Abstract:

The human voice is the most familiar and important sound in the human auditory environment, conveying large amounts of socially relevant information. Similar to face processing, there is also a functional specialization in brain for voice processing. Neuroimaging and electrophysiology studies have demonstrated that the temporal voice areas (TVAs) showed specific response to human voices. In addition, researchers have also observed the homologues of TVAs in non-human brain. Human voices can convey speech, affective and identity information, which are extracted and further processed in three interacting but partially dissociated neural pathways. To explicate these three functional pathways, researchers have proposed three corresponding models including the dual-stream model of speech processing, multi-stage model of vocal emotional processing and integrative model of voice-identity processing. In the future, researchers should further investigate whether voice-selective activity can be explained by the selective processing of specific acoustic features of voice and focus on neural mechanisms of voice processing in special populations (e.g. schizophrenia and autism).

Key words: voice processing, specialization, the temporal voice areas (TVA), speech processing, emotional prosody, voice-identity recognition