Aging and Temporal Influences on Speech Perception in Reverberation and Noise.

The present study examined the interactions of age, temporal resolution, hearing loss, and consonant perception under realistic listening conditions. Four subject groups were employed, with N = 8 per group: young adults with normal hearing or sloping sensorineural hearing loss; and elderly individuals with minimal peripheral hearing loss or presbycusis. Copies of the CUNY Nonsense Syllable Test (NST) were re-recorded under four levels of reverberation in quiet and in +10 signal-to-noise ratio of cafeteria noise. The test stimuli were presented binaurally to the subjects via TDH-49 headphones. In addition, a diotic wideband gap detection task, using a 72 dBSPL presentation level, was used as a measure of temporal resolution. Results of the present investigation demonstrated that all other subject groups performed significantly poorer than the young, normal hearing adults. Scores decreased as the amount of distortion increased, although there was very little difference between performance in the two highest reverberation conditions. For reverberation alone, the young, hearing-impaired listeners obtained the lowest scores; for reverberation + noise, the older, hearing-impaired subjects performed poorest. Large within-group variability was noted in the size of the gap thresholds, and only the two groups of young subjects differed significantly on this task of temporal resolution. Correlation analyses demonstrated a strong inverse relation between age and performance in reverberation + noise, even when degree of hearing loss was partialed out. Pure -tone average, as well as NST scores in quiet and in noise alone were related strongly to performance in reverberation. Multiple regression analyses demonstrated that nonsense syllable perception in reverberation could be predicted from a combination of age, pure-tone thresholds, scores in quiet and noise, and gap threshold. Gap threshold was the strongest predictor variable for reverberation alone, while the NST score obtained in noise was the most important variable for prediction in reverberation + noise conditions. The regression equations were less successful when the data was pooled across all elderly subjects, suggesting that other variables not measured in the present study enter into nonsense syllable perception by older individuals.