Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Shared cortical language networks with convergent hierarchical network dynamics for lexicosemantic processing in comprehension and naming
Poster Session B, Friday, October 25, 10:00 - 11:30 am, Great Hall 3 and 4
Kathryn Snyder1, Kiefer Forseth1, Oscar Woolnough1, Elliot Murphy1, Greg Hickok2, Nitin Tandon1; 1UTHSC, 2UC Irvine
Lexicosemantic processing is a fundamental component of language that describes the mapping between concepts and sensorimotor forms such as phonology or orthography. Furthermore, prominent psycholinguistic models suggest that spoken and written language engage a shared lexicosemantic processing network in both speech perception and production, yet the convergent neural mechanisms are unclear. Here, we used electrocorticography (ECoG) to identify the spatiotemporal network dynamics of lexicosemantic processing in comprehension and naming. Data were obtained from 68 patients who underwent invasive electrocorticography and completed auditory (AN) and orthographic (ON) naming (8,664 electrodes). We analyzed gamma activity (65-115Hz) using a surface-based mixed-effects multilevel analyses to identify lexicosemantic processing networks during speech comprehension and naming. Next, we isolated discrete cognitive states and associated network dynamics using autoregressive hidden Markov models (ARHMM). Lastly, we used direct cortical stimulation (DCS) to attribute causality to critical nodes. At speech onset, activation of superior temporal gyrus (pSTG) was followed by superior temporal sulcus/middle temporal gyrus (pSTS/pMTG), and posterior middle frontal gyrus (pMFG). For each written word, visual cortex activity was followed by activation of lexical and phonological reading routes (middle fusiform gyrus, mFus; pSTG; pSTS/pMTG, intraparietal sulcus, IPS; pMFG; pars triangularis, pTr; pars opercularis, pOp). Both modalities engaged pSTS/pMTG and pMFG for comprehension, and activity was correlated with the number of sentential constituents available to be merged into a meaningful phrase (p<0.01) implicating both regions in linguistic composition. The last word activated a shared network (pSTS/pMTG; mFus; IPS; pTr) for naming. ARHMM isolated 5 states for AN and 6 for ON with 3 convergent states. The first convergent state occurring at stimulus offset was characterized by outflow from pSTS/pMTG, mFus, IPS, and pTr, and state duration was correlated with reaction time (p<0.001) implicating it in lexical access. Lastly, DCS of lateral temporal cortex (LTC) showed an anterior-to-posterior gradient of modality-based functional disruption with more anterior regions preferentially disrupting AN and more posterior regions preferentially disrupting ON. Regions that disrupted both AN and ON were located primarily within the planum temporal and pSTS/pMTG. At stimulus offset, DCS of pSTS/pMTG, mFus, IPS, and pTr disrupted both AN and ON. Our results show that naming to spoken and written descriptions recruit shared cortical language networks for both speech comprehension and lexical access. Additionally, we showed that neural states and associated network dynamics are convergent for lexical access. Lastly, we found that speech processing in LTC follows an anterior-to-posterior gradient of modality-based functional specialization with PT and pSTS supporting multimodal speech comprehension. Juxtaposing network dynamics of multimodal lexicosemantic processing in speech perception and production informs our understanding of specialized and shared language networks providing new insights to facilitate designs of neural prosthetics for language disorders.
Topic Areas: Meaning: Lexical Semantics, Language Production