hide
Free keywords:
-
Abstract:
A voice region has recently been identified in the monkey auditory cortex which prefers species-specific vocalizations and is sensitive to the vocal differences among monkey callers. To better understand the sensitivity of this and other brain regions for species-specific vocalizations, we independently manipulated two components in monkey vocal sounds. After recording the ‘coo’ vocalizations from 4 rhesus monkeys, we either shifted the position of the fundamental frequency (which is established by the vocal source in the larynx) or shaped the dispersion of the higher, formant frequencies (established by the acoustical filtering that occurs in the vocal tract above the larynx). The manipulations left the sounds within the species-typical range. We then used high-resolution functional imaging (fMRI) to evaluate the sensitivity of the rhesus monkey brain to independent changes of the two vocal components. Initial results revealed that many regions along the auditory-cortical processing hierarchy were sensitive to changes in both the fundamental and formant frequencies. In the earlier stages of cortical processing, especially in the caudal auditory regions, the sensitivity of the auditory fields to changes in the fundamental and formant frequencies depended upon each field’s topography of preferred sound frequency (tonotopy), which we mapped separately. For example, the sensitivity of the auditory field A1 to manipulation of the fundamental frequency occurred over this field’s low-frequency region, while the sensitivity to manipulation of the higher formant frequencies occurred over its high-frequency preferring region. The sensitivity of the anterior cortical regions to the components in the voice showed no such dependency upon their tonotopy. Notably, the monkey voice region, near to the temporal pole, showed better sensitivity to changes in the formant frequencies than to the fundamental. The results reveal a putative cortical network for vocal-sound processing and suggest that the voice region extracts information about speaker identity from cues present in the formant frequencies.
