noise and, since delay values are dependent on the amplitude of the noise, the more rapid the changes in delay values. For example, a B W of 0.5 kHz produces more rapid changes in delay values, and therefore a greater severity of temporal jitter, than a B W of 0.1 kHz (e.g., Pass, 1998). In the older versions of the Jitter program (such as the version used by Pass, 1998), jitter parameters were applied to a single frequency band, with the upper cutoff limited by the Nyquist frequency. For example, for a sampling rate of 20 kHz, the Nyquist frequency would be 10 kHz and jitter parameters were applied to the speech signal frequencies from 0-10 kHz. The upper cutoff of the single frequency band could be input by the user and did not need to be as high as the Nyquist frequency. This type of jitter is referred to as within-band jitter. Within-band jitter was the only type of jitter used in the pilot study. Two within-band jittered conditions (jl and j2) and one unjittered condition (jO) were used in the pilot study. The jittered conditions used in this pilot were the same as the two most severe conditions of jitter used in the study by Pass (1998), i.e., the high-SD (jl) and the high-SD/BW (j2) conditions. The parameters for j l were SD=5 sample  54 points=0.25 msec, BW=0.1 kHz and for j2 they were SD=5 sample points=0.25 msec, BW=0.5 kHz. In both jittered conditions, jitter was applied to the speech signal withinband from 0 to 10 kHz.  2.2.2.2 Calibrating the Sound Level of the Stimuli On the first track of the compact disk obtained from Auditec, there is a 1.0 kHz calibration tone. The intensity of the speech signal (target word) is defined as being equivalent to the sound pressure level (SPL) in dB of a 1.0 kHz calibration tone recorded at the average level of the carrier phrase (Bess et al., 1979). The 1.0 kHz calibration tone was treated in the same fashion as the (unjittered) stimuli: it was extracted from the original CD, saved as a *.SND file using the Goldwave software, sampled at 20 kHz, resampled at 44.1 kHz, saved as a * . W A V file and then written back to the beginning of each C D used in the pilot (and main experiment). Prior to presentation of the speech material, the 1.0 kHz calibration tone on each C D was adjusted to 0 on the V U meter of the audiometer.  2.2.3 Apparatus and Physical Setting Digitized C D recordings of both jittered and unjittered NU-6 word lists were fed from a J V C XL-Z232 compact disc player, into a Grason-Stadler GSI-16 audiometer and then into TDH-50P headphones (test ear only). Masking noise was delivered from the GSI-16 audiometer through the TSH-50P headphones to the non-test ear. A l l equipment was calibrated to ANSI 3.6 1969/ISO 389 1975 standards. The pilot experimental  55  sessions took place with the participant seated in a double walled, sound attenuating IAC booth.  2.2.4 Procedure One of the major goals of the pilot was to determine if rollover could be observed under either, or both, of the two jitter conditions (j 1 and j2) and to verify that rollover was not observed in the unjittered condition (jO). If the research hypothesis is true, rollover would increase as the degree of jitter increases from jO to j l to j2. However, the number of available, equivalent word lists was limited to four full lists of 50 words each (list 1, 2, 3 & 4). Therefore, to avoid possible learning effects by testing all the conditions with a different word list, a rather complicated scheme was devised, as described below. The pilot protocol was split into two sessions, each lasting approximately 1 hour. In the first session, the listener was required to fill in a hearing and language history form (see Appendix C). Next, otoscopic screening was performed in both ears. At this point, one ear was chosen for presentation of the word lists. In this selected ear, a tympanogram and ipsilateral acoustic reflexes (0.5, 1, and 2 kHz) were measured. Pure-tone airconduction thresholds (0.25, 0.5, 1, 2, 4 and 8 kHz) were then obtained in the selected ear. For this same ear, the speech reception threshold (SRT), most comfortable listening level (MCL) and uncomfortable listening level (UCL) were obtained. The SRT was obtained according to standard procedures (e.g., Stach, 1998) but the M C L and U C L were obtained informally, using conversational speech. Finally, the PI-PB function for the jO (unjittered) condition was obtained. To obtain this function, four presentation levels were used: 40, 55, 65, and (UCL-5) dB H L . In order to prevent crossover, speech  56  noise masking was delivered to the non-test ear at 5, 20, 30, and (UCL-40) dB H L , respectively. Ascending intensity levels were used. Participants were told to repeat and write down the word and to guess if they were unsure. Participants were informed that if they felt uncomfortable at any time, they should let the experimenter know and the procedure would be terminated. In order to maximize the number of available word lists, three combinations of the word order for each list were employed (list 1 A , I B , 1C, 2A, 2B, 2C, etc.). Each 50-word list was split into two half lists of 25 words each (list 1A1, 1A2, 1B1, 1B2, 1C1, 1C2, 2A1, 2A2, 2B1, etc. Therefore, the total number of half lists available was 24. The presentation order is detailed in Table 2.1. The rationale for this procedure is based on the work of Thornton and Raffin (1978). These authors note that test-retest variability decreases as the number of items in the word discrimination list increases. Variability also increases rapidly as scores move away from the extreme scores of 100% and 0%. There is maximum variability at scores of 50%. Half-lists were presented unless the obtained score was less than 88% (22/25). The conditional use of half-lists is a compromise between minimizing variability and not re-using words between conditions (and thereby minimizing practice effects). The second session was administered on a different day than the first session. In this second session, PI-PB functions were obtained for the j l condition. The four presentation and masking levels (in ascending order of intensity) used in the first session were also employed in the second session. The same instructions and conditions for full list presentation were used. Immediately afterwards, the PI-PB function for the j2  57  condition was obtained. The presentation order for the word lists of the second session is also detailed in Table 2.1.  40 dBHL  55 d B H L  65 d B H L  UCL-5dBHL  JO  1A1  2A1  3A1  4A1  If<88%  1B1  2B1  3B1  4B1  Jl  1A2  2A2  3A2  4A2  If <88%  1B2  2B2  3B2  4B2  J2  1C1  2C1  3C1  4C1  If <88%  1C2  2C2  3C2  4C2  Table 2.1 Pilot study word list presentation order Part of the rationale behind the pilot procedure included the assumption that, ideally, every listener would obtain a score greater than 88% in the jO condition in Session #1. Then, participants would receive fresh words in the first half list at each intensity in j l , Session #2. If the pilot procedure proved to be successful, it was forecast that in the main experiment, for half the participants, the order of j l and j2 could be switched so that fresh words would be presented for the j2 condition. If no significant differences were noted, the results could then be combined. Part of the problem with the pilot procedure was that phonemic balance is not maintained when half lists are used. Furthermore, word composition of half lists is not preserved across combinations e.g., 1A1 does not contain the same words as 1B1; some are the same, some are different. Additionally, pilot results (see Section 2.3) showed that participants often did not score greater than 88% in the jO condition in Session #1.  58  Therefore, this rather complicated scheme was not used for the main experiment (see Section 2.4).  2.3 Results Results were recorded as scores out of 25 (or 50) and converted to a percentage. Scoring was done on line based on the participant's verbal response. A decision regarding a pass/fail of the 88% criterion was made based on this score. Afterwards, results were checked with the written responses, with the written responses overriding heard verbal responses when discrepancies occurred. When written responses were illegible, heard responses were used. Figure 2.1 summarizes the results from the pilot study. It can be seen that a surprising amount of rollover (average Rx=0.12) was noted in the unjittered condition for each client. For young normal listeners, one would expect Rx to be approximately 0 for jO. The high amount of rollover in the jO condition could be due to one or more of a number of factors. First, the co-articulation issue may have caused discrepancies in word difficulty across what were presumed to be equivalent word lists. Another possibility is that the cutoff for full-list presentation may not have been strict enough, e.g., perhaps a 92% criterion for acceptance of half-list results would have ensured better reliability. In fact, given 88% as the criterion, a rollover ratio of 0.12 could have been observed without a full list presentation, e.g., a participant could get 100% at a low intensity and 88% at a higher intensity without getting a full-list presentation at the higher presentation level, with a resulting Rx = 0.12.  Figure 2.1 Rollover (%) for each Pilot Participant in each Jitter Condition  MV  TA  ES Pilot Subjects  ZK  AVERAGE  Another observation from Figure 2.1 is that almost no rollover was noted in the j l condition but that substantial rollover was noted in the j2 condition. From Figure 2.2 it can be seen that, compared to jO, PBmax was substantially reduced in the j2 condition but not in the j l condition. It was found that most participants required both half lists to be presented in j l and all participants required both half lists in j2. PBmax was high (>90%) for every participant in the jO condition and >90% for every participant in the j l condition. What might be deduced from Figure 2.2, but was certainly clear from participant reactions and comments, is that ESL participants (ES and Z K ) had more difficulty with unjittered words than native English speakers.  60  Figure 2.2 PBmax (%) for each Pilot Participant in each Jitter Condition  ES  MV  •  io  TA Pilot Subjects  ZK  AVERAGE  •  *  Following completion of the pilot, participants were informally asked for their comments. Participants noted they had more problems with the onsets of unjittered words. Problem words included laud (frequently heard as blood or log), hole {pie), rag (bag), rain (brain), youth (use) and raise (braise). Some participants responded with nonwords rather than attempting to guess at a real word. Participant M V spontaneously mentioned that the loud words (UCL-5dB HL) in the j2 condition were more difficult. However, results for M V showed only a minor degree of rollover in this condition. When participant Z K was asked whether the loud words in the j2 condition were more difficult, the participant said it was not noticeably so, yet the degree of rollover observed in the j2 condition for this listener was substantial. In session #2, participant T A noted that practice effects were noticeable, such that later lists were easier. The scores, therefore,  61  did not necessarily match the subjective reports of all participants, with discrepancies in both directions being observed across cases.  2.4 Changes made to the Pilot Procedure to Develop the Experimental Procedure Based on the results of the pilot study, the following changes were made to the experimental procedure used for the main experiment:  1)  For the main experiment, it was decided not to separate the carrier phrase from the target word but to jitter the entire phrase, thereby preserving the coarticulatory effects present in the intact stimuli and eliminating any discrepancies between word lists that may have resulted from disruptions of coarticulation. Consequently, the carrier phrase was also jittered.  