Hundreds of thousands of people lose the ability to speak each year due to trauma, stroke, neuro-degeneration, and neoplasms with devastating consequences for social interaction and quality of life. Rehabilitation and treatment for these patients has been limited by fundamental deficiencies in the understanding of speech production in the brain ? specifically, how distributed cortical substrates coordinate to engender this sophisticated, fluent, and ubiquitous form of human communication. Discrete cognitive functions identified by the nature of common speech errors, chronometric studies of picture naming, and patterns of disruption in aphasia have informed hierarchical psycholinguistic models of neural processes leading from intention to articulation. Separately, motor control theorists have developed detailed computational models of articulation governed by a stable neural control system. Neurobiological instantiations of these two categories of speech production models both implicate a lateralized peri-sylvian network architecture, but no clear consensus has emerged on the dynamic behavior and functional organization of this network. This work leverages the unique and potent advantages of human invasive electrocorticography in a large cohort (n = 100) ? excellent spatiotemporal resolution, direct full spectrum recordings, and complete bilateral cortical coverage ? to study the rapid, transient and coordinated neural processes supporting speech production. First, this proposal will assess inter-regional timing and its dependence on specific response features (e.g. lexical frequency, phonological complexity). Second, dynamic causal modeling will be used to compare and optimize models of speech production as a function of parameters with direct neurophysiological significance. Third, direct closed-loop stimulation of anatomical targets will examine the response of the speech production network to external perturbation. These experiments will furnish an improved understanding of healthy language function and the breakdown of this faculty with damage to specific components of the subservient cortical network. Additionally, this work will provide the PI with the opportunity, training, and guidance to develop into a productive future clinician-scientist.
In this proposal, we will study a defining human capability ? speech. Despite general consensus on the abstract cognitive processes that must lead from the intention to speak to the articulation of a word, there remain open questions about the architecture and function of the broad network of brain regions that support speaking aloud. The development of treatment for the thousands with damaged language faculties will benefit from an improved understanding of speech production and its emergence from brain networks.