Temporal Processing in Low-Frequency Channels: Effects of Age and Hearing Loss in Middle-Aged Listeners DOI: /jaaa

Transcription

1 J Am Acad Audiol 22: (2011) Temporal Processing in Low-Frequency Channels: Effects of Age and Hearing Loss in Middle-Aged Listeners DOI: /jaaa Elizabeth D. Leigh-Paffenroth* Saravanan Elangovan* Abstract Background: Hearing loss and age interfere with the auditory system s ability to process temporal changes in the acoustic signal. A key unresolved question is whether high-frequency sensorineural hearing loss (HFSNHL) affects temporal processing in the low-frequency region where hearing loss is minimal or nonexistent. A second unresolved question is whether changes in hearing occur in middle-aged subjects in the absence of HFSNHL. Purpose: The purpose of this study was twofold: (1) to examine the influence of HFSNHL and aging on the auditory temporal processing abilities of low-frequency auditory channels with normal hearing sensitivity and (2) to examine the relations among gap detection measures, self-assessment reports of understanding speech, and functional measures of speech perception in middle-aged individuals with and without HFSNHL. Research Design: The subject groups were matched for either age (middle age) or pure-tone sensitivity (with or without hearing loss) to study the effects of age and HFSNHL on behavioral and functional measures of temporal processing and word recognition performance. These effects were analyzed by individual repeated-measures analyses of variance. Post hoc analyses were performed for each significant main effect and interaction. The relationships among the measures were analyzed with Pearson correlations. Study Sample: Eleven normal-hearing young adults (YNH), eight normal-hearing middle-aged adults (MANH), and nine middle-aged adults with HFSNHL were recruited for this study. Normal hearing sensitivity was defined as pure-tone thresholds #25 db HL for octave frequencies from 250 to 8000 Hz. HFSNHL was defined as pure-tone thresholds #25 db HL from 250 to 2000 Hz and $35 db HL from 3000 to 8000 Hz. Data Collection and Analysis: Gap detection thresholds (GDTs) were measured under within-channel and between-channel conditions with the stimulus spectrum limited to regions of normal hearing sensitivity for the HFSNHL group (i.e.,,2000 Hz). Self-perceived hearing problems were measured by a questionnaire (Abbreviated Profile of Hearing Aid Benefit), and word recognition performance was assessed under four conditions: quiet and babble, with and without low-pass filtering (cutoff frequency Hz). Results: The effects of HFSNHL and age were found for gap detection, self-perceived hearing problems, and word recognition in noise. The presence of HFSNHL significantly increased GDTs for stimuli presented in regions of normal pure-tone sensitivity. In addition, middle-aged subjects with normal hearing sensitivity reported significantly more problems hearing in background noise than the young normal-hearing subjects. Significant relationships between self-report measures of hearing ability in background noise and word recognition in babble were found. *James H. Quillen VA Medical Center, Mountain Home, TN; Department of Audiology and Speech Pathology, East Tennessee State University, Johnson City Elizabeth Leigh-Paffenroth, Audiology 126, VA Medical Center, Mountain Home, TN 37684; Phone: , ext. 7138; Fax: ; elizabeth.leigh@med.va.gov This material is based on work supported by a Research Career Development award (C4323V) funded by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Rehabilitation Research and Development (RR&D) Service, Washington, D.C., to the first author and by the RR&D Auditory and Vestibular Dysfunction Research Enhancement Award Program. Disclaimer: The contents do not represent the views of the Department of Veterans Affairs or the U.S. government. Portions of this article were presented at the American Auditory Society Annual Meeting, March 2010, Scottsdale, AZ. 393

2 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 Conclusions: The conclusions from the present study are twofold: (1) HFSNHL may have an off-channel impact on auditory temporal processing, and (2) presenescent changes in the auditory system of MANH subjects increased self-perceived problems hearing in background noise and decreased functional performance in background noise compared with YNH subjects. Key Words: gap detection, hearing loss, normal hearing sensitivity, speech perception, temporal processing Abbreviations: AC 5 auditory cortex; APHAB 5 Abbreviated Profile of Hearing Aid Benefit; AV 5 aversion; BC 5 between channel; BN 5 background noise; EC 5 ease of communication; F 0 5 fundamental frequency; F 1 5 first formant frequency; F 2 5 second formant frequency; GDT 5 gap detection threshold; HFSNHL 5 high-frequency sensorineural hearing loss; HINT 5 Hearing in Noise Test; IC 5 inferior colliculus; ILD 5 interaural level difference; ITD 5 interaural time difference; MAHL 5 middle aged, hearing loss; MANH 5 middle aged, normal hearing sensitivity; RV 5 reverberation; WC 5 within channel; YNH 5 young, normal hearing sensitivity Auditory temporal processing refers to the ability to precisely encode complex and dynamic temporal features inherent in human speech and other acoustical stimuli. Temporal processing is affected by advanced age and by hearing loss. The individual contributions of these effects to performance on temporal processing tasks are often difficult to determine given that most older subjects have some degree of highfrequency sensorineural hearing loss (HFSNHL). One way to control the confounding effects of age is to assess temporal processing abilities for low-frequency regions with normal auditory sensitivity (i.e., pure-tone thresholds are within normal limits for the range of frequencies in the test stimulus). Recent evidence, however, has shown significantly poorer gap detection for stimuli presented in regions of normal pure-tone sensitivity (Yin et al, 2008; Feng et al, 2009). Yin et al recorded electrophysiologic gap detection thresholds (GDTs) in guinea pigs before and after noise-induced HFSNHL. The stimulus spectrum for the gap detection measures was restricted to low-frequency regions where hearing sensitivity apparently was not affected by the noise. The results showed a significant reduction in GDTs after noise-induced HFSNHL compared with when the animals had normal hearing sensitivity. Further, human data revealed that subjects with HFSNHL performed poorer than subjects with normal hearing sensitivity (Feng et al, 2009) on several measures of low-frequency temporal processing (e.g., amplitude modulation detection, within-channel gap detection, and time-compressed speech perception). Thus it appears that HFSNHL