Mismatch negativity and P300 evoked response potentials: Perception and discrimination of /r/ and /I/ in Korean and English listeners

This study used behavioral measures in conjunction with electrophysiological measures to study perception and discrimination of /r/ and /l/ in native American-English and Korean listeners. MMN and P300 evoked response potentials (ERP) were obtained from 13 American-English subjects and 19 Korean subjects using audiotaped stimuli (/ra/ and /la/ syllables). All subjects, English and Korean, were given two discrimination tasks (AB and ABX paradigm) along with four taped conditions. The Korean subjects were divided into "Proficient" and "Non-Proficient" speakers of English based on their ABX scores.;It was hypothesized that the P300s elicited by tonal stimuli and the MMNs elicited by the syllables would reveal no differences in the latencies and amplitudes among the three groups. The P300s elicited by the syllables on the fourth tape were, however, expected to reveal prolonged latencies and reduced amplitudes in the Non-Proficient Korean group, due to their difficulties in perceiving /r/ and /I/.;Statistical analyses using MANOVA, ANOVA, Pearson and Spearman Correlation procedures revealed (1) no significant differences in the latencies and amplitudes of the P300s generated by tonal stimuli among the three groups, and (2) no significant differences in the latencies and amplitudes of the MMNs generated by the syllables among the three groups.;Although graphed data appeared to reveal prolonged latencies in the P300s elicited by the syllables for the Non-Proficient Koreans, a one-way MANOVA indicated that there were no statistically significant differences in the latencies or amplitudes among the three groups. Because of this finding, no relationship between the behavioral measures and the electrophysiological measures was found to exist.;These findings could be a function of the relative English proficiency of these Korean subjects (too homogeneous), the result of the subjects making discriminations on an acoustic rather than linguistic basis, and/or the result of a large degree of variability found in the groups' mean latencies and amplitudes.