2)  It was also decided that experimental participants would be given detailed instructions (see Section 3.1.5), including a warning that all words would be preceded by the phrase "Say the word..."  3)  Another major decision was that delivery of a full (50 word) list, at every presentation level, was necessary regardless of the half-list score. Full-list presentation at every intensity level was expected to reduce the anomalous rollover that occurred in the unjittered condition in the pilot study and to make the experiment more uniform across jitter conditions/intensity levels. Furthermore,  62 use of a full-list would ensure phonemic balance across word lists. The C D was paused every 25 words to allow the participant to stretch, take a sip of water, etc.  4)  The order of list presentation was counterbalanced between participants. For example, instead of presenting lists in the order 1,2,3,4, all the time, lists were presented in other orders, e.g., 3,4,1,2. Although the word lists were created to be equivalent, this design ensured that any variability due to differences between lists was evenly distributed across conditions.  5)  A few practice words, taken from the W-22 corpus of words, were presented at the participant's M C L before conducting discrimination testing using NU-6 word lists. The purpose of this practice was to familiarize participants with both the task and the talker (the same talker is used for both the W-22 and NU-6 recordings).  6)  The determination of U C L was more rigidly standardized (see Section 3.1.5). The highest presentation level, which is (UCL-5) dB HL, is crucial to accurately determining PBmin and the amount of rollover. If the highest presentation level is underestimated, significant rollover may be missed (e.g., Meyer & Mishler, 1985). Even more worrisome, if U C L is overestimated, pain to the client could result.  7)  Word discrimination lists were presented to the left ear only. The rationale for testing only one ear was to avoid having to account for variability due to right  63  ear/left ear testing differences. The reason for selecting the left ear was that, because there are more contralateral than ipsilateral projections of auditory neural fibres (e.g., Bhatnagar & Andy, 1995) and language is left-hemisphere dominant for most people (e.g., Caplan, 1987), listeners should have more difficulty for leftear presentations.  It was decided that the j l condition would not be tested in the main experiment. The main reason for this decision was that little/no rollover was found in the j l condition. Furthermore, it was decided that jitter would be applied to the frequency band of 0-1.2 kHz only. Pichora-Fuller et al., (submitted), showed that temporal jitter causes spectral splatter and that the degree of the splatter is greater at higher frequencies and at higher degrees of jitter. Thus, Pass' (1998) finding that decreases in speech intelligibility were associated with temporal jitter could have been due, at least in part, to spectral splatter and not to the simulation of neural asynchrony. In order to address this dilemma, Pichora-Fuller et al., (submitted), repeated one of Pass' (1998) conditions but instead of jittering the entire frequency band from 0-10 kHz, the authors jittered only the frequency components below 1.2 kHz. At their experimental jitter conditions (SD=0.25 msec, BW=0.5 kHz), splatter was kept 10 to 25 dB (depending on frequency) below the level of the signal. However, their results continued to show that the performance by young normals using temporally jittered stimuli resembled that of the elderly, with audiograms in the normal range, when presented with intact stimuli.  64  It was also decided that, in the main experiment, the standard deviation of the jitter would be increased from 0.25 msec, to 0.50 msec. The reason for this was to ensure that the temporal jitter would be sufficiently intense to simulate the disruption of auditory neural synchrony found in retrocochlear pathology. As noted above, the degree of spectral splatter increases not only with the frequency to be jittered, but also with the degree of jitter. Appendix D shows the spectrum of the 1 kHz pure-tone and the splatter associated with jittering a 1 kHz pure-tone at two degrees of jitter; SD=0.25 msec, BW=0.5 kHz and SD=0.50 msec, BW=0.5 kHz. It can be seen that there is a small increase in splatter in the more jittered condition but that in both conditions, the level of the splatter is kept at least 15 to 20 dB below the level of the signal.  A significant additional feature of the main experiment was the inclusion of between-band jittered word-recognition lists. W-22 word lists were jittered (SD=0.50 msec, BW=0.5 kHz) in three separate frequency bands, 0-0.4 kHz, 0.40.8 kHz and 0.8-1.2 kHz. The purpose of including the between-band jittered W22 lists was to compare the amount of rollover found to that found in the withinband jittered NU-6 word lists. Between-band jittering would reduce the degree of correlation between bands and might have an additional deleterious effect on synchrony coding across the full frequency range of the speech signal.  65  3.0 M A I N E X P E R I M E N T  3.1 Method The following sections describe the participants, materials, setting and procedures of the main experiment.  3.1.1 Participants Sixteen people participated in the main experiment; 11 females and five males. A l l participants spoke English as their first language, were between the ages of 22 and 34 (mean age = 27.3 years, standard deviation = 3.5 years) and had 16 to 25 years of education. Three of the participants were left-handed and 13 were right-handed. A l l participants had pure-tone air-conduction thresholds of 20 dB H L or better at 0.25, 0.5, 1, 2, 4, and 8 kHz, bilaterally. Ipsilateral acoustic reflexes within the normal limits (70-100 dB HL) in the test ear at 0.5, 1, and 2 kHz were obtained for all but one participant (G3S3) for whom a seal could not be maintained. Most of the participants were graduate students, but none of them were involved in the field of communication disorders. Each participant was required to give informed consent (see Appendix E) and received remuneration of $10 following completion of each experimental session.  3.1.2 Materials The original NU-6 word lists recorded on compact disc by Auditec were used for the main experiment (see sections 1.5.1.2 and 2.2.2). As in the pilot, recorded phrases  66 (carrier phrase and target word) were extracted from the original CD, sampled at 20 kHz and saved as *.SND files. However, unlike the pilot, no attempt was made to separate the carrier phrase from the target word. For the main experiment, jitter was applied to the entire phrase, for the reasons discussed in Section 2.2.4. Another major difference from the pilot was that the jitter was applied only to the frequencies below 1.2 kHz. The parameters of the jitter were SD=10 sample points=0.50 msec, and BW=0.5 kHz (the sampling rate was 20 kHz). The reasons for choosing these parameters were also discussed in Section 2.2.4. After jittering, both the unjittered and jittered stimuli were resampled at 44.1 kHz, saved as * . W A V files and written back to CD. Additionally, the Central Institute for the Deaf (CID) W-22 word lists were jittered and used in the main experiment (see Appendix F for a complete listing of W-22 words). The original W-22 materials were obtained from Auditec on the same disk as the NU-6 materials. Like the NU-6 materials, the W-22 words were preceded by the carrier phrase "Say the word..." and were recorded using the same person's voice that was used for the NU-6 words. As was done for the NU-6 material, W-22 phrases were extracted from the original CD, sampled at 20 kHz and saved as *.SND files, with no attempt made to separate the target word from the carrier phrase. Also, jitter was applied to the frequency components below 1.2 kHz for the entire phrase. However, in contrast to the NU-6 lists, jitter was applied between bands for the W-22 lists. Modifications to the Jitter program made by Jaeger (2000) allow for the jittering of up to three different frequency bands using jitter parameters (SD and BW) uniquely specified for each band. In Jitterl4.exe, a Fast Fourier Transform is used to separate the incoming signal into its component frequencies. The signal is then divided into a maximum of four frequency  67 bands according to values input by the user. The signal is then converted back to the time domain using an Inverse Fast Fourier Transform. Jitter parameters can then be applied to the lower three bands. The program jitters each band in the same way as the original jitter program (i.e., each band is jittered based on a unique L P band-limited white noise). It is not required to apply jitter to all the bands, i.e., one, two or all three bands may be jittered. The upper frequency cut-off is again limited by the sampling rate (Nyquist frequency). After jittering, all the individual bandlimited signals are added together to produce the final signal, with jitter. This type of jittering is referred to as between-band jitter. For the CID W-22 materials, the three bands to which jitter was applied were: 0-0.4 kHz, 0.4-0.8 kHz, and 0.8-1.2 kHz. The jitter parameters (SD=0.50 msec, BW=0.5 kHz) were the same for each band. The jittered W-22 stimuli were then resampled at 44.1 kHz, saved as * . W A V files and written back to C D . The word lists for each condition (intact NU-6 lists, NU-6 lists jittered withinband, and W-22 jittered between-band) were saved to a separate C D . The 1.0 kHz calibration tone was saved as the first track of each CD and was used for calibration (see Section 2.2.2.2.)  3.1.3 Apparatus and Physical Setting The facilities used for the main experiment were identical to those used for the pilot study (see Section 2.2.3).  68  3.1.4 Experimental Design The experimental design is detailed in Table 3.1 and was constructed by randomly assigning the 16 participants to one of four groups: G l , G2, G3 and G4. The first two groups (Gl and G2) received the intact NU-6 word lists in Session #1, the last two groups (G3 and G4) received the jittered NU-6 word lists in Session #1. Half of the groups ( G l and G3) received stimuli in ascending intensities; the other half (G2 and G4) received stimuli in descending intensities. There are four full NU-6 word lists (1, 2, 3, and 4) consisting of 50 words each. They were designed to be phonemically equivalent (see  Group Subject Session #1 Session #2 Jitter JO J2 40 Intensity 55 65 UCL-5 40 55 65 SI 1A 2A 3A 4A IB 2B 3B Gl S2 2A 3A 4A 1A 2B 3B 4B S3 3A 4A 1A 2A 3B 4B IB S4 4A 1A 2A 3A 4B IB 2B Jitter JO J2 Intensity UCL-5 65 55 40 UCL-5 65 55 1A 2A 3A 4A G2 SI IB 2B 3B S2 2A 3A 4A 1A 2B 3B 4B S3 3A 4A 1A 2A 3B 4B IB S4 4A 1A 2A 3A 4B IB 2B Jitter J2 JO Intensity 40 55 65 UCL-5 40 55 65 SI IB 2B 3B 4B 1A 2A 3A G3 S2 2B 3B 4B IB 2A 3A 4A S3 3B 4B IB 2B 3A 4A 1A S4 4B IB 2B 3B 4A 1A 2A Jitter J2 JO Intensity UCL-5 65 55 40 UCL-5 65 55 IB 2B 4B G4 SI 3B 1A 2A 3A S2 2B 3B 4B IB 2A 3A 4A S3 3B 4B IB 2B 3A 4A 1A S4 4B IB 2B 3B 4A 1A 2A Table 3.1 Schedule for the intact and jittered NU-6 conditions  UCL-5 4B IB 2B 3B 40 4B IB 2B 3B UCL-5 4A 1A 2A 3A 40 4A 1A 2A 3A  69  Section 1.5.1.2). However, because phonemic equivalence was based on the average performance over a large pool of listeners, slight discrepancies between lists may exist for the listeners in this study. The accurate measurement of any rollover phenomenon is highly dependent on the equivalence of the word lists used in obtaining the PI-PB function. In order to counterbalance the effects of any potential list differences, the order of list presentation was changed for each participant within a group. Over each group, and over the whole main experiment, each list was presented an equal number of times at each intensity level. Only two combinations of word order, A and B, within each NU-6 word list were needed, i.e., one for the intact condition and one for the jittered condition. Counterbalancing of intensity and word list order was also used to design the testing procedure for the jittered W-22 word lists (see table 3.2).  