Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Low-frequency power decreases in temporal lobe network underpin verbal fluency performance: Evidence from intracranial EEG
Poster E11 in Poster Session E, Thursday, October 26, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Britta U. Westner1, Shawniqua Williams Roberson2, Timothy H. Lucas II3, Brian Litt4, Vitória Piai1,5; 1Donders Centre for Cognition, Radboud University, Nijmegen, The Netherlands, 2Vanderbilt University Medical Center, Vanderbilt University, USA, 3Ohio State University, Neurotech Institute, USA, 4University of Pennsylvania, Center for Neuroengineering and Therapeutics, USA, 5Radboud University Medical Center, Donders Centre for Medical Neuroscience, Nijmegen, The Netherlands
Verbal fluency assessments are common in clinical testing of cognitive function. However, the neural underpinnings of verbal fluency are not fully understood and not widely researched. In this work, we examine the oscillatory dynamics related to word production in a verbal fluency paradigm. We re-analyzed intracranial EEG data from six epilepsy patients performing a verbal fluency task with both semantic cues (category, e.g. “animals” or “fruit”) and phonological cues (letter, e.g. “A” or “F”). Previously, this dataset has been analyzed with respect to high-frequency activity (Willams Roberson et al. 2020). In the current analysis we computed theta (4-7 Hz), alpha (8-12 Hz), and beta (13-30 Hz) power changes in an early time window (-750 to -250 ms) and in a peri-utterance time window (-250 ms to 250 ms) relative to speech onset. We included nine regions of interest in the left-hemispheric frontal, temporal (including the basal temporal language area), and occipito-parietal areas (246 contacts in total). Statistical testing against resting state activity was done at the individual participant level, using a permutation approach with a Max-T sum statistic. To derive an index of generalizability, we subjected the within-subject statistical results to a Bayesian population prevalence analysis. We observed high population prevalence for a decrease in low frequency activity in temporal regions for the early time window, with a frequency gradient from superior temporal gyrus (predominantly theta activity) to middle (predominantly theta and alpha activity) and inferior temporal gyrus (predominantly alpha and beta activity). The peri-utterance window was marked by a theta power decrease in the temporal regions, an alpha-beta decrease in inferior frontal gyrus, and theta-alpha decreases in occipito-parietal regions. An exploratory connectivity analysis of two data sets with sufficient coverage revealed a circumscribed network for one dataset (whereas no task-related connectivity at all was evident for the other data set). The temporal lobe network included parahippocampal regions and was driven by low frequencies. For the peri-utterance window, this network extended to frontal as well as occipital regions. Taken together, these results sketch a network possibly related to verbal fluency, with a high involvement of temporal-lobe regions. These findings are concordant with studies on word production using other tasks (e.g., visual and auditory naming or context-driven word production tasks), which suggest temporal regions play a role in lexical access and semantic processing. Interestingly, although the verbal fluency task did not include any visual stimulation, we find occipital involvement in both the low frequency power decreases as well as in the exploratory connectivity analysis. Whether this marks the extension of the semantic network to occipital regions or the recruitment of visual cortices for semantic processes will be an exciting topic for future research.
Topic Areas: Language Production,