Identification Of Nonsense Syllables Research Articles

Consonant and Vowel Identification in Cochlear Implant Users Measured by Nonsense Words: A Systematic Review and Meta-Analysis.

The purpose of this systematic review and meta-analysis was to establish a baseline of the vowel and consonant identification scores in prelingually and postlingually deaf users of multichannel cochlear implants (CIs) tested with consonant-vowel-consonant and vowel-consonant-vowel nonsense syllables. Six electronic databases were searched for peer-reviewed articles reporting consonant and vowel identification scores in CI users measured by nonsense words. Relevant studies were independently assessed and screened by 2 reviewers. Consonant and vowel identification scores were presented in forest plots and compared between studies in a meta-analysis. Forty-seven articles with 50 studies, including 647 participants, thereof 581 postlingually deaf and 66 prelingually deaf, met the inclusion criteria of this study. The mean performance on vowel identification tasks for the postlingually deaf CI users was 76.8% (N = 5), which was higher than the mean performance for the prelingually deaf CI users (67.7%; N = 1). The mean performance on consonant identification tasks for the postlingually deaf CI users was higher (58.4%; N = 44) than for the prelingually deaf CI users (46.7%; N = 6). The most common consonant confusions were found between those with same manner of articulation (/k/ as /t/, /m/ as /n/, and /p/ as /t/). The mean performance on consonant identification tasks for the prelingually and postlingually deaf CI users was found. There were no statistically significant differences between the scores for prelingually and postlingually deaf CI users. The consonants that were incorrectly identified were typically confused with other consonants with the same acoustic properties, namely, voicing, duration, nasality, and silent gaps. A univariate metaregression model, although not statistically significant, indicated that duration of implant use in postlingually deaf adults predict a substantial portion of their consonant identification ability. As there is no ceiling effect, a nonsense syllable identification test may be a useful addition to the standard test battery in audiology clinics when assessing the speech perception of CI users.

The Effect of Low-Pass Filtering on Identification of Nonsense Syllables in Quiet by School-Age Children With and Without Cochlear Dead Regions

The presence of cochlear dead regions (DRs) can have a significant effect on speech perception. Some studies have reported that adults do not benefit from amplification of frequencies well inside an extensive DR. However, the importance of high-frequency amplification for hearing-impaired children has been emphasized by many researchers. This study investigates the benefit of high-frequency amplification for children with various degrees of high-frequency hearing impairment, with and without DRs. The children, aged 8 to 13 years, were divided into two groups according to the severity of their hearing impairment. Group MS had moderate to severe impairment (9 ears without DRs and 3 ears with restricted DRs). Group SP had severe to profound hearing impairment (7 ears with DRs and 1 ear without a DR). The vowel-consonant-vowel stimuli were subjected to the frequency-gain characteristics prescribed by the desired sensation level fitting method and presented via headphones broadband and under various low-pass filtering conditions. Group MS benefited from high-frequency amplification whether or not a restricted DR was present. In contrast, ears in group SP with continuous extensive DRs showed limited benefit from high-frequency amplification. For the latter, performance improved with increasing cutoff frequency up to approximately 1 octave above the edge frequency of the DR and generally stayed the same, or deteriorated, with further increases in bandwidth. In one case of severe to profound hearing impairment without evidence of DRs, performance increased with increasing cutoff frequency up to 2 kHz and remained almost constant with further increases in bandwidth. For children with severe to profound hearing impairment and continuous high-frequency DRs commencing from approximately 1 kHz, applying amplification only for frequencies up to approximately 1 octave above the edge frequency of the DR may be of benefit. Tests with more participants are needed to confirm this finding.

The usefulness of the modified nonsense syllable test as a measure of speech identification

The NST is a standardized speech-identification test involving closed-set identification of nonsense syllables. Its organization permits speech-identification testing with a variety of consonants and vowel contexts, but its usefulness is limited by the long time it takes to administer and the limited confusion matrices that can be generated. A modified version of the NST (MNST) was developed by Gelfand et al. (1992). They reported data on young normal-hearing listeners at presentation levels ranging from 20 dB to 52 dB SPL in quiet and in noise. A previous study extended the work of Gelfand et al. (1992) to higher presentation levels more suited for older and/or hearing impaired listeners. This work indicated that the MNST is a potentially useful measure of nonsense syllable identification. A significant advantage of the MNST is that it allows detailed analyses of consonant confusions among all the test consonants while reducing the test time greatly. The present work describes the confusion matrices of the two subsets of the MNST for a group of young normal-hearing listeners. This work further validates the MNST and its usefulness as a tool for measuring speech-identification scores and generating detailed consonant confusion matrices in a relatively short amount of time. The above work was supported in part by NIH NIDCD 5R03DC10266-2 awarded to the first author.

