Speech sounds are the most important sounds that humans hear, yet our knowledge of how the complex
cortical network of auditory and auditory-related brain regions subserves speech perception is limited. Our
overarching goal is to understand where and how speech information is represented within this network. The
focus of this proposal is to test directly predictive coding models of speech perception. In predictive coding
hypotheses, the brain minimizes the differences between internally generated expectations and sensory
observations through an ongoing interaction between sensory and higher-order cortical regions.
We use novel combinations of complementary invasive and non-invasive experimental methods to study
these brain regions in neurosurgery patients who require placement of chronic intracranial electrodes. These
experiments involve combining direct cortical electrophysiological recording methods with non-invasive
electroencephalography, electrical stimulation techniques and anatomical and functional neuroimaging
methods. We also obtain causal evidence of network functional properties by studying the effects of selectively
disrupting the function of specific components of this network. Our investigative strategy makes use of these
unique experimental opportunities to overcome long-standing barriers to progress in this research field.
We will pursue our goals by testing hypotheses regarding: (1) the locations, functional properties and
connectivity patterns of auditory cortical fields and auditory-related cortices of the temporal and frontal lobes
that are engaged in speech processing, (2) predictive coding models of speech signaling within this network,
and (3) the behavioral consequences and neural signatures of selective disruption of the network.
These objectives are pursued by an experienced multidisciplinary group of investigators with expertise
encompassing all required clinical and research topic areas. To our knowledge, the resulting data will be the
first of its kind to directly demonstrate how speech information is processed at all levels of auditory cortical
hierarchy and provide causal evidence of the role of feedback projections on speech information processing
within canonical auditory cortex. Detailed characterization of this network will improve our understanding of the
pathophysiology of disease states affecting this system and will provide mechanistic insights that are required
to inform the design of new treatment strategies.