Sensorimotor integration during human self-vocalization: Insights from invasive electrophysiology

Effective human speech requires the neural integration of ongoing vocal production with the auditory and somatosensory feedback signals that are produced. We are using invasive electrophysiology techniques in patient volunteers undergoing neurosurgical treatment in order to gain insights into these mechanisms and underlying neural circuits. By using multi-contact electrode arrays chronically implanted over the perisylvian temporal lobe auditory cortex (e.g. area PLST) and the inferior frontal gyrus (IFG), we can examine local field potentials and frequency-specific responses from cortical areas important for both vocal production and speech sound processing. Our initial studies have found that during self-vocalization, focal areas within higher order auditory cortex on the superior temporal gyrus (STG) show response modulation compared to the responses of the same areas during passive listening. Manipulation of the auditory feedback that a speaker receives during vocalization (e.g. pitch-shifted or delayed auditory feedback) leads to further modulation of these PLST sites. Measures of functional connectivity including electrical stimulation tract tracing or phase-synchrony analysis demonstrate that portions of PLST are functionally connected to regions of the IFG. These findings support forward models for vocal control in which efference copies of premotor cortex activity modulate sub-regions of auditory cortex.