Efficacy of a Reverse Cardioid Directional Microphone

Directional microphones have been shown to improve a listener's ability to communicate in noise by improving the signal to noise ratio. However, their efficacy may be questioned in situations where the listener needs to understand speech originating from the back. The goal of the study was to examine the performance of a directional microphone mode that has an automatic reverse cardioid polar pattern. A single-blinded, factorial repeated-measures design was used to study the effect of microphone modes (reverse cardioid, omnidirectional, and front hypercardioid) and stimulus azimuths (front and back) on three outcome variables (aided thresholds, nonsense syllable identification in quiet, and sentence recognition in noise). Twenty adults with a mild-to-severe bilaterally symmetrical (±5 dB) sensorineural hearing loss participated. Audibility in quiet was evaluated by obtaining aided sound field thresholds and speech identification at an input level of 50 dB SPL presented at 0 and 180° azimuths. In addition, speech understanding in noise was also assessed with the Hearing In Noise Test (HINT) sentences presented at both azimuths (0 and 180°) with a diffuse noise. Repeated-measures analyses of variance (ANOVAs) were conducted to examine the effects of microphone mode (omnidirectional, front hypercardioid, reverse cardioid) and stimulus azimuth (0°, 180°) on aided thresholds, nonsense syllable identification, and HINT performance. Results with the reverse cardioid directional microphone in both quiet conditions were similar to the omnidirectional microphone. The results of the reverse cardioid microphone in noise were significantly better than the omnidirectional microphone and front hypercardioid microphone when speech was presented from the back (p < 0.001). These results support the possible benefits of a reverse cardioid directional microphone when used in specific listening situations.

Psychophysical tuning curves and recognition of highpass and lowpass filtered speech for a person with an inverted V-shaped audiogram.

A single subject whose audiogram resembled an inverted V shape (good hearing at 4000 Hz, poorer hearing at other frequencies) was tested. Results of the TEN(HL) test suggested that a dead region (DR) in the cochlea was present at all test frequencies from 500 to 3000 Hz, but no DR was present at 4000 Hz. Psychophysical tuning curves obtained using signal frequencies of 2000, 3000, 4000, and 6000 Hz showed upward shifted tips for the lowest two signal frequencies and a downward shifted tip for the highest frequency. The results suggested a functioning region extending from 3900 to 5100 Hz, with DRs outside that range. The identification of nonsense syllables, amplified according to the Cambridge formula, was measured as a function of lowpass or highpass filter cutoff frequency. The results suggested that useful speech information could only be extracted from a limited frequency range around 4000 Hz.

Psychophysical tuning curves and recognition of highpass and lowpass filtered speech for a person with an inverted V-shaped audiogram

A single subject whose audiogram resembled an inverted V shape (good hearing at 4000 Hz, and poorer hearing at other frequencies was tested. Results of the threshold-equalizing noise test suggested that a dead region (DR) in the cochlea was present at all test frequencies from 500 to 3000 Hz, but no DR was present at 4000 Hz. Psychophysical tuning curves (PTCs) obtained using signal frequencies of 2000, 3000, 4000, and 6000 Hz showed upward shifted tips for the lowest two signal frequencies, and a downward shifted tip for the highest frequency. The results of the PTCs suggested a functioning region extending from 3900 to 5100 Hz, with DRs outside that range. The identification of nonsense syllables, amplified according to the "Cambridge formula," was measured as a function of lowpass or highpass filter cutoff frequency. The results suggested that useful speech information could only be extracted from a limited frequency range around 4000 Hz.

Abnormal processing of temporal fine structure in speech for frequencies where absolute thresholds are normal

The identification of nonsense syllables that were lowpass filtered at 1.5 kHz was compared for subjects with normal hearing and subjects with mild-to-severe hearing loss at high frequencies but with normal or near-normal hearing at low frequencies. Absolute thresholds were mostly within the normal range (<20 dB hearing level) for both groups for frequencies below 1.5 kHz. Performance was assessed with intact speech, speech that had been processed to preserve only temporal envelope cues in a few frequency bands (E speech), and speech that had been processed to remove envelope cues as far as possible while preserving temporal fine structure cues, again in a few frequency bands (TFS speech). For the intact speech and E speech, the hearing-impaired subjects performed slightly more poorly than the normal-hearing subjects, but this effect was significant only for the intact speech. For the TFS speech, the hearing-impaired subjects performed significantly more poorly than the normal-hearing subjects, with 12 out of 16 of the former performing at chance. The results indicate that, for people with hearing loss at medium to high frequencies, the processing of the TFS of speech can be degraded for frequencies where absolute thresholds are within the normal range.

