To elucidate the central mechanisms of sound segregation, we compared responses to a harmonic sound and a mistuned sound using a whole-head magnetoencephalography system. The harmonic sound was composed of a 200-Hz tone and its 2nd to 12th harmonics. The mistuned sound had, instead of the 600-Hz harmonic, a 696-Hz tone. In the right hemisphere, the amplitude of N100m responses evoked by the mistuned sound was significantly larger and the peak latency significantly longer than that evoked by the harmonic sound, suggesting that the right hemisphere plays a more important role than the left in detecting mistuned partials.