Speech Perception Problems Research Articles

Speech perception problems of the hearing impaired reflect inability to use temporal fine structure.

People with sensorineural hearing loss have difficulty understanding speech, especially when background sounds are present. A reduction in the ability to resolve the frequency components of complex sounds is one factor contributing to this difficulty. Here, we show that a reduced ability to process the temporal fine structure of sounds plays an important role. Speech sounds were processed by filtering them into 16 adjacent frequency bands. The signal in each band was processed by using the Hilbert transform so as to preserve either the envelope (E, the relatively slow variations in amplitude over time) or the temporal fine structure (TFS, the rapid oscillations with rate close to the center frequency of the band). The band signals were then recombined and the stimuli were presented to subjects for identification. After training, normal-hearing subjects scored perfectly with unprocessed speech, and were approximately 90% correct with E and TFS speech. Both young and elderly subjects with moderate flat hearing loss performed almost as well as normal with unprocessed and E speech but performed very poorly with TFS speech, indicating a greatly reduced ability to use TFS. For the younger hearing-impaired group, TFS scores were highly correlated with the ability to take advantage of temporal dips in a background noise when identifying unprocessed speech. The results suggest that the ability to use TFS may be critical for "listening in the background dips." TFS stimuli may be useful in evaluating impaired hearing and in guiding the design of hearing aids and cochlear implants.

Speech perception in preschoolers at family risk for dyslexia: Relations with low-level auditory processing and phonological ability

We tested categorical perception and speech-in-noise perception in a group of five-year-old preschool children genetically at risk for dyslexia, compared to a group of well-matched control children and a group of adults. Both groups of children differed significantly from the adults on all speech measures. Comparing both child groups, the risk group presented a slight but significant deficit in speech-in-noise perception, particularly in the most difficult listening condition. For categorical perception a marginally significant deficit was observed on the discrimination task but not on the identification task. Speech parameters were significantly related to phonological awareness and low-level auditory measures. Results are discussed within the framework of a causal model where low-level auditory problems are hypothesized to result in subtle speech perception problems that might interfere with the development of phonology and reading and spelling ability.

Application of a central auditory test battery in normal hearing adults with auditory problems

Objective. To study the validity of a recently developed test battery for normal hearing adults with central auditory processing problems. Materials and methods. The test battery comprised a questionnaire 1 and seven central auditory tests. 2 This test battery was administered in 10 normal hearing adults who visited our clinic because of speech perception problems. Results. The test scores of the subjects was compared with those of a reference group (n = 28). On average, test scores were significantly lower for the speech-in-noise test, filtered speech test and backward masking test. In addition, individual scores were examined, based on percentile scores. A significant relationship was found between the subjects complaints (as quantified by the questionnaire) and test scores. Conclusion. First results suggest that the test battery is valid.

Can macaques perceive place of articulation from formant transition information?

An important problem in speech perception is to determine how humans extract the perceptually invariant place of articulation information in the speech wave across variable acoustic contexts. Although analyses have been developed that attempted to classify the voiced stops /b/ versus /d/ from stimulus onset information, most of the human perceptual research to date suggests that formant transition information is more important than onset information. The purpose of the present study was to determine if animal subjects, specifically Japanese macaque monkeys, are capable of categorizing /b/ versus /d/ in synthesized consonant-vowel (CV) syllables using only formant transition information. Three monkeys were trained to differentiate CV syllables with a "go-left" versus a "go-right" label. All monkeys first learned to differentiate a /za/ versus /da/ manner contrast and easily transferred to three new vowel contexts /[symbol: see text], epsilon, I/. Next, two of the three monkeys learned to differentiate a /ba/ versus /da/ stop place contrast, but were unable to transfer it to the different vowel contexts. These results suggest that animals may not use the same mechanisms as humans do for classifying place contrasts, and call for further investigation of animal perception of formant transition information versus stimulus onset information in place contrasts.

Feature extraction and feature interaction

The idea of the orderly output constraint is compared with recent findings about the representation of vowels in the auditory cortex of an animal model for human speech sound processing (Ohl & Scheich 1997). The comparison allows a critical consideration of the idea of neuronal “feature extractors,” which is of relevance to the noninvariance problem in speech perception.

Chairs’ introduction to special session on phonetic perception and word recognition

A classic statement of the speech perception problem is how to account for the perception of phonemes based on the acoustic stimulus. A classic statement of the word recognition problem is how to account for the access and selection of lexical representations, with little concern for phoneme/phonetic perception. Over the past two decades, research in spoken-word recognition has begun to explicitly model the stages of word recognition in terms of phonetic/phonemic perception and later stages involving lexical knowledge. These different models incorporate very specific proposals for the interaction of information from the bottom-up stimulus and the lexicon. They also raise issues concerning the appropriate model for phoneme perception, that is, for example, does it precede or follow the lexical processing stage? It would be convenient if theories/models of phoneme perception could ignore effects of the lexicon, and therefore, autonomously account for the ‘‘front-end’’ of word recognition. But can they, and be adequate theories/models of either phoneme perception or the front-end to word recognition? Alternatively, are there phoneme identification effects that are irrelevant to word recognition? In this session, researchers representing a variety of perspectives on the true relationship between lexical processes and phoneme/phonetic perception are brought together.

Rehabilitation of Hearing-Impaired Adults

Hearing loss produces a communication handicap which is not always predictable from audiometric results. Handicap questionnaires can be useful for identification of potential rehabilitation candidates. Hearing aids have been improved but cannot be expected to compensate for all the speech perception problems associated with sensorineural hearing loss. Other communication aids can be used to reduce hearing handicap. Effective use of devices often requires more instruction than is currently available. A comprehensive rehabilitation program takes into account individual circumstances and provides for a variety of services in addition to provision of aids.

Auditory processing in aging listeners.

Many aspects of auditory processing in aging listeners are poorly understood. Methodological problems encountered with the aging population, ranging from ear canal collapse to a conservative criterion, have rarely been considered in past research. The relation between pure-tone thresholds and other aspects of auditory functions is unclear. There have been few studies of frequency and temporal analysis in aging listeners. There is disagreement on the effect of age on speech discrimination abilities in various listening conditions, and the cause of speech perception problems in aging listeners has not been adequately investigated. Areas of importance for clinical consideration and auditory research with aging listeners are discussed.

Perceptual processing of multiple cues and contexts: effects of following vowel upon stop consonant voicing

Variations in vowel context affect values of the acoustical variables that function as perceptual cues for the voicing distinction in initial stop consonants in English. Processing these context-dependent variations is a basic problem in speech perception. Of the several acoustical correlates of voicing that may be considered cues, two major ones are examined here, previously conflated in the Voice Onset Time concept: the first is the time from the burst to the first voicing pulse and is measured as the separation of these two events; the second is the relative position in the formant transition frame at which voicing onsets and is measured as the amount of voiced first formant (F1) transition. An experiment is reported which shows that for velar stops before /a/ both cues play a role in voicing perception although the transition cue is dominant. For velar stops before /i/, where no F1 transition occurs, the separation cue appears more important than it does in /a/ context. Consideration of these results points to the need to distinguish between contextual adjustments involving a simple boundary shift and those involving a reweighting of cues according to context. The two cues used here appear to be of different types with attendant differences in the effects they generate in perception experiments. In particular a sophisticated view of the processing of different cues may be necessary to give an account of adaptation in the voicing feature in experiments where the percentage of voiced or unvoiced responses is modified following repeated exposure to a stimulus with a distinct value of the feature.