Auditory category encoding of human voice is distributed across superior temporal gyrus but converges with critical speech perception networks in posterior regions.

The posterior region of the human superior temporal gyrus and sulcus (pSTG/S) has been observed to play a critical role in a variety of neural functions. The mechanisms that underlie this functional heterogeneity and the extent to which functional subregions overlap are unknown. Leveraging the results of clinically-guided direct cortical stimulation and intracranial recordings in five patients undergoing invasive monitoring for epilepsy surgery, we identify subregions of functional overlap and divergence for auditory category encoding and speech production. Our results showed auditory category encoding sites (VOICE+) distributed across STG/S and primary auditory cortex, while sites that produced disruption to speech production on stimulation (STIM+) were located more posteriorly on STG/S, extending into the parietal lobe. We identified sites common to both functions (OVERLAP) in 4/5 participants. Among channels exhibiting functional overlap, 84.6% were located in the STG/S and, crucially, the majority were localized to the pSTG/S. A Bayes factor analysis revealed evidence in favor of similar auditory category sensitivity between VOICE+ and OVERLAP channels in 2/4 participants. Our findings highlight the particularly flexible nature of the human pSTG/S and provide support for hypotheses that a shared mechanism of dynamic encoding of acoustic information may give rise to the range of functions attributed to this region.