Mismatch negativity in a phonetic discrimination task.

The hypothesis was tested that in a discrimination task the number of phonetic features available determines the latencies and amplitudes of the components of the evoked potentials. Six conditions varying in the number and type of discriminatory cues were compared. The condition containing a three phonetic feature discrimination produced shorter latency evoked potentials as compared with the conditions containing one or two such features or an acoustic difference only. The N2 component, the so-called mismatch negativity, was particularly sensitive to the number of phonetic features.