Poster No:
1335
Submission Type:
Abstract Submission
Authors:
Pok Him Siu1, Philippa Karoly1, Artemio Soto-Breceda2, Levin Kuhlmann3, David Grayden1
Institutions:
1University of Melbourne, Melbourne, Australia, 2Université Grenoble Alpes, Grenoble, France, 3Monash University, Melbourne, Australia
First Author:
Pok Him Siu
University of Melbourne
Melbourne, Australia
Co-Author(s):
Introduction:
A foundational view of neuroscience is that, in additional to neuronal activity, the structure of the brain constrains and can explain (to an extent) brain function. An alluring formalism in computational neuroscience has been to generate eigenmodes of neural activity from a matrix representing the brain's anatomy (Pang, 2023; Mansour, 2024). Traditionally, brain connectomics has been the gold standard for the generation of these eigenmodes and the coupling between structure and function (Mansour, 2024). However, it has recently been suggested that the geometry of the brain is able to provide more explanatory power in fMRI than the much more complex connectome (Pang, 2023). An adjacent modality is the source localisation problem of EEG, which aims to identify the underlying neuronal generators of EEG recordings. The underdetermined nature of the problem necessitates sufficient constraints to produce realistic and unique solutions of source activity (He, 2018). In this work, we investigate the effectiveness of different types of structural eigenmodes for constraining the source localisation problem in a neurologically feasible manner.
Methods:
Geometric eigenmodes were constructed by solving the eigenvalue problem with the Laplace-Beltrami Operator on sources positioned on the surface mesh of the brain (Pang, 2023). Connectome eigenmodes were constructed by solving the eigenvalue problem on the graph Laplacian representation of the connectome (Mansour, 2024). An optimisation algorithm was used to find the optimal weights, in terms of reconstruction accuracy, of eigenmodes at each time step. The weighted sum of the eigenmodes and the fitted weights is thus the aggregate neural source activity, as required. A simulated dataset of seizure activity, generated from connected neural mass models, was used to benchmark the accuracy of these source localisation approaches.
Results:
Our results demonstrated that structural eigenmodes offer a computationally efficient means to compute biologically feasible sources. Both types of eigenmodes were able to reconstruct beyond 90% of the variance of an EEG signal with only a small subset of eigenmodes, 20 per hemisphere, of the 20,484 dipole source space. The results were consistent with previous work on using structural eigenmodes to reconstruct fMRI data with >80% reconstruction accuracy using 200 eigenmodes per hemisphere (Pang, 2023; Mansour, 2024). Next, we compared the accuracy when geometric and connectome eigenmodes were used to predict the underlying neural sources. We found that the two approaches offered comparable results and are, on average, slightly superior to current off-the-shelf source localisation approaches in terms of region localisation error and spatial deviation of the source.
Conclusions:
Previous work has primarily focused on fMRI; this work on EEG supports the view that, relative to the entire source space, structural eigenmodes offer a lower-dimensional and more interpretable view of the brain's mesoscale activity (Pang, 2023). In the context of EEG source localisation, our results suggest that both geometric and connectome eigenmodes offer slightly improved accuracy but are significantly faster to compute. Additionally, the simpler-to-obtain geometric eigenmodes appear to offer similar explanatory and constraining power as the more complicated connectome eigenmodes.
Source localisation is an important technique for improving the spatial accuracy of EEG and MEG measurements (He, 2018). Improved algorithms will lead to an improved understand of the brain and has direct applications in the diagnosis and treatment of neurologically diseases, such as epilepsy (Makhalova, 2022). In this work, we take the recent pivotal idea in computational neuroscience of structural eigenmodes to biologically constrain the optimisation problem of source localisation. Both geometric and connectome eigenmodes appear to perform just as well for this specific problem.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
EEG/MEG Modeling and Analysis 1
Methods Development
Keywords:
Computational Neuroscience
Electroencephaolography (EEG)
Epilepsy
MEG
Modeling
Source Localization
1|2Indicates the priority used for review
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.
I accept
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:
I do not want to participate in the reproducibility challenge.
Please indicate below if your study was a "resting state" or "task-activation” study.
Other
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Patients
Was this research conducted in the United States?
No
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.
Not applicable
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.
Not applicable
Please indicate which methods were used in your research:
EEG/ERP
Structural MRI
Diffusion MRI
Which processing packages did you use for your study?
Free Surfer
Provide references using APA citation style.
[1] Pang, J. C., Aquino, K. M., Oldehinkel, M., Robinson, P. A., Fulcher, B. D., Breakspear, M., & Fornito, A. (2023). Geometric constraints on human brain function. Nature, 1–9.
[2] Sina Mansour, L., Behjat, H., De Ville, D. V., Smith, R. E., Yeo, B. T. T., & Zalesky, A. (2024). Eigenmodes of the brain: Revisiting connectomics and geometry.
[3] He, B., Sohrabpour, A., Brown, E., & Liu, Z. (2018). Electrophysiological Source Imaging: A Noninvasive Window to Brain Dynamics. Annual Review of Biomedical Engineering, 20(1), 171–196.
[4] Makhalova, J., Medina Villalon, S., Wang, H., Giusiano, B., Woodman, M., Bénar, C., Guye, M., Jirsa, V., & Bartolomei, F. (2022). Virtual epileptic patient brain modeling: Relationships with seizure onset and surgical outcome. Epilepsia, 63(8), 1942–1955.
No