Intracranial Electrophysiology & Anatomical Connectivity of Voice-Selective Auditory Cortex

ABTRACT: The ability to recognize voice is an intricate feat of human audition. For the listener, the brain is able to seamlessly extract complex linguistic and non-linguistic cues from highly variable vocal acoustic input. Neuroimaging studies have proposed specialized regions of auditory cortex dedicated to voice perception, including superior temporal gyrus (STG) and superior temporal sulcus (STS), referred to as “temporal voice areas”. Functional neuroimaging studies also demonstrate these areas respond most strongly to vocalizations of the same-species compared to other primate vocalizations and natural sounds, further suggesting specialization of auditory cortex for vocal acoustic stimuli. It remains unknown if these regions demonstrate true selectivity for voice, or more generally function to process the spectrotemporal features of complex auditory stimuli, such as voice. The voice perception network has been partially described by neuroimaging studies and suggests temporal voice areas exhibit connectivity to inferior frontal gyrus and precentral gyrus, however these studies are limited in their ability to characterize voice areas at physiologic timescales and have largely focused on characterizing frontotemporal white matter pathways underlying speech perception and production. The proposed research aims to characterize local electrophysiologic responses to voice in temporal voice areas and will describe the frontotemporal structural connectivity of the voice perception network. I will leverage intracranial electroencephalography (iEEG) from neural populations across human auditory cortex in 15 patient-participants undergoing epilepsy surgery evaluation to examine the neural representation of voice. Neural recordings will be acquired while participants listen to a published Voice Localizer stimulus set optimized for iEEG research, as well as an engineered acoustic stimulus set from modulated noise that mimick the spectrotemporal features of voice and other natural sounds, called Gaussian Sound Patterns (GSPs). Frontotemporal connectivity of voice-selective auditory cortex will be examined across patients using clinically-acquired diffusion tensor imaging (DTI) in all patients with Voice Localizer recruited to date (n=11) and included in this proposal (n=15). Connectivity analyses will reveal regions of frontal cortex demonstrating connectivity to neuronal populations along STG and STS with the greatest voice-selective responses. Together this proposal will leverage a multimodal dataset that marries local cortical iEEG recordings at physiologic timescales and DTI structural connectivity analysis to critically examine voice selective auditory cortex.