may have an off-channel impact in the central auditory system on temporal processing in the low-frequency channels. In other words, these studies indicate that a hearing loss restricted to the higher frequencies may still affect the suprathreshold auditory processing of lowfrequency auditory channels where pure-tone thresholds are within normal limits. An off-channel impact occurs when high-frequency channels are involved in the processing of low-frequency signals (Florentine et al, 1993; Yin et al, 2008). The importance of understanding temporal processing is in its close relation to speech understanding in difficult listening environments (Tremblay et al, 2004; Healy and Bacon, 2007). Word recognition measured in the presence of background noise is a commonly used test measure to yield ecologically valid and clinically relevant information about auditory processing mechanisms in challenging listening environments. Age and the presence of hearing loss appear to interfere with the ability to perceive speech, especially in the presence of background noise (Hirsh, 1950; Carhart and Tillman, 1970; Olsen et al, 1975; Dubno et al, 1984; Gordon-Salant, 1987; Beattie, 1989; Souza and Turner, 1994; Divenyi and Haupt, 1997). Although a number of investigators have shown that older (i.e.,.60 yr of age) listeners have considerable difficulty understanding speech in the presence of competing signals, only recently have investigators (Helfer and Vargo, 2009) shown that middle-aged (45 54 yr) listeners with normal pure-tone sensitivity experience difficulties understanding speech in complex listening environments. An increase in pure-tone thresholds within the range of clinically defined normal hearing sensitivity may be a confounding variable; however, decrements in speech perception regardless of pure-tone sensitivity deserve further investigation. The onset of the decline in speech understanding abilities is not clearly understood, especially in the absence of pure-tone hearing loss. Previous studies have shown that the ability to resolve static and dynamic temporal cues in speech and nonspeech stimuli is compromised in individuals with hearing loss and with increasing age (Strouse et al, 1998; Snell and Frisina, 2000; Lister and Tarver, 2004). Only a handful of investigators (Abel et al, 1990; Snell and Frisina, 2000; Grose et al, 2006) have examined the temporal processing abilities of middle-aged listeners (i.e., yr old). Recent investigations have shown that auditory temporal processing deficits start emerging around middle age (Grose et al, 2006; Grose and Mamo, 2010) and continue to deteriorate with increasing age (Ross et al, 2007). Grose and colleagues 394

3 Temporal Processing/Leigh-Paffenroth and Elangovan have shown changes in temporal processing abilities in middle-aged subjects for gap-duration discrimination tasks (Grose et al, 2006) and interaural phase differences (Grose and Mamo, 2010) for subjects aged yr. In contrast, Dubno et al (1997) found no significant differences in performance for phonetically balanced word lists, Synthetic Sentence Identification (Speaks and Jerger, 1965), or the Speech Perception in Noise test (Bilger, 1984) when comparing middle-aged subjects (55 64 yr) to two groups of older-aged subjects (65 74 and yr). Thus, it is evident that further investigation is required to better understand the temporal processing deficits associated with middle age. One way to investigate the effects of age on measures of temporal processing is to compare temporal processing data from young normal-hearing subjects and middleaged normal-hearing subjects. Significant differences between the normal-hearing groups would indicate effects of age. The effect of HFSNHL can be assessed by comparing temporal processing data from middle-aged normal-hearing subjects and middle-aged subjects with HFSNHL. Significant differences between the middleaged groups would indicate effects of HFSNHL. The investigation of temporal processing in middleaged subjects with and without HFSNHL is important for three reasons: (1) recent work in guinea pigs and humans has shown effects of HFSNHL on low-frequency temporal processing where no pure-tone hearing loss existed (Yin et al, 2008); (2) the evidence of auditory temporal processing problems in older listeners is confounded by the ubiquitous presence of some degree of HFSNHL, which prohibits clear interpretation of results (Gordon-Salant and Fitzgibbons, 1993; Fitzgibbons and Gordon-Salant, 1996; Snell and Frisina, 2000; Humes et al, 2010); and (3) changes in temporal processing abilities have been found in middle-aged subjects (Grose and Mamo, 2010). The results from this study improve our understanding of the confounding influence of hearing loss on aging in the auditory system. The purpose of the study was twofold: (1) to examine the influence of HFSNHL and aging on the auditory temporal processing abilities of low-frequency auditory channels with normal hearing sensitivity and (2) to examine the relations among gap detection measures, self-assessment reports of understanding speech, and functional measures of speech perception in middle-aged individuals with and without HFSNHL. middle-aged adults with normal hearing (MANH), and middle-aged adults with HFSNHL (MAHL) were recruited. The YNH participants (N 5 11; from 19 to 37 yr, mean age 27 yr) and the MANH participants (N 5 8; from 46 to 59 yr, mean age 52 yr) had air- and boneconduction pure-tone thresholds #25 db HL for octave frequencies from 250 to 8000 Hz. The MAHL participants (N 5 9; from 51 to 62 yr, mean age 59 yr) had air- and bone-conduction pure-tone thresholds #25 db HL from 250 to 2000 Hz and $35 db HL from 3000 to 8000 Hz (Fig. 1). The participants with abnormal tympanograms, conductive hearing loss, air bone gaps.10 db, history or presence of otologic pathology, history of possible retrocochlear pathology, history of stroke, seizure disorder, or dementia were excluded. Self-Assessment of Hearing Problems The participants were administered the Abbreviated Profile of Hearing Aid Benefit (APHAB [Cox and Alexander, 1995]) to determine the frequency of hearing problems experienced in different listening situations. The APHAB was developed as a self-assessment inventory of the benefit a listener experienced when wearing hearing aids compared with not wearing hearing aids. The inventory consists of 24 items distributed across four subscales: ease of communication (EC), reverberation (RV), background noise (BN), and aversion (AV). Each item in the test is answered without my hearing aids and with my hearing aids. The without my hearing aids portion of the test was administered to each listener in the present study as a self-assessment tool for hearing on three of the four subscales (EC, RV, MATERIALS AND METHODS Subjects The present study was approved by the local Institutional Review Board, and all participants signed an informed consent document prior to their participation in the study. Young adults with normal hearing (YNH), Figure 1. Mean audiometric pure-tone thresholds (db HL) for participants: young, normal hearing (circles); middle aged, normal hearing (squares); and middle aged, high-frequency sensorineural hearing loss (triangles). Error bars represent 6SE of the mean. 395

