Cortical activity elicited by isolated vowels and diphthongs

Cortical activity underlying speech perception has been studied mostly by using isolated vowels with constant formant frequencies. Speech, however, is characterized by formant transitions whereby formant frequencies change as a function of time. We used magnetoencephalography (MEG) to investigate cortical activity elicited by isolated vowels and diphthongs containing formant transitions. Ten subjects were presented with two isolated vowels /a/ and /u/ and diphthongs /au/ and /ua/. Stimulus duration was 200 ms, and the diphthongs started and ended with a 50-ms constant-formant period and included a 100-ms linear transition period. Apart from studying the auditory N100m response, we examined subsequent brain activity in a 500-ms poststimulus time window, as the transitions were expected to elicit activity also in later stages of cognitive processing. All the stimuli elicited prominent N100m responses. Thereafter, both the isolated vowels and diphthongs elicited sustained brain activity lasting up to 500 ms. The present observations indicate that identification of the speech sounds as well as changes in their identity are reflected in the auditory N100m. Notably, the stimuli appeared to elicit left-hemispheric activity resembling the N400, typically obtained by using more complicated speech stimuli such as words and sentences.