Neurophysiological basis of temporally cued phonetic contrasts in Japanese and American English listeners

Abstract

The present study examined American English (AE) listeners’ perception of Japanese (JP) temporally cued contrasts of vowels and consonants. The mismatch negativity (MMN) component of event-related potentials (ERPs) was used as a measure to compare passive (visual-attention condition) and active (auditory-attention condition) discrimination in an oddball paradigm. JP nonsense temporally cued vowel ([tado] vs [taado]) and consonant contrasts ([mi■i] vs [mi■■i]) were presented to JP listeners, for whom the temporal cues are contrastive, and to AE listeners who have no knowledge of JP. The following results were obtained. (1) With respect to the type of contrast, MMNs were larger for vowel contrasts than consonant contrasts. (2) In the auditory-attention condition both groups showed a robust MMN to vowels. (3) In the visual-attention condition JP also showed a robust MMN, while AE showed a smaller amplitude or absent MMN. This group difference was significant for the vowel contrast. (4) Interestingly, in the auditory-attention condition, AE showed a larger late negativity at frontal sites for vowel contrasts than JP. The results support the hypothesis that first-language perception is an overlearned process that may interfere with second-language speech perception. [Work supported by NSF.]