Abstracts

Lexical semantic judgments are expected to depend on modality-specific, as well as amodal, lexical semantic processing. To verify the existence of these different components of lexical semantic processing, we used electrocorticographic (ECoG) recordings during a semantic relatedness task to measure event-related cortical activation with high spatial and temporal resolution. Having undergone surgical implantation of subdural electrode grids for the treatment of intractable epilepsy, six native English speakers (15-43 years old, mean 30 y.o.; 3 female, all left hemisphere dominant for language) judged whether sequentially presented pairs of words were semantically related or not, responding with a forced choice button press. Time-frequency analyses utilized matching pursuits to measure event-related ECoG spectral responses with respect to a 1-sec baseline preceding stimulus presentation. Word pairs were presented in either visual or auditory modalities during separate blocks. Event-related changes in frequency-specific signal energy were compared within subjects between conditions (visual/auditory; semantically related/non-related). Event-related increases and decreases in ECoG signal energy were observed in a variety of frequency ranges, including increases in high gamma ([gt]70 Hz) activity. Modality-specific event-related energy changes were observed in at least one recording site in all subjects. Modality-independent energy changes were observed in 5 of the 6 subjects. In 4 subjects event-related energy changes at some sites were significantly different for semantically related vs. unrelated word pairs. These findings were most prominent in left superior temporal gyrus, supramarginal gyrus, inferior frontal gyrus, and basal temporal cortex, as well as some of their right hemisphere counterparts. Lexical semantic processing occurs in both modality-specific, as well as amodal, cortical networks, distributed across cortical language regions in the dominant hemisphere, as well as in homologous regions of the nondominant hemisphere. (Supported by NINDS RO1-NS41598.)