Detecting violation in abstract pitch patterns with mismatch negativity.

The human brain automatically extracts regularities embedded in environmental auditory events. This study investigated the extraction of abstract patterns by measuring mismatch negativity (MMN). Participants watched a silent subtitled movie and ignored a sequence of auditory events comprising frequent standards and rare deviants presented in the background. Tone triplets with varying pitch (first-order property) served as the auditory events. The pitch intervals (interval 1 and interval 2) between the tones in a triplet and the ratio of interval 1 and 2 were considered second- and third-order properties, respectively. Both second- and third-order properties of the standards were kept constant in the mixed patterns block, while only the third-order property was kept constant in the ratio pattern block. Four sets of tone triplets violating the interval and ratio patterns with different deviance levels were presented as deviants in both blocks, and subtracted with physically identical stimuli in a control block to isolate the MMNs. Interval and ratio pattern deviants elicited MMNs in the mixed patterns block while only ratio pattern deviants elicited MMNs in the ratio pattern block. Larger MMNs were elicited by large deviants as compared to small deviants. These results suggest that the change detection system is sensitive to the violation of both second- and third-order abstract patterns. In addition to regularities in the abstract properties of auditory events, regularities in the relationships between abstract properties can also be extracted. This ability plays an important role in music and language perception.