Group  Subject Intensity  40  SI S2 S3 S4 SI S2 S3 S4  Gl  G2  Intensity G3  G4  SI S2 S3 S4 SI S2 S3 S4  Wl W2 W3 W4 Wl W2 W3 W4  Session #2 (De-sync ironized W-22 lists) 55 65 UCL-5  W2 W3 W4 Wl W2 W3 W4 Wl  W3 W4 Wl W2 W3 W4 Wl W2  W4 Wl W2 W3 W4 Wl W2 W3  UCL-5  65  55  40  Wl W2 W3 W4 Wl W2 W3 W4  W2 W3 W4 Wl W2 W3 W4 Wl  W3 W4 Wl W2 W3 W4 Wl W2  W4 Wl W2 W3 W4 Wl W2 W3  Table 3.2 Schedule for the jittered W-22 condition  70  3.1.5 Testing Procedure As in the pilot study, the experiment was conducted during two sessions of approximately 1 hour each. The sessions were separated by at least one week to reduce the effects of practice and/or fatigue. Before the start of the first session, each participant filled out a hearing/language history form and a consent form (see Appendix E). After otoscopic screening of both ears, a tympanogram and ipsilateral acoustic reflexes were obtained for the left ear. Bilateral pure-tone air-conduction thresholds (0.25, 0.5, 1, 2, 4, and 8 kHz) were measured followed by speech testing (SRT, M C L and U C L ) on the left ear. Participants were excluded from further testing if they had pure-tone thresholds >20 dBHL in either ear at any of the frequencies tested, abnormal reflex thresholds or an abnormal (i.e., not a Type A) tympanogram. (e.g., see Stach, 1998). The importance of obtaining an accurate measure of the U C L was stressed in Section 2.2.6. The method used in the main experiment for the estimation of U C L was adapted from Stach (1998), Mueller & Bright (1994), and Martin (1997). After obtaining the M C L in the left ear, each participant was given the following instructions:  "I want to find the level that is uncomfortably loud for you. This is a level that you would be able to tolerate for only 1 or 2 minutes. If you look at the chart posted in front of you, the uncomfortable level is somewhere between annoying and extremely uncomfortable. I'm going to start talking and increasing the volume. I want you to say Stop when the level becomes uncomfortable. Are you ready?"  71  The chart referred to appears in Appendix G and was modeled after Stach (1998, pg. 450). The experimenter read the following passage at a moderate pace, talking at the M C L (keeping the V U as close to zero as possible) and increasing the presentation level, in steps of 5 dB, every 1-2 seconds (at the locations in the text indicated by the symbol t) until the U C L was reached.  "Imagine you are in the kitchen (t) listening to the news on the radio (t). Your friend in the other room (t) wants you to turn up the volume (t) so he can hear it. However, (t) you don't want to turn up the volume (t) so much that it is uncomfortable (t) for yourself. Alternatively, (t) imagine you are in a night club (t) and there are lots of people talking (t) and the DJ is playing the music (t) very loudly. At what level (t) would it be uncomfortable for you. ( t ) Imagine you are watching T V (t) with some older people (t). They want the volume turned up (t). At what point would you ( t ) have to leave the room."  The level at which the participant said "Stop" was recorded. The participant was then instructed that the procedure would be repeated two more times. During the second estimation of U C L , the beginning level was set at M C L +10 dBHL. The procedure was repeated for a third time, with the initial level at M C L - 10 dBHL. The average of the three trials, rounded to the nearest 5 dB, was taken as the U C L . In the final stage of the first session, the PI-PB function was obtained using either jittered (SD=0.50 msec, BW=0.5 kHz) or intact NU-6 stimuli according to the schedule  72 detailed in Table 2.2. However, before administering the word discrimination lists, the following instructions were given to each participant:  "You will hear some ordinary English words. When you hear a word, I want you to repeat it and print it out in the appropriate space. Don't worry about trying to spell it correctly, just spell how it sounds if you are unsure. Each word will be preceded by the phrase 'Say the word...'. You need only repeat and write down the following target word. For example, if the guy says 'Say the word truck'. You say 'truck'. Some of the words may or may not be clear. Guessing is encouraged; say and write down what you think the word might be if you are unsure. If you have no guess, just say T don't know' and put a stroke through the appropriate space. Some of the words may or may not be loud. However, they should never be uncomfortably loud. If you feel uncomfortable at any time during this procedure, let me know and the experiment will be terminated. Do you have any questions? We will start off with three practice words to get you warmed up."  Following the instructions, three intact phrases, selected from the W-22 corpus of words, were presented to each participant, at their M C L , to familiarize him/her with both the task and the talker's voice. In the second session, the PI-PB function for the untested NU-6 condition (intact or jittered) was obtained. Finally, the PI-PB function using jittered between-band desynchronized W-22 word lists was obtained. The administration sequence of the W-22 material was made as per the schedule outlined in Table 2.3. After the presentation of two  73  full lists, participants were allowed a break of up to 5 minutes. As noted in section 2.2.6, the C D was paused every 25 words to allow the participant to stretch, take a sip of water, etc. As was done in the pilot study, speech noise masking was delivered to the non-test ear at 5, 20, 30, and (UCL-40) dBHL respectively for speech presentation levels of 40, 55, 65 and (UCL-5) dBHL.  3.1.6 Scoring Procedure Scoring (correct/incorrect) was done on-line based on the participant's verbal response. If the participant's response was believed to be incorrect, this incorrect response was written on the score sheet by the experimenter. After completion of the session, results were checked with the participant's written responses. Written responses took precedence when discrepancies occurred. When written responses were illegible, verbal responses were used.  74 4.0 RESULTS OF THE M A I N E X P E R I M E N T  4.1 Introduction This chapter is devoted to describing the effects of experimental variables (jitter condition, presentation level and group), and their interactions, on participants' scores (percent correct scores, PBmax, and percentage rollover). Participants' scores are also correlated to participant age, sex, pure-tone thresholds, UCLs, acoustic reflex thresholds, and handedness. In addition, the results of an analysis of the NU-6 words (both intact and jittered) which were the most difficult for participants, are provided. At the end of this chapter, the major results of the main experiment are summarized. Appendix H contains the important raw results obtained in this experiment.  4.2  Results of the Main Experiment The following sections review participants' percentage correct, PBmax and  rollover scores. The effects (both main and interaction) of the experimental variables on these scores are discussed in separate sections. Finally, correlations of scores to participant variables are presented.  4.2.1 Percent Correct Scores For the percent correct scores, there were three possibly significant main effects: group, jitter condition, and presentation level. In the main experiment, every participant heard 50 words at each of four intensities at each of three jitter conditions. The rationale regarding the experimental design, including the differences between groups, was reviewed in Section 3.1.4.  75  The four presentation levels were 40, 55, 65, and (UCL-5) dB H L . These intensities were labeled as Level 1, Level 2, Level 3 and Level 4, respectively. The three jitter conditions were: (1)  intact NU-6 words,  (2)  NU-6 words jittered within-band (0-1.2 kHz) using jitter parameters of SD=10 sample points (equivalent to 0.50 msec, at a sampling rate of 20 kHz) and BW=0.5 kHz, and  (3)  W-22 words jittered between-band (0-0.4 kHz, 0.4-0.8 kHz, 0.8-1.2 kHz) with SD=0.50 msec, and BW=0.5 kHz.  These three conditions are referred to as NU6intact, NU6jitter, and W22jitter, respectively. Results were recorded as scores out of 50 and converted to a percentage. A total of 12 percent correct scores (4 levels x 3 conditions) were recorded for each of the 16 participants, with 4 individuals in each of 4 groups.  4.2.1.1 Effect of Group Appendix H contains the PI-PB functions for individual listeners and for each of the four groups. It can be seen that relatively small differences in percent correct scores exist between groups. This observation is supported by an analysis of variance (ANOVA) that showed no significant main effect of group on percent correct scores F(3,12)=0.84, 2=0.50. Furthermore, there was no significant interaction effect between group and jitter condition on percent correct scores, F(6,24)=1.2, p_=0.33, or between group and  76  presentation level on percent correct scores, F(9,36)=1.7, p=0.13. Therefore, results across groups will be collapsed for further discussion/analysis.  4.2.1.2 Effect of Jitter Condition Even a cursory look at Appendix H and Figure 4.1 (see Section 4.2.1.4) will reveal that jitter condition has a substantial effect on percent correct score. This observation is supported by an A N O V A which showed a significant effect of jitter condition on percent correct score F(2,24)=237.7, p<0.001. The results of a StudentNewman-Keuls test of multiple comparisons showed the following differences between jitter conditions (p<0.001):  1.  The mean percent correct score for the NU6intact condition was significantly greater than the mean score in both the NU6jitter and W22jitter conditions.  2.  There was no significant difference in mean percent correct score between the NU6jitter and W22jitter conditions.  4.2.1.3 Effect of Presentation Level Again, Figure 4.1 would lead most to believe that there is an effect of presentation level on percent correct scores, i.e., it looks like scores are moderate at Level 1, peak at Level 2 and fall off substantially by Level 4, especially in the jittered conditions. A N O V A results confirm some of these observations by showing a significant effect of presentation level on percent correct score F(3,36)=30.7, p<0.001. The results of a  77  Student-Newman-Keuls test of multiple comparisons showed the following differences between presentation levels (p_<0.01):  1.  There was no significant difference in percent correct scores between presentation levels 1, 2 and 3.  2.  The percent correct score for presentation level 4 is significantly less than the percent correct score for presentation levels 1, 2 and 3.  4.2.1.4 Interaction of Jitter Condition and Presentation Level Table 4.1 and Figure 4.1 show the mean percent correct score and corresponding standard error for each presentation level in each jitter condition. Inspection of Table 3.1 and Figure 4.1 reveals that percent correct scores peak and then drop off as the intensity level increases. The amount of this rollover seems to be mild in the NU6intact condition and more pronounced in the jittered conditions. The amount of rollover appears to be about the same between the two jittered conditions. Results from an analysis of variance support most of these observations. There was a significant interaction effect between jitter condition and presentation level F(6,72)= 6.0, p_<0.001. The results of a StudentNewman-Keuls test of multiple comparisons showed the following significant interactions between jitter condition and presentation level on percent correct scores (2<0.01):  78 1.  