The voice is a rich source of information, which the human brain has evolved to decode and interpret. Empirical observations have shown that the human auditory system is especially sensitive to the human voice, and that activity within the voice-sensitive regions of the primary and secondary auditory cortex is modulated by the emotional quality of the vocal signal, and may therefore subserve, with frontal regions, the cognitive ability to correctly identify the speaker's affective state. So far, the network involved in the processing of vocal affect has been mainly characterised at the cortical level. However, anatomical and functional evidence suggests that acoustic information relevant to the affective quality of the auditory signal might be processed prior to the auditory cortex.
Here we review the animal and human literature on the main subcortical structures along the auditory pathway, and propose a model whereby the distinction between different types of vocal affect in auditory communication begins at very early stages of auditory processing, and relies on the analysis of individual acoustic features of the sound signal. We further suggest that this early feature-based decoding occurs at a subcortical level along the ascending auditory pathway, and provides a preliminary coarse (but fast) characterisation of the affective quality of the auditory signal before the more refined (but slower) cortical processing is completed.