The goal of this study was to analyze the time-course of sensory (bottom-up) and cognitive (top-down) processes that govern musical harmonic expectancy. Eight-chord sequences were presented to 12 musicians and 12 nonmusicians. Expectations for the last chord were manipulated both at the sensory level (i.e., the last chord was sensory consonant or dissonant) and at the cognitive level (the harmonic function of the target was varied by manipulating the harmonic context built up by the first six chords of the sequence). Changes in the harmonic function of the target chord mainly modulate the amplitude of a positive component peaking around 300 msec (P3) after target onset, reflecting top-down influences on the perceptual stages of processing. In contrast, changes in the acoustic structure of the target chord (sensory consonance) mainly modulate the amplitude of a late positive component that develops between 300 and 800 msec after target onset. Most importantly, the effects of sensory consonance and harmonic context on the event-related brain potentials associated with the target chords were found to be independent, thus suggesting that two separate processors contribute to the building up of musical expectancy.

THE AIM OF THIS SERIES OF experiments was to determine whether consonant and dissonant chords elicit similar or different electrophysiological effects out of a musical context and whether these effects are similar or different for musicians and nonmusicians. To this end, w e recorded t he e vent-related b rain potentials (ERPs) elicited by the different intervals of the chromatic scale that were classified into three categories: perfect consonances, imperfect consonances, and dissonances. Participants were to decide, on a six-point scale, whether the intervals evoked pleasant or unpleasant feelings. To test the hypothesis that the perception of dissonance results from the superposition of the partials of close frequencies (Helmholtz, 1877), two notes were either played together (harmonic intervals) or successively (melodic intervals). Since, in this latter case, the two notes are played at different points in time, the perception of roughness, if any, should be weaker than for harmonic intervals. In line with Helmholtz's hypothesis, results showed larger differences for harmonic than for melodic intervals, which were mainly found on the N1-P2 complex for musicians, on the N2 component for nonmusicians, and on a later negative component for both musicians and nonmusicians. However, these results also point to the influence of expertise and cultural factors, since different results were obtained when ERPs were averaged as a function of music theory and according to the participants' responses.

No abstract