Percent correct scores in the NU6intact condition were the same at all presentation levels and were significantly higher than the percent correct scores obtained at the other conditions, regardless of level.  Jitter Condition  Presentation Level  Mean % Score  Standard Error  NU6intact  1  98.38  0.46  NU6intact  2  99.00  0.32  NU6intact  3  98.25  0.57  NU6intact  4  95.75  1.03  NU6jitter  1  74.13  1.59  NU6jitter  2  79.75  1.36  NU6jitter  3  76.25  2.29  NU6jitter  4  66.25  2.74  W22jitter  1  76.50  1.20  W22jitter  2  80.50  1.28  W22jitter  3  79.00  1.68  W22jitter  4  66.63  2.14  Table 4.1: Mean percentage scores  2.  Percent correct scores at the highest presentation level (level 4) were not significantly different between the NU6jitter and W22jitter conditions. Additionally, these scores were significantly lower than any of the other percent correct scores recorded.  79  3.  A l l the percent correct scores recorded at levels 1, 2 and 3 in the NU6jitter and W22jitter conditions were not significantly different from each other. Additionally, these scores were significantly lower than the percent correct scores in the NU6intact condition and significantly greater than the scores recorded at presentation level 4 in the NU6jitter and W22jitter conditions.  Figure 4.1 Mean Percent Correct Score at each Presentation Level at each Jitter Condition  100  80  4.2.1.5 Interaction of Group, Presentation Level and Jitter Condition A N O V A results confirmed that there was no significant three-way interaction between group, presentation level and jitter condition on percent correct scores.  Figure 4.2 Individual PBmax Score in each Jitter Condition  G1S1G1S2G1S3G1S4G2S1G2S2G2S3G2S4G3S1G3S2G3S3G3S4G4S1G4S2G4S3G4S4  ] NU6intact  | NU6jitter  | W22jitter  4.2.2 PBmax PBmax, the maximum score of the PI-PB function, is an important component of the rollover calculation (see Section 1.6). Three PBmax scores were recorded per participant (one per jitter condition) for a total of 48 PBmax scores in this main experiment. Figure 4.2 shows the PBmax scores obtained for each listener, in each  81  condition (see also Appendix H). For the PBmax results, there were only two possible main effects: group and jitter condition.  4.2.2.1 Effect of Group A N O V A results showed no significant effect of group on PBmax scores, F(3,12)=0.95, g=0.45. Furthermore, there was no significant interaction effect of jitter condition and group on PBmax scores, F(6,24)=l .2, g=0.33.  4.2.2.2 Effect of Jitter Condition Figure 4.3 shows the mean PBmax scores, and associated standard error bars, obtained in each of the three jitter conditions. It is clear from this figure that the PBmax score obtained in the NU6intact condition is at ceiling and is substantially greater than the PBmax score in either of the jittered conditions. A N O V A results support this observation  Figure 4.3 Mean PBmax Score in each Jitter Condition  NU6intact  NU6jitter Jitter Condition |  PBmax  W22jitter  82  and show a significant effect of jitter condition on PBmax scores, F(2,24)=l 17.4, p_<0.001. The results of a Student-Newman-Keuls test of multiple comparisons showed the following differences between jitter conditions (g<0.001):  1.  PBmax in the NU6intact condition was significantly greater than the PBmax scores in the NU6jitter and W22jitter conditions.  2.  PBmax scores were not significantly different between the NU6jitter and W22jitter conditions.  4.2.3 Rollover For each PI-PB function (one per condition, 3 per participant, 48 in total), a rollover calculation was made. Figure 4.4 shows the rollover scores obtained for each participant, in each condition. For the rollover results, there were only two possible main effects: group and jitter condition.  4.2.3.1 Effect of Group A N O V A results showed no significant effect of group on rollover, F(3,12)=0.415, 2=0.75. Furthermore, there was no significant interaction effect of group and jitter condition on rollover, F(6,24)=0.73, g=0.63.  83  Figure 4.4 Individual Rollover (%) in each Jitter Condition  4.2.3.2 Effect of Jitter Condition Figure 4.5 shows the mean rollover scores, and associated standard error bars, obtained in each of the three conditions. It is clear from this figure that the amount of rollover obtained in the jittered conditions is substantially greater than the rollover in the intact condition. This is in agreement with the observations made in Section 4.2.1.4 when inspecting Table 4.1 and Figure 4.1. A N O V A results also support this finding. There is a significant effect of jitter condition on rollover scores F(2,24)=27.9, rj<0.001. The results  84  of a Student-Newman-Keuls test of multiple comparisons showed the following differences between jitter conditions (p<0.001):  1.  Rollover in the NU6intact condition was significantly less than rollover in the NU6jitter and W22jitter conditions.  2.  Rollover scores in the NUojitter and W22jitter conditions were not significantly different from each other.  Figure 4.5 Mean Rollover (%) in each Jitter Condition  NU6intact  NU6jitter | Rollover (%)  W22jitter  85  4.2.4  Correlations PBmax and rollover scores were tested for their degree of correlation to the sex,  age, years of education, left and right pure-tone average (0.5, 1, and 2 kHz), SRT (left ear), M C L (left ear), U C L (left ear), left ipsilateral acoustic reflex threshold (0.5, 1, and 2 kHz) and handedness of each participant (see Appendix I). Rollover was significantly positively correlated to U C L in the NU6intact (r=0.62, 2<0.01) and NU6jitter (r=0.61, rj<0.05) conditions but not in the W22jitter condition (r=0.36). Rollover in the NU6intact condition was significantly negatively correlated to PBmax in the NU6intact condition (r=-0.55, p<0.05) and significantly positively correlated to rollover in the NUojitter condition (r=0.57, p<0.05). Rollover in the NU6jitter condition was significantly positively correlated to rollover in the W22jitter condition (r=0.56, p_<0.05). PBmax in the W22jitter condition was significantly negatively correlated to SRT (r=-0.64, D<0.01) and significantly positively correlated to PBmax in the NU6intact (r=0.61, p_<0.05) and NUojitter (r=0.53, p_<0.05) conditions. In general, the correlation analysis shows that rollover was sometimes, but not always, significantly positively correlated to U C L and rollover and significantly negatively correlated to PBmax. PBmax was also significantly negatively correlated with SRT, in some jitter conditions. Rollover and PBmax were not significantly correlated to such participant variables as sex, age, years of education, pure-tone average, M C L , acoustic reflex threshold or handedness.  86  4.3  Word Errors The errors made in the NU6intact and NU6jitter conditions were analysed. Table  4.2 displays the total number of errors made by all participants, on each list, for each condition. The percentage of total errors (summed across all lists) is noted in brackets.  Condition  List 1  List 2  List 3  List 4  Totals  NU6unjit  11(16%)  21(30%)  28(41%)  9(13%)  69  NU6jit  213(26%)  187(23%)  224(27%)  202(24%)  826  Table 4.2  Word errors per list, per condition.  It can be observed from Table 4.2 that the errors made by participants were not evenly distributed among the four NU-6 word lists. The distribution is more uneven in the NU6intact condition than the NU6jitter condition. This may partially be due to the limited number of errors made in the intact condition. It may have been expected that the jittering of the NU6jitter condition would have decreased the interlist equivalence, but this does not seem to have been the case. In fact, the errors between lists in the NU6jitter condition are all within 4% of each other. Since, in this experiment, we are comparing scores between lists and attributing any differences in score to presentation level, the danger in word list inequality would be inaccurate estimation of PI-PB rollover. However, the risk of inaccurate rollover estimation is mitigated due to the decision to mix the order of list presentation; the finding that error differences in the NU6jitter condition are within 4% of each other is reassuring.  87  Unjittered Errors  Jittered E r r o r s  Word  # of Errors  Word  # of Errors  Pole  9  Mill  16  Youth  7  Numb  16  Mill  6  Thumb  16  Germ  5  Pole  15  Lore  5  Laud  14  Yearn  5  Limb  14  Burn  3  Met  14  Laud  2  Calm  14  Room  2  Tell  14  Goal  2  Mob  14  23 tied  1  Peg  14  Pearl  13  Goal  13  Table 4.3 Most common N U - 6 word errors A list o f the m o s t c o m m o n w o r d errors m a d e i n the N U 6 i n t a c t a n d N U 6 j i t t e r c o n d i t i o n s w a s c o n s t r u c t e d ( s e e T a b l e 4 . 3 ) . It c a n b e s e e n , u n s u r p r i s i n g l y , t h a t s o m e o f the m o s t c o m m o n errors m a d e i n the N U 6 i n t a c t c o n d i t i o n w e r e a l s o c o m m o n l y m a d e i n the N U 6 j i t t e r c o n d i t i o n (e.g., p o l e , m i l l , l a u d , g o a l ) . H o w e v e r , there w e r e s e v e r a l w o r d s c o m m o n l y i n e r r o r i n the N U 6 j i t t e r c o n d i t i o n that w e r e n o t errors i n the intact c o n d i t i o n . These include n u m b , thumb, limb, met, c a l m , tell, m o b , peg, pearl and goal. A p p e n d i x J  88 includes the participant responses that were associated with each of the target words in Table 4.3.  4.4  Summary of Results There were no significant effects of group on the results of this experiment, i.e.,  there was no significant difference between groups in measures of PBmax, rollover or percent correct scores. Jitter condition and presentation level had a significant effect on percent correct scores. Percent correct scores in the NU6intact condition were not significantly different between presentation levels, were near ceiling, and were significantly higher than the percent correct scores for other conditions and levels. Percent correct scores at the highest presentation level (level 4) were not significantly different between the NUojitter and W22jitter conditions. Additionally, these scores were significantly lower than any of the other percent correct scores recorded. A l l the percent correct scores recorded at levels 1, 2, and 3 in the NU6jitter and W22jitter conditions were not significantly different from each other. Additionally, these scores were significantly lower than the percent correct scores in the NU6intact condition and significantly greater than the scores recorded at presentation level 4 in the NU6jitter and W22jitter conditions. Rollover was significantly greater in the NU6jitter and W22jitter conditions than the NU6intact condition. Consistent with the fact that percentage correct scores were not significantly different between presentation levels, rollover was not significant in the intact condition. PBmax was found to be significantly lower in the jittered conditions, but not to differ between the NU6jitter and W22jitter conditions.  8 9  5.0 DISCUSSION A N D CONCLUSIONS  5.1 Review of Hypotheses The main purpose of this study was to determine whether young normal-hearing participants with intact acoustic reflexes would show significantly greater PI-PB rollover when presented with temporally jittered stimuli than when presented with intact stimuli. In the process of answering this question, the following null hypotheses were tested:  1.  PI-PB rollover will not be greater than zero in the intact condition for NU-6 lists presented in quiet.  2.  PI-PB rollover in the NU6jitter condition will not be greater than rollover in the NU6intact condition for NU-6 lists presented in quiet.  3.  There will be no significant difference in PBmax between the intact and jittered condition for NU-6 lists presented in quiet.  4.  There will be no significant difference in PI-PB rollover between the NU6jitter and W22jitter conditions.  5.  There will be no significant correlation between PBmax and sex, age, years of education, left-ear and right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds, or handedness.  