4 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 and BN). Scores were calculated by the APHAB software as percentage of problems experienced by subjects for each subscale. The AV subscale is a measure primarily for hearing aid users and was not used in the data analysis for the current study. Word Recognition Word recognition was measured using four 50-item lists from the recordings of the Northwestern University Auditory Test No. 6 made by a female speaker (Department of Veterans Affairs, 1998) and from the same recorded pool of words used in the Words in Noise protocol (Wilson, 2003). The word lists were presented in four conditions: quiet and babble, with and without low-pass filtering. The babble consisted of six talkers (three female and three male) reading passages that when combined are unintelligible to the listener (Sperry et al, 1997; Wilson, 2003). The word lists (lists 1A, 2A, 3A, and 4A) were counterbalanced across the four conditions (quiet, quiet low pass, babble, and babble low pass) and presented at 70 db SPL, and the babble was presented at 62 db SPL for a fixed signal-to-noise ratio of 8 db. The materials were reproduced by compact disc (Sony, Model CDP-CE375) and fed through an audiometer (Grason Stadler, Model 61) and low-pass filter (Wavetek, Model 753A) to an Etymotic Research ER-3A insert earphone. 1 The filter cutoff was at 2000 Hz, with a 115 db/octave skirt. For the unfiltered conditions the filter was turned off. Testing was conducted in a double-walled sound booth with the verbal responses of the listener scored by the examiner. Gap Detection Measurement of gap detection is one psychophysical task often used to evaluate auditory temporal processing. In a typical gap detection paradigm, a sound sequence consisting of a leading marker, a silent gap, and a trailing marker is presented. The gap detection threshold is the shortest silent gap between the two markers that a listener can detect. Depending on whether the temporal and spectral properties of the leading and trailing markers are similar (within-channel gap detection) or dissimilar (between-channel gap detection), this paradigm is likely to tap different temporal processing mechanisms (Formby et al, 1998). In many ways, between-channel listening resembles complex auditory tasks such as speech perception and results in higher GDTs compared with within-channel designs (e.g., Phillips and Hall, 2002). Stimuli The gap detection stimulus was modeled after the first two formants of the bilabial /ba/ (i.e., F Hz and F Hz) and similar to the nonspeech stimuli used by Horev et al (2007). The source filter synthesis function of the Praat software (Boersma and Weenink, 2008) was used to synthesize the fundamental frequency (F 0 ) of the source (120 Hz) and the first two formants (F 1 and F 2, with bandwidths of 60 Hz and 110 Hz, respectively). The stimulus spectra of all stimuli were sampled at 44,100 Hz and low-pass filtered (Hanning window) with a cutoff frequency of 2000 Hz with a 48 db/octave skirt. The experimental was controlled by SykofizX (version 2.0, Tucker-Davis Technologies System 3) software, and the stimuli were delivered through ER-3A insert earphones. The responses were recorded via a response pad integrated with the SykofizX software. The gap markers for the within-channel (WC) and between-channel (BC) gap conditions were synthesized from the basic two-formant stimulus and are illustrated in Figure 2. The leading and trailing markers for the WC gap condition were acoustically similar and were z150 msec in duration with a 3 msec rise/fall time. The markers contained F 0,F 1, and F 2 as described above. The leading and trailing markers for the BC condition differed in both spectrum and duration as follows: (1) the leading marker was z20 msec in duration and contained F 0 and F 2, and (2) the trailing marker duration was z280 msec and contained only F 0 and F 1.Eachstimulus trial consisted of three stimulus sequences (i.e., two control sequences and one test sequence), with an intersequence interval of 500 msec. The leading marker and trailing marker of the control sequences were separated by an inaudible gap of 1 msec. The leading and trailing markers for the test sequence were separated by a variable gap. In addition, a continuous wideband Gaussian background noise was presented 20 db below the peak level of the stimulus to reduce the possibility of perceived spectral splatter at the onset or offset of the gap. The stimulus level was calibrated daily with a sound level meter (Brüel and Kjær, Type 2250) using an ER3-A insert earphone and a 2 cm 3 coupler (Brüel and Kjær, Type DB-0138) attached to an artificial ear (Brüel and Kjær, Type 4152). The low-pass filter cutoff was confirmed by spectral analysis. Procedures Behavioral gap detection thresholds were obtained in a three-interval forced-choice adaptive procedure converging on the 70.7% correct point on the psychometric function (Levitt, 1971). The test trials continued until there were 14 reversals or 60 total trials, and the GDT was stable in the desirable direction (i.e., lower than the gap duration set at the onset of the task) for the test stimulus. The gap threshold for the block was the arithmetic mean of the duration of the gap for the last six reversals. The GDTs were specified as the duration of the silent interval exclusive of rise/fall time 396

