Syntactic control of lexical activation during speech production - Resubmission - 1

Project Summary/Abstract Language is a defining feature of being human. No other animal has the ability to systematically order words or signs so as to communicate complex thoughts, nor does such complex communication play nearly as central a role in other species. According to the NIDCD, roughly 1 million people in the United States have some form of

aphasia, or loss of linguistic ability. The consequences are devastating for individuals and families. The power of language derives from a unique evolutionary development: syntax, the abstract rules that allow us to combine finite words into infinitely many possible strings. Without syntax, communication is painstaking and

error-prone. Despite its centrality, little remains known about its neural instantiation. Previous research has attempted to experimentally isolate syntax so as to localize it in the brain, but this approach has produced mixed and inconsistent results. We take a novel approach, using electrocorticography (ECoG) to investigate a well-

defined component of syntax: sequencing, or selecting words in an order determined by syntax. A main impediment to progress in understanding syntax at the neural level is a granularity mismatch problem. Neuroscience tends to deal in broader conceptual distinctions, as evidenced by experimental manipulations that

attempt to isolate all of syntax. However, psycholinguistic models tend to be much finer grained, portraying syntactic abilities as a conglomerate of processes and representations rather than a monolithic entity that can be turned on or off. Here, we leverage computationally explicit models to make linking hypotheses between

cognitive processes and commensurate neural mechanisms. For instance, sequencing critically relies on a binding mechanism for linking particular words to positions in a syntactic tree. The central hypothesis of this proposal is that binding between syntax and words is instantiated as neural synchrony between regions encoding

syntax and those encoding words. Our approach has a number of strengths. (1) As ECoG is largely insensitive to production artifacts, we will employ a production paradigm, allowing us to more clearly isolate particular syntactic processes than traditional comprehension studies. (2) ECoG has unparalleled spatial and temporal resolution. (3) Our research draws not

only from findings in neuroscience, but from computationally explicit psycholinguistic models, allowing us to test hypotheses that have a high probability of being at least partially correct. This project leverages the applicant’s expertise in linguistics and cognitive psychology, and complements this

background with additional training in neuroscience and signal processing. The applicant will be embedded in a world-class research and clinical environment, with dedicated mentors who are leading experts in ECoG and the neurobiology of language. The findings will deepen our understanding of normally functioning communication,

in line with the NIDCD’s first priority area, by shedding light on the mechanisms affected in word order deficits, which are common in a range of disorders including aphasia and specific language impairment.