Changes in the Perception of Stop Consonants Through Enhanced Cue Training as Reflected by Categorical Boundaries and Late Auditory Evoked Potentials

Hearing-impaired listeners have difficulty in discriminating between voiced stop consonants. An important acoustic cue in this discrimination is the transition from the frequency of the consonant to the frequency of the vowel. The main purpose of this study was to investigate the effect of auditory training on the perception of the formant transition cue in the discrimination of the place of articulation of voiced stop consonants in synthetic CV stimuli of hearing-impaired listeners. Changes in perception were represented by behavioral and electrophysiological measures. Generalization effects after training and correlations between behavioral and electrophysiological measures were also measured.

Eight male and three female hearing-impaired subjects (23 - 72 years of age) participated in this study. The behavioral measures involved the participants performing categorization of stimuli as either /b/ and /d/ in a 2AFC task. Stimuli varied in onset frequency of the second formant. Electrophysiological measures, including P100, N100, P200, N200 latencies and P100-N100, N100-P200, P200-N200 peak-to-peak amplitudes, were recorded as participants sat quietly while watching a video in silence as stimuli was presented in one ear.

Subjects participated in auditory training on the formant transition cue over the period of four days. Stimuli used in auditory training were consistent with stimuli used in behavioral and electrophysiological measures, except that the formant transition cue was enhanced during training. This enhancement resulted from amplification of the transition portion of the stimulus to ensure audibility of the cue for the listener. Behavioral measures involved identifying the category (/b/ or /d/) to which the stimulus belonged. Visual feedback was given after each correct response.

Results were consistent with previous studies. Behavioral measures indicated improvement from before training to after training in the perception of the formant transition cue by some listeners. However, analysis of variance (ANOVA) and correlation procedures failed to reflect any pre versus post training change. Age and severity of hearing were not correlated with improvement due to training. However, ability to categorize prior to training correlated with improvement after training. Statistical results for electrophysiological measures did not find significant changes between pre and post training.