5 Temporal Processing/Leigh-Paffenroth and Elangovan Figure 2. Stimulus spectrum for the two-formant gap detection paradigm showing no spectral energy above 2000 Hz. (Phillips and Hall, 2002). Behavioral GDTs were recorded under two presentation levels (i.e., 55 and 80 db SPL) for both gap conditions (i.e., WC55, WC80, BC55, and BC80). Presentation levels were chosen to reflect gap detection performance at a low level (55 db SPL) and at a level consistent with conversational speech (80 db SPL). The experimental procedures for each subject were conducted in one 2 hr session. All stimuli were presented monaurally to the right ear (N 5 26)ortothe left ear in cases where audiometric exclusion criteria were identified for the right ear (N 5 3). RESULTS Frequency of Hearing Problems Reported for Unaided APHAB The frequency of problems reported in different listening situations is represented by group in Figure 3. The data for the three subject groups are shown for three subscales of the APHAB: listening in quiet, in reverberant environments, and in background noise. Overall, the most problems hearing were reported for the MAHL group, with a range of 37 56%, followed by the MANH group (z11 28%) and the YNH group (z4 8%). In addition, more problems were reported in the BN environment than in reverberant or quiet listening environments. In the BN environment, the MANH group reported more problems than the YNH group (see Table 1). A two-factor, mixed, repeated-measures analysis of variance (ANOVA) was performed to determine the effects of condition (EC, RV, and BN) and group (YNH, MANH, and MAHL) on the percentage of problems reported on the APHAB. The main effects of environment (F [2, 52] , p,.001) and group (F [2, 26] , p,.001) were significant. For condition, post hoc comparisons showed that all differences among the three subscales were significant. The two-way interaction (condition 3 group) was significant (F [4, 52] 5 3.4, p 5.015). Post hoc comparisons revealed the following significant differences: (1) the MAHL subjects reported significantly more problems in each listening environment than YNH and MANH subjects (p,.001), (2) the MANH subjects reported significantly more problems in background noise than the YNH subjects (p 5.023), (3) the percentage of problems reported by the YNH subjects was the same regardless of the listening condition ( p..05 for ear comparison), (4) the MANH subjects reported more problems in background noise than reverberant (p 5.007) or quiet environments (p 5.010), and (5) the MAHL group reported significantly more problems in reverberant environments than in quiet (p 5.001) and significantly more problems in background noise than in reverberant environments (p 5.046). In summary, the data reported here show an effect of hearing loss for the quiet, reverberant, and background noise listening conditions and an additional effect of age for the BN listening condition. Word Recognition Average word recognition performance for each stimulus condition and each subject group is shown in Figure 4. The bars represent average percent correct performance for each word recognition condition (quiet, filtered quiet, babble, and filtered babble) and for each subject group (YNH, MANH, and MAHL). Overall, 397

6 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 Figure 3. Mean frequency of problems reported for the Abbreviated Profile of Hearing Aid Benefit (APHAB) questionnaire under the ease of communication (EC), reverberation (RV), and background noise (BN) categories as a function of group. Error bars represent 6SD of the mean. word recognition performance was best for words presented in quiet and poorest for words presented in low-pass-filtered babble. The subjects with normal hearing sensitivity performed better than the subjects with hearing loss, but all of the subject groups showed a decline in performance when babble or filtering was added to the word recognition task, as seen in Table 1. The word recognition data were skewed positively (i. e., toward 100% correct) and required a rationalized arcsine transformation to be performed prior to data analysis (see Studebaker, 1985). A two-factor, mixed, repeated-measures ANOVA was performed to determine the effects of word recognition condition (quiet, filtered quiet, babble, and filtered babble) on subject group (YNH, MANH, and MAHL). The main effects of word recognition condition (F [3, 72] , p,.001) and subject group (F [2, 24] , p 5.001) were significant. For word recognition condition, post hoc comparisons revealed a significant decline in word recognition performance as task difficulty increased with the addition of low-pass filtering or the presence of babble (see Figure 4). The two-way interaction (condition 3 group) was significant (F [6, 72] , p 5.011). Post hoc comparisons were performed, and the following pertinent findings were noted: (1) the performance of the MAHL group was significantly poorer in the quiet condition than the performance of the YNH group (p,.001) and the MANH group (p 5.010); (2) the performance of the MAHL group was significantly poorer in babble than that of the YNH group (p,.001) but not the MANH group (p 5.056); (3) no significant differences among the groups were found for the quiet filtered and the babble filtered conditions (p..05); (4) within the MANH group performance was significantly different between each pair of word recognition conditions (e.g., quiet vs. quiet filtered, quiet filtered vs. babble, etc.); and (5) across conditions, the MANH group performed significantly poorer than the YNH group (p 5.035) but no different from the MAHL group ( p 5.057). In summary, the data reported here show effects of hearing loss and age for words presented in quiet and babble without low-pass filtering. Gap Detection Average GDTs in different conditions are represented by group in Figure 5. The bars represent GDT as a function of condition (WC and BC), presentation level (55 and 80 db SPL), and subject group (YNH, Table 1. Means and Standard Errors for the Subscales of the Abbreviated Profile of Hearing Aid Benefit, Word Recognition Conditions, and Gap Detection Thresholds Across Condition and Level for Each Subject Group Young, Normal Hearing Middle Aged, Normal Hearing Middle Aged with High-Frequency Sensorineural Hearing Loss Variable M SE M SE M SE Abbreviated Profile of Hearing Aid Benefit Subscale Ease of Communication Reverberation Background Noise Word Recognition Quiet Quiet Low Pass Babble Babble Low Pass Gap Detection Threshold Within Channel, 55 db SPL Within Channel, 80 db SPL Between Channel, 55 db SPL Between Channel, 80 db SPL

7 Temporal Processing/Leigh-Paffenroth and Elangovan Figure 4. Mean word recognition scores recorded under the different listening conditions as a function of group. Error bars represent 6SD of the mean. MANH, and MAHL). Overall, as expected, the WC GDTs were better (lower) than the BC GDTs for both presentation levels. Further, the performance of the MANH group varied depending on the gap condition (i.e., WC or BC). For the WC conditions, the MANH participants performed similar to the YNH group, but for the BC conditions, the average performance of the MANH group was in between those of the YNH and MAHL groups (Table 1). As can be seen in Figure 5 and Table 1, variability within each group was higher for the BC GDTs than for the WC GDTs. A three-factor, mixed, repeated-measures ANOVA was performed to determine the effects of gap detection condition (WC and BC), presentation level (55 or 80 db SPL), and subject group (YNH, MANH, and MAHL) on GDTs. The main effects of gap detection condition (F [1, 29] , p,.001), level (F [1, 29] , p 5.002), and group (F [2, 29] , p 5.026) were significant. In other words, the gap thresholds estimated by the WC conditions were significantly shorter than those estimated by the BC conditions, and the GDTs obtained at 80 db SPL were significantly shorter than the GDTs obtained at 55 db SPL. The two-way interaction (condition 3 level) was significant (F [1, 29] , p 5.027), and post hoc comparisons revealed significantly shorter GDTs for the WC condition compared with the BC condition at 55 db SPL (p,.001) and at 80 db SPL (p 5.001). In addition, the GDTs obtained at 80 db SPL were significantly shorter than the GDTs obtained at 55 db SPL for the BC condition only ( p 5.003). The three-way interaction (condition 3 level 3 group) was not significant. The WC and BC gap conditions were analyzed separately in order to examine further the possible group differences for each type of gap detection task (Fig. 5). Two independent, two-factor, mixed, repeated-measures ANOVAs were employed to examine differences in gap Figure 5. Mean gap detection thresholds recorded under within- (i.e., WC) and between-channel (i.e., BC) conditions as a function of group and presentation level (55 and 80 db SPL). Error bars represent 6SD of the mean. detection performance as a function of presentation level and group for each gap condition. The two-way (group 3 level) repeated-measures ANOVA for the WC gap condition revealed a significant main effect for group (F [1, 29] , p 5.004). Post hoc comparisons revealed that the WC GDTs from the MAHL group were significantly higher (poorer) that the WC GDTs from the YNH group (p 5.003) and the MANH group (p 5.006). No significant difference in WC GDTs was found between the YNH and the MANH groups at either level. The interaction between group and level was not significant (p..05). The two-way (group 3 level) repeated-measures ANOVA for the BC gap condition revealed no main effect for group (F [1, 29] , p 5.074). In summary, this analysis revealed a significant effect of HFSNHL for the WC gap condition but no significant effects of age or hearing loss for the BC gap condition. The BC GDTs had large variability within each group. Self-Reported Problems and Word Recognition Performance The relationship between the frequency of problems reported (%) under reverberant environments and in background noise for the APHAB and individual word recognition performances (% correct) is shown in scatter plots in Figure 6 as a function of participant group. The data are presented from the YNH subjects in the left column, the MANH subjects in the middle column, and the MAHL subjects in the right column. 2 Word recognition in babble scores are presented in the first and second rows, and the word recognition in low-passfiltered babble scores are presented in the third and fourth rows. Self-reported frequency of problems for RV listening conditions is presented in the first and third rows, and the frequency of problems reported 399

8 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 Figure 6. Bivariate scatter plots and regression lines for the frequency of problems (%) reported under reverberant environments (RV) and in background noise (BN) for the Abbreviated Profile of Hearing Aid Benefit (APHAB) as a function of individual word recognition scores (% correct) for each condition and participant group. Regression lines were fit by the method of least squares. for the BN listening condition is presented in the second and fourth rows. Pearson correlation coefficients (r) and P-values are found in each scatter plot. In general, individuals who reported fewer problems in listening in reverberant and noisy environments had better word recognition performances. However, significant negative correlations were observed only for the following comparisons: RV versus babble and BN versus babble for the MAHL group, BN versus babble for the MANH group, and BN versus babble filtered for the YNH group. No significant correlations were found between GDTs and self-reported measures or word recognition performance in the current study. T he DISCUSSION stimuli for the gap detection and word recognition tasks (i.e., quiet filtered and babble filtered conditions) were designed to stimulate the audible lowfrequency auditory channels (i.e.,,2000 Hz) for the HFSNHL group. The overall results support a possible hierarchical influence of aging and high-frequency pure-tone thresholds on auditory temporal processing and speech recognition (both objective and perceived). In other words, the measures recorded from the MANH group were intermediate to those of the YNH and MAHL groups on average, but large variability was found for some measures (e.g., between-channel gap detection). Self-Report of Hearing Abilities The subscales of the APHAB reflect the frequency of problems reported for hearing in quiet, in reverberant environments, and in background noise. In the present study, the subjects with hearing loss (MAHL) reported more problems in all listening environments than the 400

9 Temporal Processing/Leigh-Paffenroth and Elangovan subjects with normal hearing sensitivity (YNH and MANH). However, the MANH subjects reported significantly more problems hearing in background noise (28%) than the YNH subjects (8%) but not in quiet. The APHAB scores for the MAHL subjects reported in the present study are similar to data reported from previous studies. The problem frequencies reported by MAHL subjects in the present study were 31%, 43%, and 52% for EC, RV, and BN subscales, respectively. The present data are comparable to those recorded from subjects with mild hearing loss aged yr who reported the frequency of listening problems to be z25% in quiet, z45% in reverberation, and z55% in background noise (Cox et al, 2003). As expected, the MAHL group reported fewer problems than an older group with more hearing loss reported in a recent study (Roup and Noe, 2009). Speech Perception Abilities in Middle-Aged Subjects One goal of the present investigation was to determine if a high-frequency hearing loss would affect the speech recognition ability of low-frequency auditory channels. Toward that end, low-pass filtering was used to restrict the stimuli to regions of normal hearing sensitivity for all subjects in order to simulate effects of HFSNHL. Further, since a number of studies have shown that age-related deficits in speech perception are critically revealed in demanding listening environments, performances were also assessed in the presence of multitalker babble. A close examination of Figure 4 reveals that the main differences between the groups emerged with the babble conditions (i.e., babble and babble filtered conditions). In the unfiltered babble condition, a group mean difference in word recognition scores was observed between the YNH group