Group Joint ICA for EEG-fMRI Fusion: A Multimodal Integration Approach
Poster No:
1445
Submission Type:
Abstract Submission
Authors:
Souvik Phadikar1, Oktay Agcaoglu1, Lei Wu1, Vince Calhoun2
Institutions:
1Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Atlanta, GA, 2GSU/GATech/Emory, Atlanta, GA
First Author:
Souvik Phadikar
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Co-Author(s):
Oktay Agcaoglu, PhD
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Lei Wu
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Introduction:
Electroencephalography (EEG) provides high temporal resolution but limited spatial resolution, while functional magnetic resonance imaging (fMRI) offers high spatial resolution with poor temporal resolution. Integrating EEG and fMRI enables exploration of brain dynamics across space and time but remains challenging due to their distinct data representations: EEG directly captures neural activity, whereas fMRI reflects slower hemodynamic responses. Joint independent component analysis (jICA) and parallel ICA are commonly used to identify shared and modality-specific sources; however, they operate on single-subject features and cannot link multi-subject time series data (Calhoun, 2006; Eichele, 2008; Heugel, 2022; Liu, 2007). This paper proposes a multimodal fusion technique combining group ICA with joint ICA, preserving full temporal and spatial information at the subject level and estimating joint components across subjects. Key advantages include linking EEG and fMRI across space, time, and subjects while capturing subject-specific information through back-reconstruction.
Methods:
This section outlines the proposed group jICA method for fusing simultaneous EEG and fMRI data in six steps: (1) Preprocessing and convolution – EEG and fMRI data were preprocessed, and EEG was convolved with a hemodynamic response function (HRF) to approximate fMRI delays; (2) Feature extraction – EEG spectrograms were computed using a sliding window (matched to the fMRI repeat time (TR)) and concatenated across channels; (3) Modality merge and PCA – subject-specific EEG spectrograms and fMRI volumes were concatenated, and PCA was applied to extract p1 principal components (PCs); (4) Grouping data and PCA – subject-specific PCs were then concatenated across subjects, and second-level PCA was applied to reduce group data to p2 PCs; (5) Joint ICA – joint ICA was applied to the reduced group data to estimate p2 components; (6) Back reconstruction – regression-based analysis reconstructed joint EEG and fMRI components. The proposed method was applied to simultaneous EEG and fMRI data from 25 subjects, and basic preprocessing steps were performed on both the EEG and fMRI data separately, for more detail readers may refer to (Phadikar, 2024; Wu, 2010).
Results:
The proposed method estimated 45 components from simultaneous resting-state EEG and fMRI data, one of which is shown in Figure 1. The fMRI component (Figure 1(a)) presents activation in the frontal region, part of the default mode network. The corresponding EEG components (Figure 1(b)) were divided into four frequency bands: delta (0.1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), and beta (12–30 Hz) and displayed topographically, showing high frontal activation in theta and alpha bands. EEG source localization (Figure 1(c)) reveals the spatial distribution of neural activity, with strong activation in the frontal cortices as seen from top-down, lateral, and anterior views. The color scale represents neural activity intensity, with yellow and white indicating the highest values.
·Figure 1: Jointly estimated fMRI (a) and EEG (b and c) components.
Conclusions:
We introduce group jICA, an effective multimodal fusion approach to integrate simultaneous EEG and fMRI data for identifying resting-state networks (RSNs). The method transforms EEG and fMRI data into a unified feature space using PCA and estimates RSNs through ICA, capturing both temporal and frequency-specific information. Our results reveal physiologically meaningful associations, such as the default mode network linked to theta and alpha bands in the frontal cortex. This promising approach paves the way for broader analyses, with future studies focusing on larger cohorts and clinical populations to identify group differences and clinically relevant RSNs.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
EEG/MEG Modeling and Analysis 2
fMRI Connectivity and Network Modeling 1
Keywords:
Data analysis
Electroencephaolography (EEG)
FUNCTIONAL MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Eichele, T., Calhoun, V. D., Moosmann, M., Specht, K., Jongsma, M. L., Quiroga, R. Q., Nordby, H., & Hugdahl, K. (2008). Unmixing concurrent EEG-fMRI with parallel independent component analysis. International Journal of Psychophysiology, 67(3), 222-234.
Heugel, N., Beardsley, S. A., Liebenthal, E. (2022). EEG and fMRI coupling and decoupling based on joint independent component analysis (jICA). Journal of neuroscience methods, 369, 109477.
Liu, J., & Calhoun, V. D. (2007). Parallel independent component analysis for multimodal analysis: application to fMRI and EEG data IEEE International Symposium on Biomedical Imaging: From Nano to Macro,
Phadikar, S., Pusuluri, K., Jensen, K. M., Wu, L., Iraji, A., Calhoun, V. D. . (2024). Coupling Between Time-varying EEG Spectral Bands and Spatial Dynamic fMRI Networks IEEE International Symposium on Biomedical Imaging (ISBI), Athens, Greece.
Wu, L., Eichele, T., Calhoun, V. D. (2010). Reactivity of hemodynamic responses and functional connectivity to different states of alpha synchrony: a concurrent EEG-fMRI study. NeuroImage, 52(4), 1252 - 1260.