Neural responses in human superior temporal cortex support coding of voice representations

Abstract

The ability to recognize abstract features of voice during auditory perception is an intricate feat of human audition. For the listener, this occurs in near-automatic fashion to seamlessly extract complex cues from a highly variable auditory signal. Voice perception depends on specialized regions of auditory cortex, including superior temporal gyrus (STG) and superior temporal sulcus (STS). However, the nature of voice encoding at the cortical level remains poorly understood. We leverage intracerebral recordings across human auditory cortex during presentation of voice and nonvoice acoustic stimuli to examine voice encoding at the cortical level in 8 patient-participants undergoing epilepsy surgery evaluation. We show that voice selectivity increases along the auditory hierarchy from supratemporal plane (STP) to the STG and STS. Results show accurate decoding of vocalizations from human auditory cortical activity even in the complete absence of linguistic content. These findings show an early, less-selective temporal window of neural activity in the STG and STS followed by a sustained, strongly voice-selective window. Encoding models demonstrate divergence in the encoding of acoustic features along the auditory hierarchy, wherein STG/STS responses are best explained by voice category and acoustics, as opposed to acoustic features of voice stimuli alone. This is in contrast to neural activity recorded from STP, in which responses were accounted for by acoustic features. These findings support a model of voice perception that engages categorical encoding mechanisms within STG and STS to facilitate feature extraction.

Functional connectivity across the human subcortical auditory system using an autoregressive matrix-Gaussian copula graphical model approach with partial correlations

“…Primary auditory cortex receives direct input from lemniscal MGB and is the last auditory structure with fine-grained tonotopicity (Pickles, 2015). In humans, STG is hierarchically structured, with portions further away from primary auditory cortex having increasingly wider temporal integration windows (Hamilton et al ., 2018; Norman-Haignere et al ., 2022) and greater categorical specificity (Bhaya-Grossman and Chang, 2022; Feng et al ., 2021; Hamilton et al ., 2020; Keshishian et al ., 2023; Nourski et al ., 2018; Pernet et al ., 2015; Rauschecker and Tian, 2000; Rupp et al ., 2022). While invasive recordings from human STG suggest a potential direct connection between MGB and posterior STG (Hamilton et al ., 2021), we found mixed evidence for such a direct pathway in ourOur partial correlation data (in one hemisphere in only one of the data partitions).…”

The path of voices in our brain

“…Interestingly, while voice patches are observed bilaterally in the auditory associative areas [1,2], processing of familiar voice-identity recognition is largely a right-lateralized process [10]. This distinction is also observed in other cognitive domains, such as speech and melodies.…”

“…A new publication by Rupp and colleagues in PLOS Biology [1] capitalises on human intracerebral recordings of individuals with epilepsy implanted for clinical purposes to further examine how voices are categorised by the human brain. Voices constitute a distinctive auditory category that selectively activates specific "voice patches" in bilateral associative auditory cortex: the "temporal voice areas" (TVAs; see Fig 1; [2]).…”

The Temporal Voice Areas are not “just” Speech Areas

“…Yet, other results indicate that even non-speech vocal sounds induce greater TVA activity than control sounds (Bodin et al's, 2021 ) or lead to above chance classification into vocal/non-vocal categories (Rupp et al, 2022 ), suggesting a selectivity to this category of sounds in the TVAs.…”