90  6.  There will be no significant correlation between PI-PB rollover and sex, age, years of education, left-ear and right-ear pure-tone average, SRT (test ear), M C L (test ear), U C L (test ear), ipsilateral acoustic reflex thresholds, or handedness.  5.2 Summary of Results The following sections discuss whether the null hypotheses were rejected or not rejected.  5.2.1 Null Hypothesis 1: PI-PB rollover in the NU6intact condition Results (see Section 4.2.1.4) showed that percent correct scores in the NU6intact condition were the same at all presentation levels (p<0.01). In other words, rollover was not significantly different from zero in the NU6intact condition. Thus, the null hypothesis can not be rejected. The lack of difference between WRT percent scores in the NU6intact condition was an expected finding. As noted in Section 1.5.3, PBmax is typically attained at 35-40 dB H L for normal-hearing individuals (Martin, 1997). Furthermore, for normal-hearing individuals, the PB score stays at PBmax as intensity is increased (Stach, 1998). The four presentation levels (40, 55, 65 and UCL-5 dB HL) used to create the PI-PB functions of this experiment were all sufficiently intense such that PBmax would be expected.  91  5.2.2 Null Hypothesis 2: PI-PB rollover in the NU6intact versus the NU6jitter conditions Results (see Section 4.2.3.2) show that rollover in the NUojitter condition was significantly greater than rollover in the NU6intact condition (p_<0.001). Thus, the null hypothesis can be rejected and the research hypothesis is supported. At a significance level of p_<0.001, the hypothesis that there is a greater amount of rollover in the jittered than the intact condition is very strongly supported. From Figure 4.5, it may be seen that the average rollover obtained in the NU6intact condition was just over 4.1% whereas the average rollover in the NU6jitter condition was slightly more than 16.6%. These results help to support the major claim of this thesis, i.e., that temporal jitter simulates neural asynchrony in young normal-hearing listeners.  5.2.3 Null Hypothesis 3: PBmax in the NU6intact versus the NU6jitter conditions Results (see Section 4.2.2.2) show that PBmax in the NU6intact condition was significantly greater than PBmax in the NU6jitter condition (rj<0.001). Thus, the null hypothesis can be rejected and the research hypothesis is supported. The decreased PBmax associated with the NU6jitter condition was an expected finding. Recall that in the pilot study (see Chapter 2), the milder jitter condition, j l (SD=0.25 msec, BW=0.1 kHz), produced little/no rollover and the PBmax associated with j l was not substantially less than the PBmax associated with the unjittered condition (jO). It was thought that the PBmax score may be an indicator of how much rollover to expect, i.e., if the auditory system was sufficiently stressed in terms of the asynchrony of neural timing, then this would be reflected in the PBmax score. As noted in Section 2.4, one of the reasons for increasing the SD of the jitter, from 0.25 msec, in the pilot to 0.50  92  msec, in the main experiment, was to ensure an adequate disruption of neural synchrony. Correspondingly, it was expected that PBmax would be substantially (and significantly) decreased in the jittered conditions of the main experiment. From Figure 4.3 it can be seen that whereas PBmax in the NU6intact condition averaged nearly 100%, it dropped to just over 80% in the NU6jitter condition.  5.2.4 Null Hypothesis 4: PI-PB rollover in the NU6jitter versus the W22jitter conditions Results (see Section 4.2.3.2) show that rollover scores in the NU6jitter and W22jitter conditions were not significantly different from each other (p_<0.001). Thus, the null hypothesis can not be rejected. Figure 4.5 reveals that rollover in the NU6jitter condition was just over 16.6% and rollover in the W22jitter condition was just over 16.0%. The lack of difference in rollover between the two jitter conditions is somewhat surprising. The NU6jitter and W22jitter conditions are different because 1) they are composed of different words and 2) the NU6jitter is applied within-band only whereas the W22jitter is applied both withinband and between-band. Because the NU-6 and W-22 lists are composed of different words, they have a different frequency composition. Thus, applying jitter within the band 0-1.2 kHz may jitter the NU-6 and W-22 word lists to different extents depending on how much of their frequency composition falls within the 0-1200 Hz band. This difference in jitter would also be expected to lead to a different amount of neural asynchrony and a different amount of PI-PB rollover. It is possible that the frequency differences (within the 0-1.2  kHz band) between the NU-6 and W-22 word lists were so slight that they led to no significant differences in rollover. The main factor that was expected to result in rollover differences between the two jitter conditions of this experiment is that the W22jitter condition additionally contains between-band jitter. Greenberg and Arai (1998) divided the speech signal up into frequency bands and then time shifted those bands with respect to each other. This type of manipulation can be thought of as a type of between-band desynchrony. They found that speech intelligibility, in quiet, remained high (above 75%), even for shifts of up to 140 ms. Their results suggest that between-band desynchronization, obtained by time-shifting frequency bands with respect to each other, does not substantially reduce a listener's ability to understand spoken language in quiet. The jitter used in the W22jitter condition of the present study is similar to the type of between-band desynchronization introduced in the Greenberg and Arai (1998) study insofar as frequency bands were temporally desynchronized with respect to each other. Although the methods of betweenband desynchronization are different, the findings of this experiment are consistent with those of Greenberg and Arai (1998) insofar as no evidence was found that between-band desynchronization reduced listeners' ability to understand words in quiet to a greater extent than within-band desynchronization. Both the jitter conditions of the present study contain within-band jitter. The major differences between the two conditions are that their lists are composed of different words and the W22jitter condition contains betweenband (as well as within-band) jitter. These differences did not lead to a significant difference in PBmax score between the NU6jitter and W22jitter conditions.  94  5.2.5 Null Hypothesis 5: Correlations between PBmax and Participant Variables Results show that the only significant correlation was a negative correlation between PB max in the W22jitter condition and SRT (r=-0.64, g<0.01). This result is somewhat unexpected and is considered anomalous (e.g., one would also expect a significant negative correlation in the NU6jitter condition). Except for this isolated significant correlation, there were no other significant correlations between PBmax and participant variables. A total of 48 correlations between PBmax and participant variables were examined and the finding of one significant correlation is therefore likely to be due to chance. For all the other participant variables besides SRT, the null hypothesis can not be rejected.  5.2.6 Null Hypothesis 6: Correlations between PI-PB rollover and Participant Variables Rollover was significantly positively correlated to U C L in the NU6intact (r=0.62, p_<0.01) and NU6jitter (r=0.61, p_<0.05) conditions. This is not a surprising finding because Meyer and Mishler (1985) noted that if the highest presentation level is underestimated, significant rollover may be missed. Because U C L estimation relies on subjective judgments by the participant, it is possible that those who were conservative in their judgments of U C L may have shown a correspondingly conservative amount of PIPB rollover. It was expected that rollover would be significantly correlated to U C L in the W22jitter condition but this was not the case. However, the observed significance level (r=0.36) was in the right direction and close to the minimum correlation (r=0.437) necessary for significance (p<0.05).  95  Except for the significant correlations with U C L noted above, there were no other significant correlations between rollover and participant variables. Thus, for all the other participant variables besides U C L , the null hypothesis can not be rejected. This was an expected finding.  5.3 Conclusions and General Discussion In the present study, when young normal-hearing participants were presented with temporally jittered PB word lists in quiet, they showed significantly more PI-PB rollover than when presented with intact stimuli in quiet. Temporal jitter was able to simulate the type of speech perception deficits associated with retrocochlear disorder, i.e., the drop in performance at high intensities. PI-PB rollover is rare in the general clinical population (e.g., Jerger & Jerger, 1971). It is more commonly exhibited by acoustic neuroma patients (e.g., Jerger & Jerger, 1971), the elderly (e.g., Shirinian & Arnst, 1980), and stapedectomized patients (e.g., McCandless & Goering, 1974). Findings of PI-PB rollover have been attributed to neural causes, mechanical factors associated with the absence of acoustic reflexes, and both neural and mechanical factors (e.g., Hannley & Jerger, 1981). The PI-PB rollover found in the present study can not be attributed to mechanical factors because the participants had intact acoustic reflexes. Therefore, the rollover found in the jittered conditions is more likely due to simulated neural factors. Clinically, PI-PB rollover is associated with neural asynchrony based on findings in the acoustic neuroma, elderly and MS populations (see Section 1.6.1.2). Theoretically, the A L S R predicts the presence of rollover if there is neural asynchrony (see Section 1.4.4.4.3). Thus, there is  96 both clinical and theoretical backing to support the claim that the temporally jittered stimuli of the present study simulated neural asynchrony, of the type found in acoustic neuroma patients who exhibit rollover, in young normal-hearing listeners. Pichora-Fuller et al., (submitted), were the first to argue that external, temporal jitter simulates internal, neural asynchrony in young normals. In their studies, high- and low-context SPIN-R sentences were temporally jittered and used to test young, normal hearing adults in a range of S/N ratios. The performance, especially in low S/N ratios, of the young normals resembled the performance on intact SPIN-R stimuli by older adults with pure-tone audiograms in the normal range. Because one of the possible explanations for the relatively poor performance of these older adults is the existence of a temporal processing deficit (e.g., neural asynchrony) (Pichora-Fuller & Schneider, 1992), the authors concluded that the performance by the young-normals may be a result of the simulation of internal neural asynchrony. It is possible that the results of Pichora-Fuller et al. (submitted) may be due to something other than the simulation of neural asynchrony. One possibility that was addressed by the authors is the introduction of spectral splatter. Temporal jitter introduces spectral splatter that is greater at higher frequencies and for higher degrees of jitter. To counter the effects of spectral splatter, the authors performed a study in which jitter was applied to the frequencies below 1.2 kHz only. The authors argue that the noise introduced by splatter was low enough such that it would not have significantly affected experimental results. Another possibility is that temporal jitter may be simulating a type of auditory processing deficit other than, or in addition to, neural asynchrony and it is this other deficit that may be causing the decreased performance by young normals.  