Presenter Name:Joshua Hoddinott

School/Affiliation:University of Western Ontario

Co-Authors:Jessica Grahn

Abstract:

Humans spontaneously perceive an underlying pulse, or “beat,” arising from the rhythmic structure in music. Functional magnetic resonance imaging (fMRI) studies show that motor regions of the brain, such as the basal ganglia, supplementary motor area, and premotor cortices have increased activity when people listen to rhythms with a strong beat (Grahn & Brett, 2007). However, previous fMRI studies have generally used univariate analyses, which investigate activity averaged over voxels in a region, whereas multivariate techniques can identify patterns of covariance across voxels that allow greater sensitivity and better identification of stimulus features that predict brain–behavior relationships. Thus, here we use multivariate pattern analysis to compare neural activity patterns elicited by rhythms with strong, weak, or no beat. The results enable us to determine which neural regions are sensitive to beat strength and will allow us to characterize how these regions respond to varying degrees of beat strength. One possibility is that motor areas ‘tune’ activity patterns to each strong-beat rhythm, exhibited by high dissimilarity between activity patterns elicited by individual strong-beat rhythms, and low dissimilarity between patterns elicited by individual weak- and non-beat rhythms; alternatively, motor areas may activate in highly-correlated patterns for strong-beat rhythms, but may exhibit highly-dissimilar activity patterns for rhythms with weak or no beat. Preliminary results reveal that multivariate patterns do indeed appear sensitive to individual rhythms, but may also distinguish between rhythm types, building on previous univariate work.