Locating Melody Processing Activity in Auditory Cortex with MEG

Roy D. Patterson, Martin Andermann, Stefan Uppenkamp, Andre Rupp

This paper describes a technique for isolating the brain activity associated with melodic pitch processing. The magnetoencephalograhic (MEG) response to a four note, diatonic melody built of French horn notes, is contrasted with the response to a control sequence containing four identical, “tonic” notes. The transient response (TR) to the first note of each bar is dominated by energy-onset activity; the melody processing is observed by contrasting the TRs to the remaining melodic and tonic notes of the bar (2-4). They have uniform shape within a tonic or melodic sequence which makes it possible to fit a 4-dipole model and show that there are two sources in each hemisphere – a melody source in the anterior part of Heschl's gyrus (HG) and an onset source about 10 mm posterior to it, in anterior planum temporale (PT). The N1m to the initial note has a short latency and the same magnitude for the tonic and the melodic sequences. The melody activity is distinguished by the relative sizes of the N1m and P2m components of the TRs to notes 2-4. In the anterior source a given note elicits a much larger N1m-P2m complex with a shorter latency when it is part of a melodic sequence. The TRs are accompanied by a sustained field (SF) originating from a source in HG near the melody source. However, the SF magnitude is of comparable size for the tonic and melodic bars. This study shows that melodies evoke activation in a separate, anterior part of auditory cortex.