The present study addresses both of these concerns. First, the claim of this thesis (i.e., that temporal jitter simulates neural asynchrony) should not be affected by the issue of spectral splatter. That is, PI-PB rollover is evidenced by changes in WRT score with presentation level. Any potential decrease in W R T score caused by frequency distortion (and not temporal asynchrony) should not affect the calculation of rollover. Although an increased amount of spectral splatter would be expected to decrease the PBmax percent score, it would not be expected to differentially affect W R T scores at higher intensities because the S/N ratio would remain constant despite changes in presentation level. The relative inability of spectral distortion to affect the rollover calculation is one of the key strengths that the rollover method has in the investigation of the ability of temporal jitter to simulate neural asynchrony. The second concern with the Pichora-Fuller et al., (submitted), study is that temporal jitter could be simulating a deficit other than neural asynchrony. The advantage of the rollover method is that there is both clinical and theoretical data that suggest a strong association between rollover and neural asynchrony.  5.4 Future Research Directions In the Pichora-Fuller et al., (submitted), study, the performance by young adults (using jittered stimuli) was most severely affected in low S/N ratios. That is, supposedly, noise introduced stress into the auditory system and forced it to rely on synchrony information. This fits with the observation that synchrony information is more robust in the presence of background noise than firing rate information (e.g., Greenberg, 1996). The present study has its backing in the hypothesis (i.e., the ALSR) that the human auditory system is more reliant on synchrony information at high, versus low,  98  intensity levels. Instead of using noise, intensity was used as a parameter by which to stress the auditory system and force the reliance on synchrony information. Future studies should look at the combination of the effects of high intensity and low S/N ratios on the perception of speech that is jittered both within-band and between-bands. For example, as part of an M.Sc. thesis, PI-PB functions could be created for young normals at different S/N ratios using jittered and unjittered NU-6 word lists. It is predicted that they will show an even greater degree of rollover than was observed in the present study. The present study, together with the findings of Pichora-Fuller et al., (submitted), provides substantial support for the claim that external temporal jitter simulates internal neural asynchrony. Future studies should look at determining thresholds for detecting externally introduced jitter in a signal in young normal-hearing listeners. Older adults, or any other population with presumed neural asynchrony, would be expected to have relatively poor jitter detection thresholds. Once normal limits are established for the normal-hearing population, other populations (e.g., older adults) can be tested to see if their level of neural synchrony is normal or abnormal. In the future, the ability to identify patients with neural asynchrony could have consequences for their aural rehabilitation.  99 REFERENCES  Abel, S.M., Krever, E . M . , & Alberti, P.W. (1990). Auditory detection, discrimination and speech processing in aging, noise-sensitive and hearing-impaired listeners. Scandinavian Audiology, 19,43-54. Beattie, R.C., Svihovec, D.A., & Edgerton, B.J. (1978). Comparison of speech detection and spondee thresholds and half- versus full-list intelligibility scores with M L V and taped presentation of NU-6. Journal of the American Auditory Society, 3, 267-272. Bhatnagar, S.C., & Andy, O.J. (1995). Neuroscience for the Study of Communicative Disorders. Baltimore, M D : Williams & Willkins. Bess, F.H., Josey, A.F., & Humes, L.E. (1979). Performance intensity functions in cochlear and eighth nerve disorders. American Journal of Otolaryngology, 1 (1), 27-31. Blackburn, C.C., & Sachs, M . B . (1990). The representation of the steady-state vowel sound [£] in the discharge patterns of cat anteroventral cochlear nucleus neurons. Journal of Neurophysiology. 63, 1191-1212. Borg, E. (1968). A quantitative study of the effect of the acoustic stapedial reflex on sound transmission through the middle ear of man. Acta. Otolaryngologica. 66, 461-472. Brown, S. (2000). Temporal jitter mimics effects of aging on word identification and word recall in noise. [M.Sc. thesis]. Vancouver: The University of British Columbia. Campbell, R.A. (1965). Discrimination and word test difficulty. Journal of Speech and Hearing Research, 8, 13-22. Caplan, D. (1987). Cerebral dominance and specialization for language. In D. Caplan, Neurolinguistics and linguistic aphasiology: A n introduction (pp.345-368). New York: Cambridge University Press. C H A B A : Committee on Hearing, Bioacoustics, and Biomechanics, Working Group on Speech Understanding and Aging. (1988). Speech understanding and aging. Journal of the Acoustical Society of America, 83, 859-895. Chmiel, R., & Jerger, J. (1996). Hearing aid use, central auditory disorder, and hearing handicap in elderly persons. Journal of the American Academy of Audiology, 7, 190-202. Craig, S. (2000). Goldwave Version 4.12 [computer program]. St. John's, NF: Goldwave. Dallos, P. (1996). Overviewxochlear neurobiology. In P. Dallos, A . N . Popper, & R.R. Fay (Eds.), The Cochlea (pp. 1-43). New York: Springer.  100  Davis, H., Morrical, K.C., & Harrison, C E . (1949). Memorandum on recording characteristics and monitoring of word and sentence tests distributed by Central Institute for the Deaf. The Journal of the Acoustic Society of America, 21, 552-553. Delgutte, B. (1996). Physiological models for basic auditory percepts. In R.R. Fay & A . N . Popper (Eds.), Auditory computation (pp. 157-220). New York: Springer. Dirks, D.D., Kamm, C , Bower, D., & Betsworth, A . (1977). Use of performanceintensity functions for diagnosis. Journal of Speech and Hearing Disorders, 42, 408-415. Dorman, M.F., Marton, K., Hannley, M.T., & Lindholm, J.M. (1985). Phonetic identification by elderly normal and hearing-impaired listeners. Journal of the Acoustical Society of America, 77, 664-670. Dorman, M.F., Raphael, L.J., & Liberman, A . M . (1979). Some experiments on the sound of silence in phonetic perception. Journal of the Acoustical Society of America, 74,11361144. Drullman, R. (1995). Temporal envelope and fine structure cues for speech intelligibility. Journal of the Acoustical Society of America, 97 (1), 585-592. Dubno, J.R., Dirks, D.D., & Morgan, D.E. (1984). Effects of age and mild hearing loss on speech recognition in noise. Journal of the Acoustical Society of America, 76 (1), 8796. Durlach, N.I. (1972). Binaural signal detection: Equalization and cancellation theory. In J.V. Tobias (Ed.), Foundations of Modern Auditory Theory, Vol.2 (pp.374-462). New York: Academic Press. Egan, J.P. (1948). Articulation testing methods. Laryngoscope, 58, 955-991. Eldert, E., & Davis, H., (1951). The articulation function of patients with conductive deafness. Laryngoscope, 61, 896-909. Elpern, B.S. (1961). The relative stability of half-list and full-list discrimination tests. Laryngoscope, 70, 1560-1565. Fletcher, H . (1953). Speech and Hearing in Communication. Princeton, NJ: Van Nostrand. Fletcher, H., & Steinberg, J.C. (1930). Articulation testing methods. Journal of the Acoustical Society of America, 1,1-97. Franks, J., & Daniloff, R. (1973). A review of the audiological implications of testing vowel perception. Journal of Auditory Research, 13, 355-368.  101  French, N.R., & Steinberg, J.C. (1947). Factors governing the intelligibility of speech sounds. Journal of the Acoustical Society of America, 19, 90-119. Gang, R.P. (1976). The effects of age on the diagnostic utility of the rollover phenomenon. Journal of Speech and Hearing Disorders, 31, 63-69. Geisler, C D . (1998). From sound to synapse. New York: Oxford University Press. Gelfand, S.A. (1975). Use of the carrier phrase in live voice speech discrimination testing. Journal of Auditory Research, 15, 107-110. Gordon-Salant, S., & Fitzgibbons, P.J. (1995). Recognition of multiply degraded speech by young and elderly listeners. Journal of Speech and Hearing research, 38,1150-1156. Greenberg, S. (1996). Auditory Processing of Speech. In N.J. Lass (Ed.), Principles of Experimental Phonetics (pp. 362-407). St. Louis: Mosby. Grose, J.H. (1996). Binaural performance and aging. Journal of the American Academy of Audiology, 7,168-174. Grose, J.H., Poth, E.A., & Peters, R.W. (1994). Masking level differences for tones and speech in elderly listeners with relatively normal audiograms. Journal of Speech and Hearing Research, 37, 422-428. Handel, S. (1989). Listening. Cambridge, M A : MIT Press. Hannley, M . , & Jerger, J. (1981). PB rollover and the acoustic reflex. Audiology, 20, 251-258. Hannley, M . , & Jerger, J. (1985). Patterns of phoneme identification error in cochlear and eighth-nerve disorders. Audiology, 24, 157-166. Haubert, N . (1999). Speech perception by older listeners: Contributions of auditory temporal processes. [M.Sc. thesis]. Vancouver: The University of British Columbia. Hirsh, I.J., Reynolds, E.G., & Joseph, M . (1954). Intelligibility of different speech materials. Journal of the Acoustical Society of America, 26, 530-538. Houtgast, T., & Steeneken, H.J. (1973). The modulation transfer function in room acoustics as a predictor of speech. Acustica, 28(1), 66-73. Humes, L.E. (1996). Speech understanding in the elderly. Journal of the American Academy of Audiology, 7,161-167.  102  Jacobson, J.T. (1994). Prelude to auditory evoked potentials. In Jacobson, J.T. (Ed.), Principles and Applications in Auditory Evoked Potentials (pp.3-19). Needham Heights: Allyn & Bacon. Jacobson, G.P., Jacobson, J.T., Ramadan, N . , & Hyde, M . (1994). Auditory brainstem response measures in acoustic nerve and brainstem disease. In Jacobson, J.T. (Ed.), Principles and Applications in Auditory Evoked Potentials (pp.387-426). Needham Heights: Allyn & Bacon. Jaeger, C P . (2000). Jitterl4.exe [computer program]. Vancouver, CJA Consulting. Jerger, J., & Jerger, S. (1971). Diagnostic significance of PB word functions. Archives of Otolaryngology, 93.573-580. Konig, E. (1957). Pitch discrimination and age. Acta Otolaryngologica, 48,475-489. Kricos, P.B. (1995). Characteristics of the Aged Population. In Kricos, P.B. & Lesner, S.A. (Eds). Hearing care for the older adult: Audiologic rehabilitation(pp. 1-21). Boston: Butterworth-Heinemann. Lehiste, I., & Peterson, G.E. (1959). Linguistic considerations in the study of speech intelligibility. Journal of the Acoustical Society of America, 31, 280-286. Lieberman, P. (1977). Speech Physiology and Acoustic Phonetics. New York: Macmillan. Lynn, J.M., & Brotman, S.R. (1981). Perceptual significance of the CID W-22 carrier phrase. Ear and Hearing, 2, 95-99. Martin, F.N. (1997). Introduction to Audiology (sixth edition). Boston, M A : Allyn and Bacon. Martin, F.N., & Forbis, N.R. (1978). The present status of audiometric practice: a followup study. Journal of the American Speech and Hearing Association, 20,531-541. Martin, F.N., Hawkins, R.R., & Bailey, H.A.T. (1962). The nonessentiality of the carrier phrase in phonetically balanced (PB) word testing. Journal of Auditory Research, 2,319322. McCandless, G.A., & Goering, D . M . (1974). Changes in loudness after stapedectomy. Archives of Otolaryngology, 100,344-350. McCandless, G.A., & Schumacher, M . H . (1979). Auditory dysfunction with facial paralysis. Archives of Otolaryngology, 105,271-274.  