and the MAHL group (p,.001) and between the YNH group and the MANH group ( p 5.035). The difference between the MANH and MAHL group approached statistical significance (i.e., p 5.057) and may prove statistically significant in larger samples of MANH and MAHL subjects. In an earlier investigation, Horwitz et al (2002) measured speech perception performance in quiet and in speech-shaped noise in subjects with normal hearing sensitivity through 2000 Hz with the intention of better understanding the role of high-frequency auditory channels in low-frequency speech perception. Half the subjects had a HFSNHL (66 79 yr), and the other half had normal high-frequency hearing (21 35 yr). Their results revealed significant differences in speech perception in noise conditions between groups even when the speech was low-pass filtered and restricted to regions of normal pure-tone hearing. The authors concluded that these results support the hypothesis that high-frequency auditory channels code meaningful low-frequency speech cues and that peripheral hearing loss alone does not account for all the speech perception problems experienced by middle-aged subjects. Although Horwitz et al (2002) acknowledge the age differences between their control and test groups and that age could be a potential confounding variable, they primarily attributed the reduced speech performance to the impaired high-frequency auditory channels and highlighted the role of high-frequency auditory channels toward the perception of low-frequency cues. The word recognition findings from the present study concur with those reported by Helfer and Vargo (2009), who measured speech recognition for sentences presented in noise. Their results showed significant differences in sentence recognition between young (19 22 yr) and middle-aged (45 54 yr) subjects for speech presented with speech maskers at a 4 db signal-to-noise ratio. All subjects had normal pure-tone thresholds from 250 to 4000 Hz. The speech perception data in the present study show that subjects with normal hearing sensitivity (YNH and MANH) performed significantly better than subjects with HFSNHL (MAHL) for word recognition in quiet. The young-aged subjects performed significantly better than the middle-aged subjects with HFSNHL for word recognition in noise performance. Further, the two middle-aged groups differed considerably in word recognition in babble performance (62% vs. 44% for MANH and MAHL, respectively), albeit not a statistically significant difference (p 5.056). The data from the present study taken together with data from Horwitz et al (2002) and Helfer and Vargo (2009) suggest that highfrequency channels contribute to speech perception and that age-related deficits reduce word recognition in babble performance. Relations Between Self-Report and Word Recognition Significant relations were found between selfreported measures of listening problems and word recognition performance in babble for each subject group (Fig. 6). Word recognition performance in babble improved as the frequency of listening problems in background noise decreased for both groups of middle-aged subjects (MANH and MAHL). In addition, word recognition performance improved for the MAHL subjects when the subjects perceived fewer problems hearing in reverberant environments. No significant correlations were found either between GDTs and selfreported measures or between GDTs and word recognition performance in the current study. In summary, middle-aged subjects showed significantly poorer speech perception abilities when materials were presented in challenging tasks. In tests where younger subjects were challenged (e.g., word recognition in speech 401

10 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 maskers), middle-aged subjects performed significantly poorer than their younger counterparts. In speech perception tasks where performance was near ceiling effects for young subjects, the middle-aged subjects performed equally as well as younger subjects (e. g., word recognition in quiet), as seen in Figure 4. When hearing loss was present in middle-aged subjects, performance was significantly poorer than that of both the younger subjects and the middle-aged subjects without hearing loss. Gap Detection Thresholds Temporal processing is an important component of speech perception performance (Tremblay et al, 2004; Healy and Bacon, 2007), but the effects of age and hearing loss on the processing of temporal cues have not been clearly discerned. Auditory gap detection is a commonly used measure of temporal acuity and has been shown to be sensitive to effects of age and hearing loss (Lister and Roberts, 2005; Humes et al, 2010), especially if disparate gap markers are involved (i.e., a BC condition). In the present investigation, GDTs were recorded under both WC and BC conditions at low and high presentation levels with the stimuli spectra limited to the normal low-frequency auditory channels of all subject groups. The results from the present study showed that although a high-frequency hearing loss appears to influence the gap detection performance of low-frequency auditory channels under both gap conditions, the BC gap paradigm may prove to be particularly sensitive to the effects of aging. The data from the present study showed considerable variability for the BC GDTs, and further investigation is warranted to identify the factors that contribute to this variability (e.g., attention to the task, differences in pure-tone thresholds within the range of clinically normal, etc.). Recent evidence supports the contention that mechanisms involved in BC gap detection are perceptually and physiologically different from that mediating WC gap detection and likely tap different coding mechanisms in the central auditory systems. In many ways, BC listening resembles complex auditory tasks such as speech perception. GDTs in BC designs are longer than GDTs in WC designs, indicating that BC (i.e., across) processing (e.g., different frequencies before and after the silent gap) is more difficult than WC processing and results in a higher threshold (e.g., Pichora-Fuller et al, 2006). The present gap detection data demonstrate a difference in gap thresholds (i.e., BC thresholds were consistently poorer than WC thresholds), supporting the idea that the mechanisms mediating these gap tasks are different (Phillips and Smith, 2004; Elangovan and Stuart, 2008). In addition, the current results suggest that the WC gap detection task may be sensitive to hearingloss-related temporal processing deficits, whereas the BC gap detection task may be sensitive to both hearingloss- and age-related deficits (see the differences among groups in Figure 5). Significant differences in gap detection performance for higher-order processing