Neural Mechanisms of Language Perception in Human Intracranial Neurophysiology

Language has been the subject of academic fascination for centuries, and the ability to communicate abstract notions through speech and writing allows humans to interact in ways that would not otherwise be possible. While the mechanisms of language processing have been studied extensively with behavioral and noninvasive neuroimaging methods, much about how the brain encodes language remains unknown. In this dissertation, I describe experiments using intracranial neurophysiology in humans to interrogate the mechanisms of language perception at high spatiotemporal resolution. First, I explore the neural mechanisms of visual word recognition in a large human intracranial dataset. By analyzing population sensitivity to a hierarchy of word features, I create a high-resolution map of stimulus encoding during single-word reading that reveals the early influence of lexical features in lingual and fusiform gyri followed by a cascade of lexical, orthographic, and semantic information in temporal and frontal lobes. Along with clustering analyses that show stimulus encoding in anatomically distributed populations, these results demonstrate that feed-forward, feed-back, and distributed processing mechanisms underlie visual word recognition. Second, I describe the development of an artificial language task designed to characterize the neural mechanisms of auditory word segmentation. The task is designed in three phases to probe how the brain tracks distributional regularity and the neural mechanisms of word segmentation with and without lexical access. Taken together, this work expands our understanding of the neural mechanisms of language processing using human intracranial neurophysiology.