103  Merrell, H.B., & Atkinson, C J . (1965). The effect of selected variables upon discrimination scores. Journal of Auditory Research, 5,285-292. Meyer, D.H., & Mishler, E.T. (1985). Rollover measurements with Auditec NU-6 word lists. Journal of Speech and Hearing Disorders, 50,351-356. Miller, M . L , & Sachs, M . B . (1983). Representation of stop consonants in the discharge patterns of auditory-nerve fibers. Journal of the Acoustical Society of America, 74, 502512. Moore, B.C.J. (1989). A n introduction to the psychology of hearing (third edition). San Diego, C A : Academic Press Inc. Moore, B.C.J., & Peters, R.W. (1992). Pitch discrimination and phase sensitivity in young and elderly subjects and its relationship to frequency selectivity. Journal of the Acoustical Society of America, 91(5), 2881-2893. Mueller, H.G., & Bright, K . E . (1994). Selection and verification of maximum output. In M . Valentine (Ed.), Strategies for selecting and verifying hearing aid fittings (pp. 38-63). New York: Thieme Medical Publishers, Inc. Nabelek, A . K . , & Robinson, P.K. (1982). Monaural and binaural speech perception in reverberation for listeners of various ages. Journal of the Acoustical Society of America, 71(5), 1242-1248. Nelson, D.A., & Chaiklin, J.B. (1970). Writedown versus talkback scoring and scoring bias in speech discrimination testing. Journal of Speech and Hearing Research, 13,645654. Northern, J.L., & Hauler, K . W . (1974). Evaluation of four speech discrimination test procedures on hearing impaired patients. Journal of Auditory Research, Supplement 1,137. Owens, E., Talbott, C , & Schubert, E. (1968). Vowel discrimination of hearing impaired listeners. Journal of Speech and Hearing Research, 11,648-655. Pass, H.E. (1998). The importance of temporal synchrony in word recognition. [M.Sc. thesis]. Vancouver: The University of British Columbia. Patuzzi, R. (1996). Cochlear micromechanics and macromechanics. In R.R. Fay & A . N . Popper (Eds.), The cochlea. New York: Springer-Verlag. Penrod, J.P. (1994). Speech threshold and word recognition/discrimination testing (pp. 147-164). In J. Katz (Ed.), Handbook of Clinical Audiology (fourth edition). Baltimore, M D : Williams & Wilkins.  104  Phillips, D.P. (1995). Central auditory processing: A view from auditory neuroscience. The American Journal of Otology, 16(3), 338-352. Phillips D.P., & Carr M . M . (1998). Disturbances of loudness perception. Journal of the American Academy of Audiology, 9.371 -379. Pichora-Fuller, M . K . (1997). Language comprehension in older listeners. Journal of Speech-Language Pathology and Audiology, 21,125-139. Pichora-Fuller, M . K . , Schneider, B.A., Pass, H.E., & Brown, S. (Submitted). Temporal jitter disrupts speech intelligibility and memory. Pichora-Fuller, M . K . , & Schneider, B.A. (1991). Masking-level differences in the elderly: A comparison of antiphasic and time-delay dichotic conditions. Journal of Speech and Hearing Research, 34.1410-1422. Pichora-Fuller, M . K . , & Schneider, B.A. (1992). The effect of interaural delay of the masker on masking-level diferences in young and old adults. Journal of the Acoustical Society of America. 91(4). 2129-2135. Pichora-Fuller, M . K . , Schneider, B.A., & Daneman, B . A . (1995). How young and old adults listen to and remember speech in noise. Journal of the Acoustical Society of America. 97(1). 593-608. Pickles, J. (1988). A n introduction to the physiology of hearing (second edition). San Diego: Academic Press Inc. Rappaport, J.M., Gulliver, J.M., Phillips, D.P., VanDorpe, R.A., Maxner, C.E., & Bhan, V . (1994). Auditory temporal resolution in multiple sclerosis. Journal of Otolaryngology, 23,307-324. Rintelmann, W.F. (1985) Monaural speech tests in the detection of central auditory processing disorders. In Pinheiro, M . L . , & Musiek F.E. (Eds.), Assessment of central auditory dysfunction: Foundations and clinical correlates. Baltimore: Williams and Wilkins. Ross, M . , & Huntington, D.A. (1962). Concerning the reliability and equivalency of the CID W-22 auditory tests. Journal of Auditory Research. 2. 220-228. Sachs, M.B., Winslow, R.L., & Blackburn, C.C. (1988). Representation of speech in the auditory periphery (pp. 747-774). In Edelman, G . M . , Gall, W., E., & Cowan, W . M . (Eds.), Auditory function: Neurobiological bases of hearing. New York: Wiley. Schneider, B.A. (1997). Psychoacoustics and aging: implications for everyday listening. Journal of Speech-Language Pathology and Audiology, 21, 111-124.  105  Schneider, B.A. (1997). Jitspin2.exe [computer program]. Toronto: University of Toronto. Schneider, B.A., & Hamstra, S.J. (1999). Gap detection as a function of tonal duration for younger and older listeners. Journal of the Acoustical Society of America, 106(1), 371380. Schneider, B.A., Pichora-Fuller, M . K . , Kowalchuk, D., & Lamb, M . (1994). Gap detection and the precedence effect in young and old adults. Journal of the Acoustical Society of America, 95(2), 980-991. Schuknecht, H., & Woellener, R. (1955). A n experimental and clinical study of deafness from lesions of the cochlear nerve. Journal of Laryngology, 69,75-97. Schumaier, D.R., & Rintelmann, W.I. (1974). Half-list vs full-list discrimination testing in a clinical setting. Journal of Auditory Research, Supplement 2,16-17. Shamma, S. (1985). Speech processing in the auditory system: I. representation of speech sounds in the responses of the auditory nerve. Journal of the Acoustical Society of America, 78, 1612-1621. Shannon, R.V., Zeng, F., Kamath, V., Wygonski, J., & Ekelid, M . (1995). Speech recognition with primarily temporal cues. Science, 270, 303-304. Shirinian M.J., & Arnst D.J. (1980). PI-PB rollover in a group of aged listeners. Ear and Hearing 1,50-53. Snell, K . B . (1997). Age-related changes in temporal gap detection. Journal of the Acoustical Society of America. 101(4), 2214-2220. Stach, B.A. (1998). Clinical audiology: A n introduction. San Diego: Singular Publishing Group, Inc. Stach, B.A., Delgado-Vilches, G., & Smith-Farach, S. (1990). Hearing loss in multiple sclerosis. Seminars in Hearing, 11,221-230. Stockard, J.J., & Rossiter, V.S. (1977). Clinical and pathological correlates of brain stem auditory response abnormalities. Neurology, 27,316-325. Thorndike, E.L., & Lorge, I. (1944). The Teacher's Word Book of 30,000 Words. New York: Columbia University Press. Thornton, A.R., & Raffin, M.J.M. (1978). Speech-discrimination scores modeled as a binomial variable. Journal of Speech and Hearing Research, 21,507-518.  106  Tillman, T.W., & Carhart, R. (1966). A n expanded test for speech discrimination utilizing C N C monosyllabic words: Northwestern University Auditory Test No. 6. Technical report no. SAM-TR-66-55. U S A F School of Aerospace Medicine, Brooks Air Force Base, T X . Tillman, T.W., Carhart, R., & Wilber, L. (1963). A test for speech discrimination composed of C N C monosyllabic words: Northwestern University Auditory Test No. 4. Technical report no. SAM-TDR-62-135. U S A F School of Aerospace Medicine, Brooks Air Force Base, T X . Turner, C.W., Souza, P.E., & Forget, L . N . (1995). Use of temporal envelope cues in speech recognition by normal and hearing-impaired listeners. Journal of the Acoustical Society of America, 97(4), 2568-2576. Viemeister, N.F., & Plack, C.J. (1993). Time analysis (pp.116-154). In W.A. Yost, A . N . Popper, R.R. Fay (Eds.), Human Psychophysics. N Y : Springer-Verlag New York, Inc. Von Bekesy, G. (1960). Experiments in Hearing. New York, N Y : McGraw-Hill. Weaver M . , & Northern J. (1976). The acoustic nerve tumor (pp. 185-186). In J. Northern (Ed.), Hearing Disorders. Boston: Little Brown & Co. Willot, J.F. (1991). Aging and the auditory system: Anatomy, physiology and psychophysics. San Diego, C A : Singular. Wingfield, A. (1996). Cognitive factors in auditory performance: context, speed of processing, and constraints of memory. Journal of the American Academy of Audiology, 7,175-182. Wormald P.J., Rogers C , & Gatehouse S. (1995). Speech discrimination in patients with Bell's palsy and a paralysed stapedius muscle. Clinical Otolaryngology, 20,59-62. Yost, W.A. (1994). Fundamentals of hearing: A n introduction (third edition). San Diego, C A : Academic Press, Inc. Young, E.D., & Sachs, M . B . (1979). Representation of steady-state vowels in the temporal aspects of the discharge patterns of auditory-nerve fibers. Journal of the Acoustical Society of America, 66, 1381-1403.  107 APPENDIX A  AUDITEC OF ST. LOUIS NU Auditory Test #6 LIST IB  1. burn  2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50.  lot  sub  home dime which keen yes boat sure hurl door kite sell nag take fall week death love tough gap moon choice king size pool vine chalk laud goose shout fat puff jar reach rag mode tip page raid raise bean hash limb third jail knock whip met  LIST 2B  .1. l i v e •2. voice .3. ton •4. l e a r n •5. match . 6 . chair •7. deep •8. pike • 9 . room •10. read *11. calm 12. book -13. dab •14. l o a f •15. goal •16. shack •17. far •18. witch •19. rot •20. pick •21. f a i l •22. s a i d • 23. wag •24. haze •25. white •26'. hush •27. dead •28. pad •29. mill • 30. merge •31. j u i c e •32. keg •33. gin • 34. nice •35. numb •36. chief •37. gaze •38. young .39. keep •40. tool • 41. soap •42. hate •43. turn •44. rain •45. shawl •46. bought •47. thought •48. b i t e •49. lore •50. south  LIST 3B  1. sheep 2. cause 3. rat 4. bar 5. mouse 6 . talk 7. h i r e 8. search 9. luck 10. cab 11. rush 12. f i v e 13. team 14. p e a r l 15. soup 16. h a l f 17. chat 18. road 19. pole 20. phone 21. l i f e 22. pain 23. base 24. mop 25. mess 26. germ 27. thin 28. name 29. ditch 30. t e l l 31. cool 32. seize 33. dodge 34. youth 35. h i t 36. l a t e 37. jug 38. wire 39. walk 40. date 41. when 42. r i n g 43. cheek 44. note 45. gun 46. beg 47. void 48. s h a l l 49. l i d 50. good  LIST 4B . 1. rose 2. dog 3. time 4. such 5. have 6. mob 7. bone 8. s a i l 9. rough 10. dip 11. j o i n 12. check 13. wheat 14. thumb 15. near 16. lease 17. yearn 18. kick 19. get 20. l o s e 21. k i l l 22. f i t 23. judge 24. should 25. pass 26. back 27. h a l l 28. bath 29. t i r e 30. peg 31. perch 32. chain 33. make 34. long 35. wash 36. food 37. mood 38. neat 39. tape 40. ripe 41. hole 42. gas 43. came 44. vote 45. lean 46. red 47. d o l l 48. s h i r t 49. sour 50. wi f e  108  APPENDIX B Jitter 14.exe Main Screen  flfcA I i i f WINDOWS  Qj Start Menu H i SYSTEM iM T E M P  : :  'HH Temporary Internet Flat £gtwain_32  |||  1  © BACKUP H i COMMAND p CONFIG  ;  »  I!  I3©namc fnocfiRof '  f l ^  E x t e n d {3 char* m3x) j  srK  Eft tfffenttwmoditief « L « t " « i He sound *nd produces output Be Lsound snd  MNHi  j  '4ita Parameter? Standajd deviation  Boundary value* for the f rtquency range* (HjJ:  Band *1l§fToi  ttf Boundary J1200"  8ar<d2  jo  2nd Boundary jo  Band 3  fo  3rd Boundary [cT*  a seconds batae seritarce t-egm* ;  Nutob* of Ife* , — toproee** |50  Rut.  Noise  rwndwtdttiiii^ 5anyfingRata(H#:  ^20000  liiiii  I  109  APPENDIX C  Hearing and Language History Date: First Name: Birthday (day/mo/yr):  Last Name:  What is your first language? Do you speak any other languages fluently? Are you right or left handed? How many years of formal schooling have you had?  L  R  Do you know how to play a musical instrument? Have you had any training in music?  Y Y  N N  Do you think you have normal hearing? Do you hear better in one ear versus another Do people ever complain about your hearing? Have you ever had a hearing test? Are you exceptionally sensitive to loud sounds?  Y L Y Y Y  N R N N N  Has anyone in your family had a hearing loss before old age? Who?  Y  N  Do you often get colds? Do you have one now? Do you have allergies? Are you bothered by one now? Do you often get ear infections? Do you have one now?  Y Y Y Y Y Y  N N N N N N N  Have you ever had ear surgery? What kind? Do you have ringing in your ears? When? Which ears? Have you been exposed to excessive noise? What kinds? Do you regularly take any medication? What kind? Do you have trouble with your vision? What kind? W i l l your participation be affected by a health problem? What kind?  