Combining Temporal-Envelope Cues Across Channels: Effects of Age and Hearing Loss

The goal of this study was to examine the ability to combine temporal-envelope information across frequency channels. Three areas were addressed: (a) the effects of hearing loss, (b) the effects of age and (c) whether such effects increase with the number of frequency channels. Twenty adults aged 23-80 years with hearing loss ranging from mild to severe and a control group of 6 adults with normal hearing participated. Stimuli were vowel-consonant-vowel syllables. Consonant identification was measured for 5 conditions: (a) 1-channel temporal-envelope information (minimal spectral cues), (b) 2-channel, (c) 4-channel, (d) 8-channel, and (e) an unprocessed (maximal spectral cues) speech condition. Performance of listeners with normal hearing and listeners with hearing loss was similar in the 1-channel condition. Performance increased with the number of frequency channels in both groups; however, increasing the number of channels led to smaller improvements in consonant identification in listeners with hearing loss. Older listeners performed more poorly than younger listeners but did not have more difficulty combining temporal cues across channels than in a simple, 1-channel temporal task. Age was a significant predictor of nonsense syllable identification, whereas amount of hearing loss was not. The results support an age-related deficit in use of temporal-envelope information with age, regardless of the number of channels.

The effect on speech intelligibility of varying compression time constants in a digital hearing aid

The identification of nonsense syllables in quiet and in three types of background (babble, cafeteria and single female speaker) was measured using four hearing aid compression algorithms differing in attack and release time constants, and using linear amplification. The speech level was always 65 dB SPL. The compression algorithms, which were implemented in a Phonak Claro ITE hearing aid, were: (1) ‘very fast’—the attack time was 8 ms and the release time was 32 ms, for all 20 channels; (2) ‘slow–fast’—the attack and release times decreased from 500 ms for low frequencies to about 100 ms for high frequencies; (3) ‘fast–slow’—the attack and release times increased from about 50 ms for low frequencies to 500 ms for high frequencies; and (4) ‘slow+fast’—a very slow-acting gain control signal was combined with a fast-acting gain control signal, for each channel in a 10-channel system. Acoustical stimuli were presented monaurally via a circumaural headphone mounted over the hearing aid. The linear condition did not use the Claro aid; instead, the signal was digitally filtered to implement the Cambridge formula prior to delivery via the earphone. Five subjects with moderate sensorineural hearing loss were tested in a counter-balanced order across conditions. In quiet, performance was best for linear amplification and worst for the slow+fast algorithm. In the presence of background sounds, the highest scores were obtained with the linear-gain Cambridge formula implemented via headphones; a supplementary experiment suggested that this was due to the greater high-frequency gain resulting from the use of this formula. No significant differences were found between scores for the different compression algorithms. We conclude that the intelligibility of speech at a fixed level, presented in background sounds, is not markedly affected by rather substantial variations of the time constants in a multichannel compression system.

Predicting the perception of words from the perception of phonemes

The present paper explores the degree to which the perception of CVC syllables can be factored into the perception of its constituent phonemes. Categorization experiments with synthetic speech typically manifest such factorability to a remarkable degree (regardless of lexical status of the syllables.) Such factorability is also compatible with intelligibility studies of English CVCs [ e.g., A. Boothroyd and S. Nittrouer, J. Acoust. Soc. Am. 84, 101–114 (1988)]. These show that the identification of nonsense syllables can be predicted as the product of the probabilities of identification of its phoneme parts, while word identification is systematically higher. Simulation studies are reported here involving a ‘‘factored’’ perceptual model that consists of a set of phoneme-likelihood estimators whose outputs are modulated by prior probabilities related to lexical status. This model can approximate the patterns observed in human perception. Simulations were also run with nonfactorable models, where syllables and words involve information about unique stimulus properties that cannot be predicted from their constituent phonemes. Consistent with a conjecture of Allen [J. Allen, IEEE Trans. Speech Audio Process. 2, 567–577 (1994)], such syllable template models do not produce behavior compatible with human performance. [Work supported by SSHRC.]