deficits have been shown to be task dependent (i.e., WC or BC) for aged individuals (Lister et al, 2002; Lister and Tarver, 2004) and for auditory processing deficits in children (Phillips et al, 2010). Only recently have investigators examined the temporal processing abilities of middle-aged subjects. Grose, Hall, and Buss (2006) measured temporal processing with gap-duration discrimination tests for within- and across- (between-) channel markers in young (18 27 yr), middle-aged (40 55 yr), and elderly (65 83 yr) subjects. The pure-tone thresholds of the young and middle-aged subjects were within normal limits from 250 to 8000 Hz. The results revealed significantly poorer gap-duration discrimination performance for the middle-aged group compared with the young group for both fixed and random duration markers. Similar findings were reported in a follow-up investigation of temporal fine structure (Grose and Mamo, 2010). These investigators measured interaural phase discrimination in young (18 27 yr), middle-aged (40 55 yr), and older subjects (63 75 yr) with normal hearing sensitivity for the stimulus frequency range (i.e., Hz). The results revealed that the processing of temporal fine structure cues was reduced in middle-aged subjects compared with young subjects but not as much as in older subjects. Similar evidence for early age-related deficits in temporal processing abilities also has been reported by Babkoff et al (2002), who measured lateralization acuity based on interaural time differences (ITDs) and interaural level differences (ILDs) for young (22 29 yr), middle-aged (34 49 and yr), and older (66 73 yr) subjects with normal hearing sensitivity for their age. Their results revealed that ITD-based lateralization acuity changed significantly as a function of increasing age across groups. In contrast, the ILD-based lateralization did not significantly differ across these age groups. Further, several studies have reported significant differences between middle-aged and young normal-hearing subjects for duration difference limens, frequency difference limens (Abel et al, 2000), and homophasic and antiphasic thresholds for a speech masking level difference task (Wilson et al, 1985). The importance of high-frequency coding of lowfrequency signals is supported by recent animal (Yin et al, 2008) and human (Feng et al, 2009) data. Yin et al (2008) investigated the effects of HFSNHL on gap-evoked responses in the inferior colliculus (IC) and the auditory cortex (AC) in guinea pigs before and after noise-induced HFSNHL. Gap thresholds were measured for a low-frequency stimulus at 45, 65, and 85 db SPL in the IC and in the AC. Responses were recorded prior to acoustic trauma and up to eight weeks after acoustic trauma. At four weeks after acoustic trauma 402

11 Temporal Processing/Leigh-Paffenroth and Elangovan significant differences were observed in gap thresholds in the IC at 85 db SPL. The high-frequency hearing loss resulted in elevated (i.e., poorer) GDTs compared with the GDTs for normal high-frequency sensitivity. In the AC, GDTs were elevated by eight weeks after acoustic trauma, even at 45 db SPL. A similar study in humans (Feng et al, 2009) showed that adults with steeply sloping HFSNHL had poorer performance on a modulation detection task and a gap detection task compared with age-matched subjects with normal hearing sensitivity through 8000 Hz. Time-compressed word recognition in noise performance was measured by the Hearing in Noise Test (HINT), and higher signal-to-noise ratios were required for the subjects with HFSNHL compared with the subjects with normal hearing sensitivity for equivalent speech recognition performance. Further, significant correlations were found between modulation detection performance and the HINT scores. The gap detection data from the current study are in agreement with the Yin et al and Feng et al conclusions, which showed significant differences in temporal processing measures for low-frequency stimuli between subjects with HFSNHL and subjects with normal high-frequency pure-tone sensitivity. Summary Earlier investigations on temporal processing in subjects with hearing loss were not restricted to regions of normal hearing sensitivity, as most experiments included subjects with some degree of high-frequency hearing loss (e.g., Bacon and Viemeister, 1985; Lister and Roberts, 2005). A clear understanding of the potential effects of HFSNHL on auditory perception is important because it affects up to 93% of males in the United States aged yr (Agrawal et al, 2008). In addition, Agrawal et al found that the prevalence of high-frequency hearing loss is increasing in young adults, particularly those with a history of noise exposure or poor cardiovascular health. The current study reduces presentation level variability, which plagues many studies of hearing impairment, by presenting all stimuli in the frequency range of normal hearing sensitivity for all subjects. Middle-aged subjects with normal hearing sensitivity reported significantly more problems hearing in background noise ( p 5.042) and had poorer word recognition performance across conditions than young subjects with normal hearing sensitivity ( p 5.035). The middle-aged subjects with HFSNHL had poorer word recognition than either group with normal hearing sensitivity. The presence of high-frequency hearing loss significantly increased gap detection thresholds, decreased word recognition performance in quiet, and increased self-perceived hearing problems. Significant effects of aging have been well studied in elderly subjects, and this study extends the effects of aging to include middle-aged persons. In addition, the effects of high-frequency hearing loss have been extended to include a reduction in performance for auditory temporal processing and word recognition tests in regions of normal hearing sensitivity as well as self-reported problems hearing. These results provide further evidence of the confounding effects of hearing loss and aging. NOTES 1. The output impedance of both the filter and the insert earphone was 50V. 2. Word recognition data points were missing for two YNH subjects and two MAHL subjects. REFERENCES Abel SM, Giguere C, Consoli A, Papsin BC. (2000) The effect of aging on horizontal plane sound localization. J Acoust Soc Am 108: Abel SM, Krever EM, Alberti PW. (1990) Auditory detection, discrimination and speech processing in ageing, noise-sensitive and hearing-impaired listeners. Scand Audiol 19: Agrawal Y, Platz EA, Niparko JK. (2008) Prevalence of hearing loss and differences by demographic characteristics among US adults: data from the National Health and Nutrition Examination Survey, Arch Intern Med 168: American National Standards Institute. (2004) Methods for Manual Pure-Tone