Always L Y  N Sometimes R N  Y  N  Y  N  Y  N  110  APPENDIX D Spectral Splatter of a 1 kHz tone This appendix contains three figures. The first figure is the spectrum of an unjittered 1 kHz pure-tone. This tone was the calibration tone obtained on the first track of the Auditec C D . It was extracted from the CD, sampled at 20 kHz and then resampled at 44.1 kHz. The second figure shows the spectrum of a 1 kHz pure-tone jittered at SD=0.25 msec, BW=500 Hz. The third figure is of a 1 kHz pure-tone jittered at SD=0.50 msec, BW=500 Hz.  Unjittered 1 kHz tone  caa  EiaiaEiiargB3Bigi^;Eint3BiBiaana  19.H 17.6  -BB -64  E  -BB  Q3.Z  U C  -75  ' B . B  k  -75  z  -B3  t.«  -B7  2.2 B.B  5B2  12B3  15B4 T i n t <M«)  21 BB  24B7  27BT  aigniaBiBiggnn  lKHZ-Uaveforw P5.BI  U B.B o-S.B 4 B.B  3B1  BB2  5BZ  12B3  15B4 T i m (Hi)  lKHZ-Data Processed  21 BE,  24B7  27B7  lKHZ-Cross-Sect i on  P5.B  B  CURSOR AT Tine : 1443.3 ms Freq : N/A Mag. : M/H  -71  N'-B  »-a.B  • SPECTROGRftH I Status; looz  w  w  w  w  v  w  £  PARAMETERS : AC : 512 pts : 256 : 50 Z : MANNING : 44.10 kHz : 98.0 X J 15  5 PEAKS F(Hz> 947.0 18949.0 20930.0 N/A N/A  M(dB> -5G.50 -79.30 -84.70 N/A N/A  MARKER AT F(Hz) M(dB)  39  -9fl 1444 1 44E 1 4471 M B 1 445 145B 1 451 1 453 1 (5414' TIM <H«>  FILE Proc. Hind. Bands DVlp. Func. SF Pre. Order  2.2  4.4 6.6 B.B 1 1 . B 1 3 . 2 1 5 . 4 1 7 . B 1 5 . B 2 a rt-»qu«ncH (kHz)  Jittered 1 kHz tone, SD=0.25 msec, BW=500 Hz  • SPECTROGRAH I Status: 100Z CURSOR AT Time : 1443.2 ms Freq : N/A Mag. : N/A  F I L E PARAMETERS Proc. AC Hind. 512 pts Bands 256 Ovlp. 50 Z MANNING Func. 44.10 kHz SF 98.0 Z Pre. 15 Order  5 PEAKS F(Hz> 1464.0 G632.0 9560.0 18260.0 219G3.0 MARKER AT F(Hz)  M<dB> -54.00 -72.30 -75.30 -77.60 -77.40  M<dB>  112  Jittered 1 kHz tone, SD=0.50 msec, BW=500 Hz  ™  aai3ia'wr^BaBLiit?EinHasaann  •  SPECTROGRflH  status:  IOOZ  CURSOR AT Tine : 1442.9 ns Freq : NVfl Hag. : N/A  F I L E PARAMETERS Proc. ! AC Hind. ! 512 pts Bands ; 256 Oulp. : 50 z Func. ', HANNING SF !: 44.10 kHz Pre. j 98.0 X Order J 15 JIKSDIO-Uaveform 5 PEAKS F(Hz>  JIKSDIO-Data Processed w— n  A  rs.B yB.B 9-5.B  B 1444 1445144G1 4471 4411 458 1 451 1452 1 453 1 4 Tine (Hi >  -Hi -TO  -72.90 -78.80  18174.0 21963.0  -78.80 -83.40  HARKER AT F(Hz> Z.i  4.4 B.B B.B l l . B 1 3 . Z 1 5 . 4 n . 6 U . B Z a rgggugngj (kHz)  H<dfi> -53.80  1722.0 8890.0 11972.0  H(dB)  113  APPENDIX E INFORMED CONSENT F O R M  116  APPENDIX F  AUDITEC  of St. Louis  CID AUDITORY TEST list 1A 1. an 2. 3. 4. 5. 6. 7.  yard carve  us day toe felt  8. stove 9. hunt 10. ran 11. knees 12. not 13. m e w 14. low 15. owl 16. it 17. she 18. high 19. there 20. earn 2 1 . twins 22. could 2 3 . what 24. bathe 2 5 . aco 2 6 . you 2 7 . as 2 8 . wet 29. c h e w 3 0 . see 3 1 . deaf 3 2 . them 3 3 . give 3 4 . true 3 5 . isle 3 6 . or 3 7 . law 3 8 . me 3 9 . none 4 0 . jam 4 1 . poor 4 2 . him 4 3 . skin 4 4 . east 4 5 . thing 4 6 . dad 4 7 . up 4 8 . bells 4 9 . wire 50. ache  List 2 A 1. yore 2. bin 3 . way 4. chest 5. then 6. ease 7. smart 8. gave 9. pew 10. ice 1 1 . odd 12. knee 13. move 14. now 15. jaw 16. one 17. hit 18. send 19. else 20. tear 2 1 . does 2 2 . too 2 3 . cap 24. with 25. air 26. and 27. young 28. cars 29. tree 30. dumb 3 1 . that 3 2 . die 3 3 . show 34. hurt 3 5 . own 36. key 37. oak 3 8 . new 39. live 40. off 42. ill 4 2 . rooms 4 3 . ham 44. star 4 5 . eat 46. thin 47. flat 48. well 49. by 50. ail  W-22 List 3 A 1. bill 2. add 3. west 4. cute 5. start 6. ears 7. tan 8. nest 9. say 10. if 1 1 . out 12. lie 13. three 14. oil 1 5 . king 16. pie 17. he 18. smooth 19. farm 20. this 2 1 . done 2 2 . use 2 3 . camp 2 4 . wool 2 5 . are 26. aim 27. when 28. book 29. tie 30. do 3 1 . hand 3 2 . and 3 3 . shove 34. have 3 5 . owes 3 6 . jar 37. no 3 8 . may 39. knit 40. on 4 1 . is 4 2 . raw 4 3 . glove 44. ten 4 5 . dull 4 6 . though 47. chair 48. we 49. ate 50. year  List 4 A 1. all 2. w o o d 3. at 4. where 5. chin 6. they 7. dolls 8. so 9. nuts 10. ought 11. in 12. net 1 3 . my 14. leave 15. of 16. hang 17. save 18. ear 19. tea 20. cook 2 1 . tin 2 2 . bread 23. why 2 4 . arm 25. yot 26. darn 27. art 28. will 29. dust 30. toy 3 1 . aid 3 2 . than 3 3 . eyes 3 4 . shoe 3 5 . his 3 6 . our 3 7 . men 3 8 . near 39. few 40. jump 4 1 . pail 4 2 . go 4 3 . stiff 44. can 4 5 . through 4 6 . clothes 47. who 48. bee 49. yes 50. am  APPENDIX G  Loudness Levels Painfully loud Extremely uncomfortable Uncomfortably loud Loud - annoying but O.K. Comfortable, but slightly loud Comfortable Comfortable, but slightly soft Soft Very soft  APPENDIX H Raw Results This appendix contains individual participant results as well as grouped results. The first four tables contain individual data, sorted by group. This is followed by a table containing results averaged across each group and averaged across all of the participants of the main study. Following these tables are graphed PI-PB functions of the same data, i.e., individual, grouped and total participant PI-PB functions.  INDIVIDUAL PARTICIPANTS' RAW DATA Subject: SS Level (dB HL) 40 55 65 100 98 98 76 84 76 80 80 84  G1S1  G1S2  NU6intact NU6jitter . W22jitter  Subject: HS Level (dB HL) 40 55 98 96 84 80 74 82  NU6intact NU6jitter W22jitter  Subject: LC Level (dB HL) 40 55 96 100 72 78 76 76  NU6intact NU6jitter W22jitter  Subject:GH Level (dB HL) 40 55 100 98 84 66 76 66  NU6intact NU6jitter W22jitter  65 100 94 80  75 100 78 74  95 96 76 56  PBmax 100 84 84  Pbmin 100 76 74  Rollover (%) 0 8 10  Rx 0.00 0.10 0.12  PBmax 100 94 82  Pbmin 96 76 56  Rollover (%) 4 18 26  Rx 0.04 0.19 0.32  PBmax 100 80 84  Pbmin 98 76 70  Rollover (%) 2 4 14  Rx 0.02 0.05 0.17  PBmax 100 84 78  Pbmin 90 52 60  Rollover (%) 10 32 18  Rx 0.10 0.38 0.23  G1S3 65 100 80 84  80 98 76 70  G1S4 65 98 78 78  100 90 52 60  INDIVIDUAL PARTICIPANTS' RAW DATA G2S1  NU6intact NU6jitter W22jitter  Subject: B B Level (dB HL) 40 55 65 100 98 100 78 78 82 80 80 . 70  G2S2  NU6intact NU6jitter W22jitter  Subject: MJ Level (dB HL) 40 55 65 100 98 96 80 84 84 78 74 80  G2S3  NU6intact NU6jitter W22jitter  Subject: CT Level (dB HL) 40 55 65 96 98 100 74 76 54 72 84 70  G2S4  NU6intact NU6jitter W22jitter  Subject: ES Level (dB HL) 40 55 65 100 98 100 82 86 82 78 86 90  80 100 70 64  90 92 76 78  75 96 66 62  80 94 72 82  PBmax 100 82 80  Pbmin 100 70 64  Rollover (%) 0 12 16  Rx 0.00 0.15 0.20  PBmax 100. 84 80  Pbmin 92 76 78  Rollover (%) 8 8 2  Rx 0.08 0.10 0.03  PBmax 100 76 84  Pbmin 96 54 62  Rollover (%) 4 22 22  Rx 0.04 0.29 0.26  PBmax 100 86 90  Pbmin 94 72 82  Rollover (%) 6 14 8  Rx 0.06 0.16 0.09  INDIVIDUAL PARTICIPANTS' RAW DATA  NU6intact NU6jitter W22jitter  Subject: EF Level.(dB HL), 40 55 65 98 100 96 84 68 76 84 86 76  G3S1 PBmax 100 84 86  85 100 68 66  Pbmin 100 68 66  Rollover (%) 0 16 20  Rx 0.00 0.19 0.23  Pbmin 98 64 74  Rollover (%) 2 12 6  Rx 0.02 0.16 0.08  PBmax 100 , 78 86  Pbmin 98 64 62  Rollover (%) 2 14 24  Rx 0.02 0.18 0.28  PBmax 100. 88 90 .  Pbmin 98 64 66  Rollover (%) 2 24 24  Rx 0.02 0.27 0.27  • •  NU6intact NU6jitter W22jitter  Subject: MS Level (dB HL) 40 55 65 100 100 98 70 74 76 72 80 78  G3S2 PBmax 100 76 80  80 98 64 74  •  G3S3  NU6intact NU6jitter W22jitter  Subject: MB Level (dB HL) 40 55 65 98 100 100 64 78 64 80 86 76  G3S4  NU6intact NU6jitter W22jitter  Subject: LR Level (dB HL) 40 55 65 100 100 98 76 88 78 80 90 86  90 98 72 62  85 98 64 66  '  INDIVIDUAL PARTICIPANTS' RAW DATA G4S1  NU6intact NU6jitter W22jitter  Subject: C G Level (dB HL) 40 55 65 100 100 100 66 68 68 78 74 72  G4S2  NU6intact NU6jitter W22jitter  Subject:AB Level (dB HL) 40 55 65 94 92 98 70 72 72 72 70 68  G4S3  NU6intact NU6jitter W22jitter  Subject: SI Level (dB HL) 40 55 65 98 100 96 82 84 76 82 82 88  G4S4  NU6intact NU6jitter W22jitter  Subject: G C Level (dB HL) 40 55 65 98 100 100 84 74 78 74 80 84  95 94 40 62  90 86 48 52  90 92 64 60  80 100 74 78  PBmax 100 68 78  Pbmin 94 40 62  Rollover (%) 6 28 16  Rx 0.06 0.41 0.21  PBmax 98 72 72  Pbmin 86 48 52  Rollover (%) 12 24 20  Rx 0.12 0.33 0.28  PBmax 100 84 88  Pbmin 92 64 64  Rollover (%) 8 20 24  Rx 0.08 0.24 0.27  PBmax 100 84 84  Pbmin 100 74 78  Rollover (%) 0 10 6  Rx 0.00 0.12 0.07  GROUPED RAW DATA GROUP 1 Level 1 NU6intact NU6jitter W22jitter  Level 2 Level 3 Level 4 Rollover(%) 98 98.5 99 96 3 74.5 81.5 82 70.5 . 11.5 74 78.5 81.5 65 16.5  GROUP 2 Level 1 NU6intact NU6jitter W22jitter  Level 2  99 78.5 77  Level 3 98 81 81  99 75.5 77.5  Level 4 Rollover(%) 95.5 3.5 71 10 71.5 9.5  GROUP 3 Level 1 NU6intact NU6jitter W22jitter  99 69.5 79  Level 2 Level 3 Level 4 Rollover(%) 100 98 98.5 2 81.5 73 67 14.5 85.5 79 67 18.5  GROUP 4  NU6intact NU6jitter W22jitter  Level 1 Level 2 Level 3 Level 4 Rollover(%) 97 97.5 99.5 93 2 74 75 74.5 56.5 18.5 76 77 78 63 15  A L L SUBJECTS  NU6intact NU6jitter W22jitter  Rollover(%) Level 1 Level 2 Level 3 Level 4 98.38 99.00 98.25 95.75 3.25 74.13 79.75 76.25 66.25 13.50 76.50 80.50 79.00 66.63 13.88  124  LO  o o  on o c oo o  o o o CD LO ^t  (%) ajoos IdAA  125  (o/ ) OJOOS ± U M 0  LO  (o/ ) ajoos ±JdM 0  133  o o  o  CD  o oo  o  o CD  (o/„) ajoos ± U M  o LO  o ^j-  («•/„) ajoos ±y/w  LO  o O  o  O  o  o  00  Is  o  CO  (o/ ) OJOOS ± U M 0  o  LO  o  (o ) ajoos ±UM /o  (o/ ) ejoos l ^ M 0  145 APPENDIX I Individual Subject Characteristics Subject  Sex  Age  Years o f  (Years)  Education  PTA(dBHL) Left  Right  SRT (dBHL) Left  G1S1  M  29  20  -1. 7  0  -5  G1S2  F  29  19  3.3  1. 7  0  G1S3  F  29  16  -1. 7  -5  0  G1S4  M  34  25  0  -1. 7  0  G2S1  F  33  18  0  3.3  0  G2S2  F  26  20  0  1.7  0  G2S3  F  27  17  -5  -5  -5  G2S4  F  27  18  -8.3  -5  -5  G3S1  F  23  18  3.3  3.3  0  G3S2  F  30  21  -5  -1.7  0  G3S3  F  22  16  -1.7  -1. 7  -5  G3S4  F  28  18  -1.7  1.7  0  G4S1  M  25  20  5  3.3  5  G4S2  F  22  16  0  1. 7  5  G4S3  M  25  20  -1.7  -1. 7  0  G4S4  M  28  18  3.3  5  0  146 Indivdual Subject Characteristics (continued) Subject  MCL  UCL  Left  Left ,  I p s i l a t e r a l Acoustic Reflex 500 Hz  Threshold(dBHL) 1 kHz  Handedness  2 kHz  G1S1  45  80  90  90  90  R  G1S2  40  100  80  . 85  90  L  G1S3  40  85  . 90  85  85  R  G1S4 •  40  105  100  100  95  R  G2S1  45  85  85  80  85  R  G2S2  50  95  90  95  85  L  G2S3  35  80  95  95  95  L  G2S4  40  85  90  95  90  R  G3S1  50  90  95  95  90  R  G3S2  45  85  95  95  100  R  G3S3  45  95  CNT  CNT  CNT  R  G3S4  45  90  90  90  90  R  G4S1  45  80  80  85  R  G4S2  45  95  90 .  85  90  R  G4S3  40  95  80  75  80  R  G4S4  50  85  90  85  90  R  100 ,  APPENDIX J Participant Responses for Most Common NU-6 Word Errors E r r o r s Made i n NU6intact C o n d i t i o n # of E r r o r s  I n c o r r e c t responses  Pole  9  Pull(9)  Youth  7  Mill  6  Nill(6)  Germ  5  Jarm(3), Jam(2)  Lore  5  Lure(3),  Yearn  5  Yarn(5)  Burn  3  Barn(3)  Laud  2  Log(l),  Room  2  Roam(2)  Goal  2  Gaul(1),  Target  Word  '  Use (6), Youths(1)  E r r o r s Made i n N U 6 j i t t e r Target Mill  Word  Lower(2)  Lodge(1)  Gull(1)  Condition  # of E r r o r s  Incorrect  Responses  16  Null(6), Mull(5), N o ( l ) , Mow(l),  Nail(l),  Low(l),  Mall(1) Numb  16  None(9), Men (4),  Mum(l),  When(l) , N i n ( l ) Thumb  16  Thin(5),  Fun(3), From(3),  F r i e n d ( 2 ) , Phone(1), Bum(l) Fen(l)  Pole  15  Pull(13),  Pearl(1),  Laud  14  Loud(7), Love(2) , Led(2), Lard(l), Lot(l),  Limb  Lung(7),  14  Ball(l)  Lob(l)  Ling(2),  Lend(1), Lynn(1),  Long(l), Lun(1),  Loon(l) Met  14  '  M i t t (8), K n i t ( 3 ) ,  Net(2),  Lit(l) Calm  14  Cone(4),  C a r ( 2 ) , Kern(2),  Come(1), Karn(1), Comb(1)  Tell  14  Cram(1), Curb(1),  Curve(1)  Tail(ll),  Turn(l),  Tag(l),  Tongue(1), Mob  14  Knob(7), Nab(3), N o t ( l ) , Name(l), N o ( l ) ,  Mop(l)  Peg  14  Pig(14)  Pearl  13  G i r l ( 2 ) , G o a l ( 2 ) , B a l l (2), Pall(l),  Curl(l),  Pail(l),  Crawl(1),  Gaul(l), All(l),  NR(1) Goal  13  Gaul(5), B u l l ( 2 ) , G u l l ( 2 ) , B a l l ( 2 ) , Gone(l),  * Note: NR  = No Response  Gong(l)  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089575/manifest

Comment

Related Items