A Novel Spatio-Temporal Event Network Information Mining Method for Resting-state EEG
Poster No:
1641
Submission Type:
Abstract Submission
Authors:
Qiwei Dong1,2,3, Runchen Yang2,4, Zongwen Feng2,4, Li Dong*2,4, Qi Xu1,3, Dezhong Yao2,3,4
Institutions:
1Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences & Peking Union, Medical College (CAMS & PUMC), Beijing, China, 2The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation, University of Electronic Science and Technology of China, Chengdu, China, 3Research Unit of NeuroInformation, Chinese Academy of Medical Sciences, Chengdu 2019RU035, China, 4Sichuan Institute for Brain Science and Brain-Inspired Intelligence, Chengdu, China
First Author:
Qiwei Dong
Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences & Peking Union|The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences
Medical College (CAMS & PUMC), Beijing, China|University of Electronic Science and Technology of China, Chengdu, China|Chengdu 2019RU035, China
Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences & Peking Union|The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences
Medical College (CAMS & PUMC), Beijing, China|University of Electronic Science and Technology of China, Chengdu, China|Chengdu 2019RU035, China
Co-Author(s):
Runchen Yang
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
Zongwen Feng
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
Li Dong*
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu, China
Qi Xu
Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences & Peking Union|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences
Medical College (CAMS & PUMC), Beijing, China|Chengdu 2019RU035, China
Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences & Peking Union|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences
Medical College (CAMS & PUMC), Beijing, China|Chengdu 2019RU035, China
Dezhong Yao
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu 2019RU035, China|Chengdu, China
The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation|Research Unit of NeuroInformation, Chinese Academy of Medical Sciences|Sichuan Institute for Brain Science and Brain-Inspired Intelligence
University of Electronic Science and Technology of China, Chengdu, China|Chengdu 2019RU035, China|Chengdu, China
Introduction:
In face of the complex brain information, it is necessary to explore the spatiotemporal dimension in order to elucidate the precise temporal coordination and spatial connectivity patterns (Iraji et al., 2022; Park and Friston, 2013). However, there is presently a significant shortfall in the development of EEG-based method for mining spatiotemporal information of brain activity. This study introduces a novel framework, named EEG Spatio-Temporal Event Network Analysis (ESENA), to map rhythmic activity related events into a network structure according to their temporal sequence to mine spatiotemporal patterns of the brain. The purpose of this study was to introduce ESENA for validating the mapping and interpretation of dynamic brain activity patterns during resting-state condition.
Methods:
Data were collected from 50 subjects (age range = 18-22 years) during a resting-state with their eyes closed. All data were preprocessed and analyzed using the WeBrain toolbox (Dong et al., 2021). The main procedures included: filtering (1-60Hz), quality assessment, artifacts removal using ICA, re-referencing to REST and bad channels interpolating. Based on the extraction of recurrent spatiotemporal patterns from EEG data (Ferreira et al., 2020), the ESENA analysis contains following steps: 1) segmenting clean EEG data into epochs of equal length, 2) calculating the relative power spectrum (Delta: 1-4Hz, Theta: 4–8 Hz, Alpha: 8–12.5 Hz, Beta: 12.5–30 Hz, Gamma: 30–60 Hz) for each epoch, 3) defining events of channels that exceed a given threshold (event threshold) are identified as detected events for each epoch, and 4) constructing network by accumulating links of adjacent epoch channels with events and normalizing to 0-1. The parameters of event threshold and data length were investigated, and a one-sample T-test is conducted on all the links (FDR<0.05). After testing, connections that are consistently present in 60% of the dataset are retained as the final connections.
Results:
As shown in Figure 1, the optimal parameter of event threshold was set at 1 standard deviation, and data length was set at 90s. In the delta band, links are mainly located in the fronto-parietal regions. The theta band networks are main links between central frontal and parietal areas. The alpha band network shows links in the inferior parietal-occipital regions. For the beta and gamma band, the ESENA highlights connections predominantly between the bilateral temporal lobes. In addition, the links of those networks are mainly located in the regions of relative power spectrum.
Conclusions:
ESENA perhaps is a promised method to mine the spatio-temporal network information for resting-state EEG data. This method may provide new insight for enhancing our understanding of brain activity patterns.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
EEG/MEG Modeling and Analysis 1
Methods Development
Keywords:
Electroencephaolography (EEG)
Other - EEG network; spatio-temporal information; resting-state
1|2Indicates the priority used for review
Provide references using author date format
Ferreira, L.N. (2020), 'Spatiotemporal data analysis with chronological networks', Nature Communications, vol. 11, no.1, pp. 4036.
Iraji, A. (2022), 'Moving beyond the 'CAP' of the Iceberg: Intrinsic connectivity networks in fMRI are continuously engaging and overlapping', Neuroimage, vol. 251, pp. 119013.
Park, H.J. (2013), 'Structural and functional brain networks: from connections to cognition', Science, Vol. 342, no. 6158, pp. 1238411.