Binaural Fusion of Virtual Pitch in the Human Auditory Cortex

Abstract

The vowels of human speech and the sounds produced by many musical instruments are harmonic complex tones. These are comprised of pure tones that are integer multiples of a fundamental frequency. The perceived pitch of complex tonal stimuli is qualitatively different from the one of pure tones. The pitch of pure tones is a basic auditory percept which is monotonically related to its frequency, a phenomenon known as spectral pitch. The pitch of complex tones is termed virtual pitch. It can be assessed by auditory matching to pure tones. The virtual pitch of harmonic complex tones corresponds to a frequency approximating the frequency difference between the components, i.e. to the fundamental frequency, even if the fundamental frequency is missing. The fact that the first six to eight harmonics of a complex tone can be perceived as separate spectral pitches led Terhardt [1] to conclude that the virtual pitch is extracted by a type of “Gestalt” recognition, i.e. virtual pitch is an auditory attribute that is established on a higher level of processing than spectral pitch. The distinction between spectral pitch and virtual pitch can be likened to the distinction between primary visual contours and virtual (“illusory”) contours in vision [2]. Just as a visual virtual contour can be regarded as being “inferred” from a set of primary contours, virtual pitch is regarded as the result of a two-stage process which involves peripheral frequency analysis and virtual-pitch formation by a pertinent “pitch processor”[3]. It is believed that the pitch processor is located in more central regions of the auditory system. An important feature is binaural fusion, i.e. the ability to combine spectral components regardless of the ear to which they are presented. It was therefore hypothesized that neural activity ascending from both ears would fuse into one percept [4].