Authors :
Presenting Author: Masaki Sonoda, MD, PhD – Yokohama City University Graduate School of Medicine
Ethan Firestone, MD Candidate – Children’s Hospital of Michigan, Detroit Medical Center, Wayne State University
Keita Fujii, MD – Yokohama City University Graduate School of Medicine
Naoto Kuroda, MD, PhD – Wayne State University
Kazuki Sakakura, MD, PhD – Rush University
Tetsuya Yamamoto, MD, PhD – Yokohama City University Graduate School of Medicine
Aimee Luat, MD – Children’s Hospital of Michigan
Eishi Asano, MD, PhD – Wayne State University
Rationale: This study aimed to determine the directional and temporal characteristics of language-related neural information flow through the left arcuate fasciculus.
Methods: We analyzed data from 136 patients with drug-resistant focal epilepsy who performed an auditory descriptive naming task (Figure A) during extraoperative intracranial EEG (iEEG) recordings at Detroit Medical Center, USA. Time–frequency analysis of iEEG data characterized the spatiotemporal profile of task-related high-gamma activity (70–110 Hz). Transfer entropy analysis identified task-related neural information flow between frontal and temporal cortical regions connected via the left arcuate fasciculus (Figure B). Neural information flow was defined as present when high-gamma activity in one cortical region at a given time predicted subsequent activity in another region within 10–50 ms. Furthermore, cortico-cortical spectral responses (CCSRs ), elicited by single-pulse electrical stimulation, were measured to quantify endogenous effective connectivity between frontal and temporal regions. Mixed-effects regression models were used to assess whether the strength of task-related neural information flow correlated with the strength of endogenous effective connectivity, as defined by CCSR amplitudes.
Results:
Bidirectional neural information flow was observed throughout the task. Before the offset of auditory stimuli, flow from the temporal to frontal lobe was dominant (Figure C). After stimulus offset, the direction reversed, with frontal-to-temporal flow becoming dominant. Following the response, temporal-to-frontal flow regained dominance. Mixed-effects regression models demonstrated a positive correlation between the strength of task-related neural information flow and the strength of endogenous effective connectivity (Figure D).
Conclusions: Neural information flow between the left frontal and temporal lobes via the arcuate fasciculus alternates during auditory language processing. Frontal–temporal lobe pairs with stronger endogenous effective connectivity were preferentially engaged in neural communication during auditory naming. This study exemplifies how invasive research in epilepsy patients can help clarify fundamental questions in cognitive neuroscience.
Funding:
This work was supported by NIH grant NS064033 (to E.A.) and KAKENHI Grant JP24K19533 (to M.S.).