Transfer Entropy-based Analysis of Dynamics of Neural Information During Auditory Naming
Abstract number :
2.198
Submission category :
3. Neurophysiology / 3G. Computational Analysis & Modeling of EEG
Year :
2024
Submission ID :
1157
Source :
www.aesnet.org
Presentation date :
12/8/2024 12:00:00 AM
Published date :
Authors :
Presenting Author: Masaki Sonoda, MD, PhD – Yokohama City University
Ethan Firestone, MD-PhD Candidate – Wayne State University
Naoto Kuroda, MD – Wayne State University
Kazuki Sakakura, MD, PhD – Rush University
Tetsuya Yamamoto, MD, PhD – Yokohama City University
Eishi Asano, MD/PhD – Wayne State University
Rationale: This study aims to employ a transfer entropy-based effective connectivity analysis on intracranial EEG (iEEG) data to visualize the dynamics of cortico-cortical neural information flows during auditory naming.
Methods: We studied 136 patients with drug-resistant focal epilepsy who underwent an auditory descriptive naming task during extraoperative iEEG recording as part of the presurgical evaluation at Detroit Medical Center, Detroit, USA. Time-frequency analysis determined the spatiotemporal characteristics of naming-related cortical high gamma augmentation (HGA) at 70-110 Hz. We employed transfer entropy analysis to identify significant neural information flows, defined as instances where HGA in a given region at a specific time point predicted sustained HGA in another region within 10-50 ms. We visualized transfer entropy-based neural information flows across 35 cortical regions of interest (ROIs) defined by the Desikan-Killiany Standard Brain Atlas.
Results: A total of 14,827 intracranial non-epileptic electrode sites (36,319 bipolar channels) were analyzed. Significant neural information flows were identified in 4,663 ROI pairs, with the majority occurring bidirectionally and without direction specificity (Figure 1). Conversely, 455 ROI pairs (9.8%) exhibited direction-preferential neural information flows, most frequently originating from the superior frontal gyri in both hemispheres. Additionally, 482 ROI pairs demonstrated significant hemispheric differences in neural information flows, with 93.2% showing higher flows on the left side. Left-hemispheric neural information flows were notably bidirectional between the superior temporal and middle/inferior temporal gyri, and between the middle/inferior temporal and inferior frontal gyri around stimulus offset. Similar bidirectional flows were observed between the inferior frontal gyrus and Rolandic area prior to overt responses.
Conclusions: Our transfer entropy-based analysis successfully quantified naming task-related neural information flows at the whole-brain level. The observed spatiotemporal dynamics align with the notion that left hemispheric-dominant, bidirectional neural information flows across perisylvian cortical structures support sound-semantic transformation, lexical retrieval, and articulatory planning.
Funding: This work was supported by NIH grant NS064033 (to E.A.) and AMED grant JP22he2202018 (to M.S.) as well as KAKENHI Grant JP24K19533 (to M.S.).
Neurophysiology