Threshold Audiometry. ANSI S New York: American National Standards Institute. Babkoff H, Muchnik C, Ben-David N, Furst M, Even-Zohar S, Hildesheimer M. (2002) Mapping lateralization of click trains in younger and older populations. Hear Res 165: Bacon SP, Viemeister NF. (1985) Temporal modulation transfer functions in normal-hearing and hearing-impaired listeners. Int J Audiol 24: Beattie RC. (1989) Word recognition functions for the CID W-22 test in multitalker noise for normally hearing and hearingimpaired subjects. J Speech Hear Disord 54: Bilger RC, Nuetzel JM, Rabinowitz WM. (1984) Standardization of a test of speech perception in noise. J Speech Hear Res 27: Boersma P, Weenink D. (2008) Praat: doing phonetics by computer. Carhart R, Tillman TW. (1970) Interaction of competing speech signals with hearing losses. Arch Otolaryngol 91: Cox RM, Alexander GC. (1995) The Abbreviated Profile of Hearing Aid Benefit. Ear Hear 16: Cox RM, Alexander GC, Gray GA. (2003) Audiometric correlates of the unaided APHAB. J Am Acad Audiol 14: Department of Veteran Affairs. (1998) Speech Recognition and Identification Materials. Mountain Home, TN: VA Medical Center. Divenyi PL, Haupt KM. (1997) Audiological correlates of speech understanding deficits in elderly listeners with mild-to-moderate hearing loss. II. Correlation analysis. Ear Hear 18: Dubno JR, Dirks DD, Morgan DE. (1984) Effects of age and mild hearing loss on speech recognition in noise. JAcoustSocAm76:

12 Journal of the American Academy of Audiology/Volume 22, Number 7, 2011 Dubno JR, Lee FS, Matthews LJ, Mills JH. (1997) Age-related and gender-related changes in monaural speech recognition. J Speech Lang Hear Res 40: Elangovan S, Stuart A. (2008) Natural boundaries in gap detection are related to categorical perception of stop consonants. Ear Hear 29: Feng Y, Yin S, Kiefte M, Wang J. (2009) Temporal resolution in regions of normal hearing and speech perception in noise for adults with sloping high-frequency hearing loss. Ear Hear 31(1): Fitzgibbons PJ, Gordon-Salant S. (1996) Auditory temporal processing in elderly listeners. J Am Acad Audiol 7: Florentine M, Reed CM, Rabinowitz WM, Braida LD, Durlach NI, Buus S. (1993) Intensity perception. XIV. Intensity discrimination in listeners with sensorineural hearing loss. J Acoust Soc Am 94: Formby C, Gerber MJ, Sherlock LP, Magder LS. (1998) Evidence for an across-frequency, between-channel process in asymptotic monaural temporal gap detection. J Acoust Soc Am 103: Gordon-Salant S. (1987) Age-related differences in speech recognition performance as a function of test format and paradigm. Ear Hear 8: Gordon-Salant S, Fitzgibbons PJ. (1993) Temporal factors and speech recognition performance in young and elderly listeners. J Speech Hear Res 36: Grose JH, Hall JW, III, Buss E. (2006) Temporal processing deficits in the pre-senescent auditory system. J Acoust Soc Am 119: Grose JH, Mamo SK. (2010) Processing of temporal fine structure as a function of age. Ear Hear 31(6): Healy EW, Bacon SP. (2007) Effect of spectral frequency range and separation on the perception of asynchronous speech. J Acoust Soc Am 121: Helfer KS, Vargo M. (2009) Speech recognition and temporal processing in middle-aged women. J Am Acad Audiol 20: Hirsh IJ. (1950) Binaural hearing aids: a review of some experiments. J Speech Disord 15: Horev N, Most T, Pratt H. (2007) Categorical perception of speech (VOT) and analogous non-speech (FOT) signals: behavioral and electrophysiological correlates. Ear Hear 28: Horwitz AR, Dubno JR, Ahlstrom JB. (2002) Recognition of lowpass-filtered consonants in noise with normal and impaired high-frequency hearing. J Acoust Soc Am 111: Humes LE, Kewley-Port D, Fogerty D, Kinney D. (2010) Measures of hearing threshold and temporal processing across the adult lifespan. Hear Res 264: Levitt H. (1971) Transformed up-down methods in psychoacoustics. J Acoust Soc Am 49(Suppl. 2): Lister J, Besing J, Koehnke J. (2002) Effects of age and frequency disparity on gap discrimination. J Acoust Soc Am 111: Lister J, Tarver K. (2004) Effect of age on silent gap discrimination in synthetic speech stimuli. J Speech Lang Hear Res 47: Lister JJ, Roberts RA. (2005) Effects of age and hearing loss on gap detection and the precedence effect: narrow-band stimuli. JSpeech Lang Hear Res 48: Olsen WO, Noffsinger D, Kurdziel S. (1975) Speech discrimination in quiet and in white noise by patients with peripheral and central lesions. Acta Otolaryngol 80: Phillips DJ, Souter MA, Vitkovitch J, Briggs RJ. (2010) Diagnosis and outcomes of middle cranial fossa repair for patients with superior semicircular canal dehiscence syndrome. J Clin Neurosci 17: Phillips DP, Hall SE. (2002) Auditory temporal gap detection for noise markers with partially overlapping and non-overlapping spectra. Hear Res 174: Phillips DP, Smith JC. (2004) Correlations among within-channel and between-channel auditory gap-detection thresholds in normal listeners. Perception 33: Pichora-Fuller MK, Schneider BA, Benson NJ, Hamstra SJ, Storzer E. (2006) Effect of age on detection of gaps in speech and nonspeech markers varying in duration and spectral symmetry. J Acoust Soc Am 119: Ross B, Fujioka T, Tremblay KL, Picton TW. (2007) Aging in binaural hearing begins in mid-life: evidence from cortical auditoryevoked responses to changes in interaural phase. J Neurosci 27: Roup CM, Noe CM. (2009) Hearing aid outcomes for listeners with high-frequency hearing loss. Am J Audiol 18: Snell KB, Frisina DR. (2000) Relationships among age-related differences in gap detection and word recognition. J Acoust Soc Am 107: Souza PE, Turner CW. (1994) Masking of speech in young and elderly listeners with hearing loss. J Speech Hear Res 37: Speaks C, Jerger J. (1965) Method for measurement of speech identification. J Speech Hear Res 8: Sperry JL, Wiley TL, Chial MR. (1997) Word recognition performance in various background competitors. JAmAcadAudiol8: Strouse A, Ashmead DH, Ohde RN, Grantham DW. (1998) Temporal processing in the aging auditory system. J Acoust Soc Am 104: Studebaker GA. (1985) A rationalized arcsine transform. J Speech Hear Res 28: Tremblay KL, Billings C, Rohila N. (2004) Speech evoked cortical potentials: effects of age and stimulus presentation rate. JAm Acad Audiol 15: , quiz 264. Wilson RH. (2003) Development of a speech-in-multitalker-babble paradigm to assess word-recognition performance. J Am Acad Audiol 14: Wilson RH, Civitello BA, Margolis RH. (1985) Influence of interaural level differences on the speech recognition masking level difference. Int J Audiol 24: Yin SK, Feng YM, Chen ZN, Wang J. (2008) The effect of noiseinduced sloping high-frequency hearing loss on the gap-response in the inferior colliculus and auditory cortex of guinea pigs. Hear Res 239: