Context: Deep band modulation (DBM) improves speech perception in individuals with learning disability and older adults, who had temporal impairment in them. However, it is unclear on perception of DBM phrases at quiet and noise conditions in individuals with auditory neuropathy spectrum disorder (ANSD) and sensorineural hearing loss (SNHL), as these individuals suffer from temporal impairment. Aim: The aim is to study the effect of DBM and noise on phrase perception in individuals with normal hearing, SNHL, and ANSD. Settings and Design: A factorial design was used to study deep-band-modulated phrase perception in quiet and at noise. Materials and Methods: Twenty participants in each group (normal, SNHL, and ANSD) were included to assess phrase perception on four lists of each unprocessed (UP) and DBM phrases at different signal-to-noise ratios (SNRs) (−1, −3, and −5 dB SNR), which were presented at most comfortable level. In addition, a temporal processing was determined by gap detection threshold test. Statistical Analysis: A mixed analysis of variance was used to investigate main and interaction effects of conditions, noise, and groups. Further, a Pearson product moment correlation was used to document relationship between phrase perception and temporal processing among study participants in each experimental condition. Results: In each group, a significant improvement was observed in DBM phrase perception over UP phrase recognition in quiet and noise conditions. Although a significant improvement was observed, the benefit of recognition from DBM over UP is negligible at −5 dB SNR in both SNHL and ANSD groups. In addition, as expected, a significant improvement in phrase perception in each condition was found in normal hearing than SNHL followed by ANSD. Further, in both atypical groups, a strong negative correlation was found between phrase perception and gap detection threshold in each of the experimental condition. Conclusion: This is to conclude that temporal envelope cues from DBM were made available for phrase perception in those individuals who have temporal impairment.
Keywords: Enhancement, hearing, noise, perception
|How to cite this article:|
Shetty HN, Kooknoor V. Deep band modulated phrase perception in quiet and noise in individuals with auditory neuropathy spectrum disorder and sensorineural hearing loss. Noise Health 2017;19:174-82
|How to cite this URL:|
Shetty HN, Kooknoor V. Deep band modulated phrase perception in quiet and noise in individuals with auditory neuropathy spectrum disorder and sensorineural hearing loss. Noise Health [serial online] 2017 [cited 2018 Mar 24];19:174-82. Available from: http://www.noiseandhealth.org/text.asp?2017/19/89/174/212948
| Introduction|| |
A hearing loss of cochlear origin is consistent with damaged outer hair cells in individuals who have air conduction thresholds of 40 dB hearing loss (HL) or less. However, hearing loss of at least 60 dB HL is likely to involve the destruction of inner hair cell (IHC) and/or afferent nerve dysfunction. The agreement between audibility and speech identification is relatively stronger in damage to cochlear structures than compared to retrocochlear pathologies. It has been observed in clinic that hearing impaired individuals often complain about difficulty to follow spoken message, especially in noisy situations. Apart from problem with audibility Moore et al. have reported that individuals with cochlear hearing loss suffer from reduced frequency resolution and near or normal temporal resolution. In contrast, Madden and Feth have documented that temporal resolution was significantly reduced in the hearing impaired individuals in whom the hearing thresholds reach greater than moderate degree.
Another type of sensorineural hearing loss (SNHL), who suffers from temporal asynchrony, is auditory neuropathy spectrum disorder (ANSD). Starr et al. operationally defined ANSD by severely distorted auditory brainstem response with preserved otoacoustic emissions. Starr et al. speculated the possible site of pathology in neuronal axon, ganglion of auditory conducting fibers, synaptic junction between IHCs and auditory nerve, auditory nerve itself, and thin myelination on the surviving auditory nerve fibers. Due to which they often complain that they can hear but are unable to follow a speech. Besides audibility problem, some patient demonstrated temporal processing impairment. Animesh and Yathiraj investigated temporal processing using gap detection test (GDT) in individuals with ANSD. The results revealed that gap required to detect was larger in ANSD group than normal hearing group at both 10 and 40 dB SL. Such temporal processing impairment underlines the coding of rapidly varying components of speech. In a similar line of study, it was empirically proven by Kumar and Jayaram, who observed that speech perception deficit in ANSD was strongly correlated with temporal processing abilities.,,,,,,
It is well-established that speech enhancement approaches showed improvement in speech perception among individuals with hearing impairment. Langhans and Strube have used the principle of temporal modulation transfer function to develop the power law expansion scheme for envelope enhancement. In this strategy, envelope of speech was extracted from the series of twenty low-pass filters. The peaks from extracted envelope were enhanced and the amplitudes of troughs were reduced. They used expansion scheme to a speech stimuli before and after applying speech shaped noise. Results revealed that speech perception was relatively better when it was applied to speech before it is mixed with the noise than compared to after mixing with a noise. However, Fu and Shannon reported that output from expansion scheme distorts a signal due to modified consonant and vowel ratio. To solve this issue, Apoux and Bacon applied envelope power law to enhance the low-amplitude speech components and compressed the high-amplitude speech components. They observed a significant perception improvement in noise by the cochlear hearing loss individuals. Similarly, in ANSD, speech enhancement strategies have shown an improvement on speech perception that includes modification of temporal scaling, envelope enhancement, and combined spectrotemporal enhancement. These signal enhancement strategies were found beneficial in both syllable level and word level.
Yet another strategy is deep band modulation (DBM) that is prepared based on the principle of compression and expansion schema. The DBM enhances the modulation depth of vocalic sound and increase the time scale of entire duration. In DBM, the extracted temporal envelope bandwidth between 3 and 30 Hz from each channel is enhanced by 15 dB, which significantly increases the modulation depth such that masking of a consonant by a vowel is minimized. It was found that DBM improved speech perception score in learning disabled children who typically demonstrates temporal processing deficit. In yet another similar line of study by Hemanth and Akshay, who utilized DBM scheme to study phrase perception among older adults at noise conditions. Older adults were selected because of having temporal resolution impairment in them. It was observed that phrase perception scores improved in noise. This is because increased amplitude modulation depth caused by DBM was less energetically masked by noise.
It is believed that impaired auditory system requires an additional time to access words in the auditory lexicon, especially when the spectral and temporal contents of speech altered either by an impaired system or a noise. This process may results in either misperception or may take more time to follow a message. Significantly, in DBM, apart from enhancing an envelope it also rescales entire duration of speech. Thus, speech output from the DBM provides cue for the impaired auditory system to access the content of speech. Hence, the effect of DBM on perception of speech in noise is necessarily important among individuals with hearing impairment who have temporal processing deficit. From the literature,, it is understood that the speech perception was improved in their study participants who had temporal impairment in them. If this is true, then after correcting audibility the envelope enhancement by DBM should improve speech perception in ANSD and SNHL patients, who also have temporal impairment. Thus, it is hypothesized that participants of the study may access higher amplitude modulation which exceeds the noise level. In addition, rescaling of entire duration provides a sufficient processing time for them for retrieving information. Thus, the aim of the study is to investigate the effect of DBM on phrase perception in noise upon individuals with ANSD and SNHL. The following specific objectives were formulated to study the aim: (1) to compare phrase perception between the DBM and unprocessed (UP) conditions at different signal-to-noise ratios (SNRs), in each group, (2) to compare between groups in deep band modulated and UP phrase perception at each SNR, and (3) to find the relationship between temporal processing ability and speech perception from the participants of the study.
| Materials and methods|| |
A factorial research design was utilized to investigate the phrase perception of deep band modulated phrases at different SNRs in individuals with normal hearing, SNHL, and late onset ANSD. The following procedures were utilized to collect the data.
A total of 60 participants in three groups were included in the study. Each group comprised of 20 participants within the age range of 15 to 40 years. All the participants had normal middle ear status as indicated by “A” type tympanogram, adequate speech and language skills, and they were fluent in speaking Kannada.
None of the participants have complained of psychological and cognitive problems.
In group-1, those participants who had normal hearing sensitivity were included. The hearing sensitivity from 250 to 8 kHz in octave frequencies was within 15 dB HL or less. The speech recognition score was at least 90%. In group-2, participants who had bilateral moderate flat sensorineural hearing impairment were involved. The pure tone average ranged from 41 to 55 dB HL and had flat configuration of audiogram. The flat contour was operationally defined in which those participants in whom hearing sensitivity were less than 25 dB between the lowest and highest thresholds within octave frequencies from 0.25 to 8 kHz. The speech recognition score was at least 70%. All participants had measurable reflexes at least in any of the octave frequencies from 0.25 to 2 kHz in ipsilateral and contralateral ears. The reflex decay result revealed negative in all participants of group-2. In group-3, individuals with late onset ANSD were involved. The proposed criteria for diagnosing ANSD given by Berlin et al. was used. As per the above criteria, clients who had preserved cochlear amplification, impaired neural response (absent or abnormal brainstem responses and middle ear reflexes), normal otological function, and no space occupying lesion (identified based on clinical neurological examination) were included. The pure tone average ranged from 41 to 55 dB HL. Minimum speech recognition score of at least 30% is to be present to include as study participants.
The four lists of standardized phrases were used. Each list comprised 10 phrases. The phrases in first list retained as it is without adding noise. The second phrase list was digitally mixed with phrase shaped spectrum noise at −1 dB SNR using the AUX viewer (Machine Learning, Maryland, USA). The noise onset proceeded by 300 ms from the onset of each phrase and continued 300 ms after the end of phrase. The noise was ramped using the Cosine square function with ramp duration of 30 ms. The onset of the noise before the onset of each phrase is believed to guard against unintended onset effects. A similar procedure was carried out to digitally mix the third and fourth phrase lists with the phrase shaped spectrum noise at −3 and −5dB SNR, respectively.
Further, same four phrase lists were utilized to prepare the deep band modulated (DBM) version of phrase lists using Praat software (Praat-version 5.3.56, developed by the Institute of Phonetic Science, University of Amsterdam, Netherlands), in which the algorithm utilized was adopted by Nagarajan et al. The following procedure was used to prepare DBM version of phrases. The each phrase was passed through a 20-s order Butterworth filters by filter bank method. The center frequencies from these 20 filters were logarithmically spaced between 100 Hz and 10 kHz. From the output of each narrow band channel an envelope was extracted (i.e., Hilbert transform was computed using Fast Fourier Transformer). The envelope in each narrow band channel was filtered using second order Butterworth filters with cutoff frequencies between 3 and 30 Hz. Further, in each channel the envelope was rectified. Before summating the envelope from each channel, a gain of 15 dB was provided for the channels within the frequency range of 1 to 4 kHz to obtain the resultant deepen band modulated phrase. The DBM phrases in first list retained as it is without adding noise. The deep band modulated version of each phrase in list-2 was digitally mixed with deep band modulated phrase spectrum shaped noise at −1 dB SNR using AUX viewer. A similar procedure was carried out to digitally mix the noise with the deep band modulated phrase list-3 and list-4 at −3 and −5dB SNR, respectively. From [Figure 1], it can be observed that amplitudes of consonants were relatively increased in the deep band modulated phrase than the UP phrase in quiet and at different SNRs. In addition, the modulation depth is more in DBM than UP condition.
|Figure 1: UP speech in quiet (a), −1 dB SNR (b), −3 dB SNR (c), and −5 dB SNR (d) are represented in first row. In the second row the deep band modulated phrases in quiet (e), −1 dB SNR (f), −3 dB SNR (g), and −5 dB SNR (h) are represented|
Click here to view
The participants were blinded regarding the experimental conditions. Four phrase lists of each in UP and DBM phrases in quiet and at different SNRs were utilized to investigate the phrase perception scores by each group of the study participants. A total of eight lists were randomized and counterbalanced across study participants. The phrases stored in personal laptop were routed through the auxiliary input of audiometer. The output of the audiometer was delivered in both ears through headphones at participant’s most comfortable level. Each participant was instructed to repeat the phrase heard. Further, succeeding phrase was presented once after repeating the phrase was heard. A score of one mark was assigned for the correct repetition of whole phrase, and a score of zero was allotted if the participant failed to repeat or partly repeat the phrase.
Gap detection threshold
Temporal processing ability from each group of the study participants was obtained from the gap detection test (GDT). Apex (Neurosciences, Leuven, Belgium) (version 3) software stored in personal laptop was used as platform to deliver the gap in three interval forced choice method. The output of laptop was routed through the audiometer. The output of audiometer was presented in both ears through the headphones at participant’s most comfortable level. The paradigm comprised three blocks of broad band noise having duration of 500 ms each. The interblock interval was 200 ms. In one among three blocks of noise, the gap used to be present at center. The noise were shaped with a cos2 window to create a 1-ms rise–fall time around the gap and a more gradual 10-ms rise-fall time on the onset of the first noise and the offset of the second noise. Further, occurrence of gap in one of the block was randomized in each trial. The duration of gap in succeeding trial was decided based on participant response in the previous trial. An adaptive two down one up procedure of eight reversals was utilized. The last eight reversals were averaged to obtain the participant 70.3% ability to detect the minimum gap present in one among the three blocks.
| Results|| |
The data collected were subjected to statistical analyses using SPSS (version 17) (IBM Corp, Armonk, New York). The analyses performed under each objective of the study are reported as follows.
Comparison of phrase perception between unprocessed and deep band modulated conditions at different SNRs, in each group
Descriptive statistical analyses were performed to document the mean and standard deviation of phrase recognition scores for the UP and deep band modulated phrases in quiet and at different SNRs by the three different groups [Table 1]. It was noted in quiet condition, in group-1, the mean phrase recognition scores were same in both UP and DBM conditions. However, it was noted that in each group the phrase perception score was better in DBM condition than UP condition at each SNR.
|Table 1: Phrase perception scores obtained in each condition from quiet and at each SNR, in different groups|
Click here to view
A mixed analysis of variance (ANOVA) was conducted to see the main and interaction effects of conditions (UP and DBM) and SNRs, in participants of within group and between groups. The results revealed that a significant main effect for the conditions [F (1, 57) = 155.26, P = 0.000], SNRs [F (3, 171) = 364.40, P = 0.000], and the interaction effect of conditions × groups [F (2, 57) = 13.09, P = 0.000], SNRs × groups [F (6, 171) = 48.44, P = 0.000], conditions × SNR [F (3, 171) = 14.88, P = 0.000], and conditions × SNR × groups [F (6, 171) = 9.88, P = 0.000], such that, the mean phrase perception scores in DBM condition was significantly better than the UP condition in each SNR. This was true in each group.
To further evaluate the conditions (UP vs. DBM) in which the phrase perception was significantly better in both quiet and at different SNRs, in each group; paired samples t tests were conducted. In group-1, the phrase perception scores are same for both UP and DBM condition without SD; thus, it was not subjected to statistical analysis. A three paired comparisons were performed to evaluate in which SNRs caused the significant difference between conditions. These comparisons resulted, a power of significance 0.016 instead of 0.05. The results revealed that though the phrase perception was better in DBM condition than that of UP condition in each SNR, the significant difference was noted only in −5 dB SNR [Figure 2].
|Figure 2: Phrase perception comparison between UP and DBM in each SNR from participants of group-1|
Click here to view
Further, four paired comparisons were conducted to evaluate, in which SNRs caused the significant difference between conditions in group-2 and group-3 separately. These comparisons result a power of significance 0.012 instead of 0.05. The phrase perception comparisons between UP and DBM in each SNR are represented in [Figure 3] and [Figure 4]. In group-2, the results of paired samples t tests [Figure 3] revealed that phrase perception scores were significantly better in DBM than UP in −1 dB SNR [t (19) = −3.83, P < 0.001], −3 dB SNR [t (19) = −5.57, P < 0.000], and −5 dB SNR [t (19) = −5.99, P < 0.000].
|Figure 3: Phrase perception comparison between UP and DBM in each SNR from participants of group-2|
Click here to view
|Figure 4: Phrase perception comparison between UP and DBM in each SNR from participants of group-3|
Click here to view
In group-3, it was observed except in quiet, standard deviation was more than that of mean phrase perception in UP condition at each SNR [Figure 4]. In DBM condition, although the mean phrase perception is higher than its standard deviation, the scores are low in 3.45, 2.20, and 1.05 at −1, −3, and −5 dB SNR, respectively. The results of paired samples t tests revealed that phrase perception scores were significantly better in DBM than UP in −1 dB SNR [t (19) = −6.89, P < 0.000], −3 dB SNR [t (19) = −6.05, P < 0.000], and −5 dB SNR [t (19) = −5.10, P < 0.000].
Phrase perception between groups, in each experimental condition
A significant main effect for group was noted in a mixed ANOVA [F (1, 18) = 67.97, P = 0.000]. Thus, so as to know in which experimental conditions caused significant difference between groups, a multivariate ANOVA (MANOVA) was performed. The results from MANOVA are displayed in [Table 2]. The result revealed that irrespective of conditions and SNRs, the mean phrase perception scores were significantly better in group-1 than group-2 followed by group-3.
|Table 2: F ratio and P value of MANOVA obtained from groups in each experimental condition|
Click here to view
Further, to know in between which groups caused a significant difference on phrase recognition, a Duncan post hoc test was performed on each experiment condition. The result of Duncan test revealed that in quiet condition, the phrase perception score in group-3 was significantly reduced compared to that of group-1 and group-2 in both UP and DBM conditions (P < 0.05). However, though the phrase perception score in quiet condition was better in group-1 than compared to group-2, this difference did not reach significant. Further, it was noted in each SNR (i.e., −1, −3, and −5) the phrase perception scores were significantly different (P < 0.05) between groups for both of the conditions (UP and DBM), such that reduced perception scores were noted in group-3 than compared to group-2 and group-1, respectively. Further, the perception score reduced in group-2 compared to that group-1.
Relationship between gap detection threshold and phrase perception scores in each experimental condition
The mean gap detection threshold (SD) was higher in group-3 [34.15 (12.02) in ms] than group-2 [11.55 (5.70) in ms] followed by group-1 [3.84 (0.95) in ms]. The data of gap detection threshold (GDT) obtained from three groups were subjected to a one way ANOVA [F (2, 57) = 83.75, P = 0.000]. The result revealed a significant difference between groups in GDT. Besides, a post hoc Duncan test was performed to see which groups had caused significant difference. The results revealed that each group was significantly different from each other in GDT.
The correlation coefficient [Table 3] was computed between GDT and phrase perception score in each experimental condition using Pearson product-moment correlation. The Pearson correlation coefficient results revealed that there was a significant strong negative correlation between GDT and phrase perception score in each experimental condition. It indicates the higher GDT was noted in individuals having lower phrase perception scores [Figure 5].
|Figure 5: Representing scatter plot between GDT and phrase perception score at each SNR in both (a) UP and (b) Deepen band modulated conditions|
Click here to view
| Discussion|| |
Phrase perception in quiet and at noise in UP and DBM conditions
In all study participants, phrase perception scores in DBM were better than UP phrase perception scores in quiet and at different SNRs. In hearing group-1 participants of normal hearing, DBM phrase perception scores at −5 dB SNR reached significantly better than UP condition. In UP phrases the noise at −5 dB SNR mask temporal dips and alters the spectral content in it. Although, in DBM phrases, though the noise at −5 dB SNR modify spectral parameters, higher amplitude of modulation depth in phrases are less energetically masked, this provides cues for the normal hearing individuals to follow the phrases even in noise. In addition, normal temporal resolution in them is advantageous to detect subtle changes in the amplitude of modulation depth embedded in noise.
In group-2 participants of SNHL, DBM significantly improved the phrase perception at quiet and at −1, −3, and −5 SNRs. Improved perception of DBM can be reasoned that enhancing the consonant portion of speech and compressing the vocalic portion may have improved the listeners’ ability to process the slow modulation envelope of phrases (3–30 Hz). The modulation depth enhanced by 15 dB was found beneficial and made the temporal dips cue more resistant though obscured by noise. Though phrase recognition was significantly improved in DBM than UP at each SNR, the amount of benefit received from DBM over UP condition was negligible at reduced SNR, especially at −5 dB SNR. This is because available healthy filters are taxed at reduced SNR and experienced a difficulty to recognize phrase. In addition, mean temporal resolution in hearing impaired individuals is 11.55 ms. Those individuals who obtained high gap detection threshold have received reduced perception abilities in both UP and DBM conditions. This indicates from 15 dB modulation gain in phrases, an available amplitude of temporal dips in noise at −5 dB SNR are unable to process due to temporal impairment in them.
In group-3 participants of ANSD, phrase perception in DBM condition was better than UP condition. This is due to enhancing the consonant portion of speech and compressing vowel portion of speech thereby increases a consonant vowel distinction. This is in line with previous findings that higher consonant to vowel ratio improves the word intelligibility. Although at −5 dB SNRs relative to other SNRs and quiet conditions, the amount of improvement obtained from DBM over UP is negligible. This is because the −5 dB SNR significantly reduces the depth of modulation in phrase. In addition, noise distorts the temporal fine structures and obliterates envelope cues for identification.. Liu and Zeng reported that in adverse listening condition, normal hearing individuals are able to follow speech by extracting the fine structure cues. Whereas, individuals with ANSD were unable to process the accessible fine structure temporal cues obscured by noise. The detrimental effect of noise at reduced SNR (−dB SNR) on increased modulation depth in phrase by temporal enhancement is unknown. One possible speculation could be that individuals with ANSD failed to extract available higher amplitude envelope which are not masked by noise is due to asynchronous firing. The explanation can be substantiated by the results of gap detection threshold, in whom temporal resolution impairment is evident with the GDT of 34.15 in ms.
Phrase perception ability between groups of study participants
At quiet in unprocessed condition
In ANSD, phrase perception scores from UP condition was significantly reduced than SNHL and normal hearing group. The reason could be due to the greater temporal asynchrony in ANSD, who might have failed to precisely encode an ongoing variation of available temporal modulation depth in phrases. Yet another reason for reduced phrase perception could be a neural involvement rather at a synaptic junction between IHCs and auditory nerve. This conjecture is substantiated by the research reports of Starr et al. and Sininger and Oba who showed their 20 to 30 percent of ANSD study participants had speech perception ability close to the range of perception ability by SNHL. They speculated their finding to the possible etiology in synaptic junction between IHCs and auditory nerve in them.
At quiet in deep band modulated condition
In DBM condition, the phrase perception score was significantly reduced in ANSD than SNHL. The significant difference was due to higher standard deviation in ANSD. Otherwise, the mean speech perception scores were 8.30 and 9.25 for ANSD and SNHL participants, respectively. This result might be attributed to the improved accessibility of temporal cues in the DBM condition for both groups. The one possible reason could be the impaired auditory system tracks the temporal changes caused by increased modulation depth by 15 dB gain which may lead to increased phase-locking of neural discharges.
At noise in unprocessed condition
Phrase perception scores at different SNRs were reduced in SNHL participants than normal hearing patients and the reduction is dramatic at lesser SNRs. This result may be because in cochlear hearing loss patients auditory filters are broadened. A wider auditory filter does not mean it removes information from speech rather it impedes the transfer of spectral and temporal information. It can be expected that spectral peaks and valleys in phrase stimulus are smoothed out in those individuals with SNHL. In addition, downward spread of masking, that is, the high frequency components of speech shaped noise masks the low frequency components in phrase stimulus, which is found to be more in SNHL than normal participants. It can also be assumed that only a few auditory filters are available for analysis. But the noise accompanied with phrase stimulus taxes these available filters such that noise accumulates in functioning filters leading to reduced recognition in lesser SNRs. Another explanation, evident from the present study, is temporal resolution that is impaired in SNHL patients. That is, they require larger gap to detect. It seems reasonable to expect SNHL patients enable to detect the subtle changes when the spectral peaks and valleys are masked by noise.
The phrase perception scores in UP condition at quiet and noise were significantly reduced in ANSD than SNHL. It clearly indicates that site of pathology in these two groups are different. The reason could be due to the greater temporal asynchrony in ANSD, who might have failed to precisely encode the ongoing rapid variation of available spectrotemporal cues. This dyssynchronous firing in the auditory nerve impairs processing of rapidly varying signals at auditory subcortical level. These physiological changes have concomitant impairment in the encoding of spectrotemporal cues at auditory cortical level. Whereas, in noise at different SNRs the phrase perception scores were reduced in ANSD than SNHL and it was observed to be dramatically reduced at lesser SNRs. Temporal envelope cues, a slow modulated low frequency component are found resistant in noise than temporal fine structure cues to retrieve information. On the contrary, participants of ANSD have failed to identify phrase though the envelope cues are partly available to them. Further, auditory masking might have interrupted linguistic processing of the target phrase. Thus, the asynchronous auditory system is unable to extract the linguistic information, especially in lesser SNR and or in the absence of redundancy cues.
At noise in deep band modulated condition
In DBM condition, at −1 dB SNR, phrase perception in participants of SNHL was found closer to normal hearing patients but due to higher standard deviation in SNHL, this difference reached significant. It infers that increased modulation depth by gain of 15 dB have made it possible for some of SNHL patients to utilize the high amplitude modulated fluctuations that are less energetically masked by noise. However, at −5 dB SNR, existed modulation depth may not be sufficient for the SNHL patients to recognize the phrases. Further, experiments are warranted to vary the bandwidth of modulation by increasing the gain and its effects on perception. In addition, it was noted significant reduction of phrase perception scores in ANSD than SNHL patients. It infers DBM may be of limited help for ANSD participants in noise. As noted earlier, ANSD participants exhibited dyssynchronous neural activity along different auditory pathway in which they are unable to fire precisely to the high amplitude of slow modulated envelope, which was masked by noise. Thus, temporal processing impairment in ANSD and introduction of noise obscures temporal dips and spectral content of the phrase has failed to recognize phrase.
| Conclusion|| |
In each group, the phrase recognition in DBM was found significantly better than UP. It indicates the gain of 15 dB has increased the modulation depth which made the envelope cues available to the study participants. However, at −5 dB SNR, the phrase recognition benefit from DBM over UP is negligible in SNHL and ANSD groups. In addition, DBM phrase recognition was significantly improved in SNHL than ANSD at quiet and noise and this attributed to different site of lesion.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Rosen S, Faulkner A, Smith DAJ. The psychoacoustics of profound hearing impairment. Acta Otolaryngol Suppl 1990;469:16-22.
Moore BC, Glasberg BR, Simpson A. Evaluation of a method of simulating reduced frequency selectivity. J Acoust Soc Am 1992;91:3402-23.
Madden JP, Feth LL. Temporal resolution in normal hearing and hearing-impaired listeners using frequency-modulated stimuli. J Speech Hear Res 1992;35:436-42.
Starr A, Picton TW, Sininger Y, Hood LJ, Berlin CI. Auditory neuropathy. Brain 1996;119:741-53.
Starr A, Michalewski HJ, Zeng FG, Fujikawa-Brooks S, Linthicum F, Kim CS et al.
Pathology and physiology of auditory neuropathy with a novel mutation in the MPZ gene (Tyr145 → Ser). Brain 2003;126:1604-19.
Animesh B, Yathiraj A. Pyscho-Acoustic Profile in Normals and individuals with Auditory Dys-Synchrony. Unpublished Thesis: Submitted to Mysuru University; 2007.
Kumar AU, Jayaram M. Speech perception in individuals with auditory dys-synchrony. J Laryngol Otol 2011;125:236-45.
Humes LE, Humes LE, Wilson DL. A comparison of single-channel linear amplification and two-channel wide-dynamic-range-compression amplification by means of an independent-group design. Am J Audiol 2004;13:39-53.
Kochkin S. MarkeTrak VI: 10-Year customer satisfaction trends in the U.S. Hearing Instrument Market. Hear Rev 2002;9:14-25, 46.
Baer T, Moore BCJ. Effects of spectral smearing on the intelligibility of sentences in the presence of noise. J Acoust Soc Am 1993;94:1229-41.
Amlani AM. Efficacy of directional microphone hearing aids: A meta-analytic perspective. J Am Acad Audiol 2001;12:202-14.
Van Tasell DJ. Hearing loss, speech, and hearing aids. J Speech Hear Res 1993;36:228-44.
Sininger YS, Hood LJ, Starr A, Berlin CT, Picton TW. Hearing loss due to auditory neuropathy. Audiol Today 1995;7:16-8.
Rance G, Cone-Wesson B, Wunderlich J, Dowell R. Speech perception and cortical event related potentials in children with auditory neuropathy. Ear Hear 2002;21:239-53.
Langhans T, Strube HW. Speech enhancement by nonlinear multiband envelope filtering. IEEE International Conference on Acoustic, Speech and Signal Process, 1982, pp. 156-9.
Fu QJ, Shannon R. Effect of amplitude nonlinearity on phoneme recognition by cochlear implant users and normal-hearing listeners. J Acoust Soc Am 1998;104:2570-7.
Apoux F, Bacon SP. Relative importance of temporal information in various frequency regions for consonant identification in quiet and in noise. J Acoust Soc Am 2004;116:1671-80.
Nagarajan SS, Wang X, Merzenich MM, Schreiner CE, Johnston P, Jenkins WM et al.
Speech modifications algorithms used for training language learning-impaired children. IEEE Trans Rehabil Eng 1998;6:257-68.
Hemanth N, Akshay M. Deep band modulation and noise effects: Perception of phrases in adult. Hear Balance Commun 2015;13:111-7.
Berlin CI, Hood LJ, Morlet T, Wilensky D, St John P, Montgomery E et al.
Absent or elevated middle ear muscle reflexes in the presence of normal otoacoustic emissions: A universal finding in 136 cases of auditory neuropathy/dys-synchrony. J Am Acad Audiol 2005;16:546-53.
Houtgast T, Steeneken HJM. A review of the MTF concept in room acoustics and its use for estimating speech intelligibility in auditoria. J Acoust Soc Am 1985;77:1069-1077.
House AS, Williams CE, Hecker MH, Kryter KD. Articulation testing methods: Consonantal differentiation with a closed response set. J Acoust Soc Am 1965;37:158-66.
Drullman R. Temporal envelope and fine structure cues for speech intelligibility. J Acoust Soc Am 1995;97:585-92.
Liu S, Zeng FG. Temporal properties in clear speech perception. J Acoust Soc Am 2006;120:424-32.
Sininger Y., Oba S. Patients with auditory neuropathy: Who are they and what can they hear? Auditory Neuropathy: A New Perspectives on Hearing Disorders. San Diego: Singular Publishing; 2001.
Zeng FG, Kong YY, Michalewski HJ, Starr A. Perceptual consequences of disrupted auditory nerve activity. J Neurophysiol 2005;93:3050-63.
Kumar AU, Jayaram M. Auditory processing in individuals with auditory neuropathy. Behav Brain Funct 2005;1:21.
Hemanth Narayan Shetty
Department of Audiology, All India Institute of Speech and Hearing, Mysore, Karnataka
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3]