Speechreading sentences I: Development of a sequence comparator

Previous research suggests that many lexical errors in the speechreading of sentences can be explained in terms of visual phonemic errors. However, description and quantification of perceptual errors at the phonemic level requires specification of stimulus-to-response alignments. Because speechreading produces numerous errors, including phoneme insertions, deletions, and/or substitutions, alignment is a nontrivial problem. This paper describes development of a sequence comparator that can be used to obtain alignments automatically for phonemically transcribed sentences. The comparator employs a weights matrix that reflects presumed visual distances between all possible segmental stimulus-response pairs to find the alignment that minimizes overall stimulus-response distance. Initially, the comparator used weights based on viseme groupings, but these weights resulted in multiple, equal-distance, alternative alignments. More effective weights were obtained empirically via multidimensional scaling of phonemic confusions. Vowel data were obtained from Montgomery and Jackson [J. Acoust. Soc. Am. 73, 2134–2144 (1983)] and consonant data from a nonsense syllable identification task, which employed 22 consonants spoken by the same talkers who produced the sentence stimuli for this study [Bernstein et al., J. Acoust. Soc. Am. 85, 397–405 (1989)]. [Work supported by NIH.]

A frequency importance function for continuous discourse.

Normal hearing subjects estimated the intelligibility of continuous discourse (CD) passages spoken by three talkers (two male and one female) under 135 conditions of filtering and signal-to-noise ratio. The relationship between the intelligibility of CD and the articulation index (the transfer function) was different from any found in ANSI S3.5-1969. Also, the lower frequencies were found to be relatively more important for the intelligibility of CD than for identification of nonsense syllables and other types of speech for which data are available except for synthetic sentences [Speaks, J. Speech Hear. Res. 10, 289-298 (1967)]. The frequency which divides the auditory spectrum into two equally important halves (the crossover frequency) was found to be about 0.5 oct lower for the CD used in this study than the crossover frequency for male talkers of nonsense syllables found in ANSI S3.5-1969 and about 0.7 oct lower than the one for combined male and female talkers of nonsense syllables reported by French and Steinberg [J. Acoust. Soc. Am. 19, 90-119 (1947)].

Importance of various frequencies to perception of continuous discourse

Normal‐hearing subjects estimated intelligibility of continous discourse passages spoken by three talkers under various conditions of filtering and noise interference. The relationship between the intelligibility of continuous discourse and the articulation index (transfer function) was different from any found in ANSI S3.5‐1969. Also, the lower frequencies were found to be relatively more important for the intelligibility of continuous discourse than for identification of nonsense syllables and other types of speech for which data are available except for synthetic sentences.

Effects of intensity and filtering on identification of nonsense syllables

The identification of nonsense syllables was investigated under various conditions of filtering, noise, and intensity level. The syllables were filtered and recorded simultaneously as two separate bands in a background of speech spectrum noise. The low‐pass band (200–400 Hz) was recorded at a nominal signal to noise ratio of 20 dB, the high‐pass band (1–5 kHz) at 7 dB. The low‐pass band was presented at 60, 80, and 96 dB SPL, the high‐pass band at 60 and 80 dB SPL (nominal levels represent coupler level prior to filterings. Each band was presented alone and in every possible combination of intensity level with the other band to four normal hearing subjects. The best performance resulted when both bands were at 60 dB. Increasing the level of the high‐pass band by 20 dB and of the low‐pass band by 20 and 36 dB relative to the best condition resulted in a reduction in performance. Poorest performance resulted when the low‐pass band was at the highest level (96 dB) and the high‐pass band was at the lowest level (60 dB).

An auditory discrimination test battery

A battery of eight auditory discrimination tests was designed to measure the primary dimension along which simple and complex sounds can be discriminated. Six of the tests, selected on the basis of data collected in an earlier 22-subtest version of the battery [D. M. Johnson, J. K. Jensen, and C. S. Watson, J. Acoust. Soc. Am. Suppl. 1 67, S52 (1980)], use tonal stimuli: tests of frequency, intensity, and duration discrimination, using single tones, and tests of rhythm, temporal order, and tonal-pattern discrimination using multitone sequences. Two additional tests employ speech sounds: a syllable order discrimination test and a modification of one portion- of a nonsense syllable identification test [J. R. Dubno and H. Levitt, J. Acoust. Soc. Am. 69, 249–261 (1981)]. The tape-recorded tests were designed to be administered to groups of listeners in a free field, moderately low-reverberation environment. with responses recorded on printed forms. Seventy-two modified 2AFC trials, distributed among eight stimulus values, are presented in each of the discrimination tests. Performance of groups of listeners given successive versions of the test battery show it to be a reliable screening instrument for the measurement of auditory discrimination abilities. [Work supported by NIH/NINCDS.]