Seizure Propagation Pathway Analysis through Latent Embedding of iEEG Signals
Poster No:
1522
Submission Type:
Abstract Submission
Authors:
Kexin Lou1, Markus Barth2, Quanying Liu3
Institutions:
1The University of Queensland, Brisbane, QLD, 2The University of Queensland, Brisbane, Australia, 3Southern University of Science and Technology, Shenzhen, Guangdong
First Author:
Co-Author(s):
Introduction:
Characterizing the temporal and spatial patterns in intracranial electroencephalography (iEEG) signals is a crucial aspect of pre-surgical evaluation for patients with drug-resistant epilepsy. Although deep neural networks have demonstrated high accuracy in detecting and classifying pathological events, they often fall short in revealing the underlying mechanisms driving the state changes to seizure dynamics. Thus, a tool that can not only detect but also interpret the disrupted neural dynamics observed in epilepsy is of high need. The objective of this research: (1) To characterize seizure propagation pathway by providing a novel latent dynamical model-based framework; (2) To interpret the state transition in latent space during seizure progression.
Methods:
We first extracted latent data embeddings using a Graph Neural Network (GNN), which were then segmented into discrete bins representing distinct brain states. Next, we computed the energy landscape of these states and mapped the time points corresponding to different states onto this energy surface. To further analyze the system's dynamics, we quantified the transitions between these states and calculated the entropy production rate, which characterizes the irreversibility and energetic inefficiency of these transitions. We used the HUP iEEG Epilepsy Dataset, which includes both interictal and ictal sessions (Bernabei, 2022). The iEEG signals were processed and labeled. We constructed a GNN to extract node embeddings from the power of time bins. The network was trained using a portion of the inter-ictal and ictal sessions (leaving one session out for each), and evaluated across all sessions to achieve optimal performance. By clustering time points in the latent space into distinct states, we derived a state transition matrix and quantified the disruption in balance through the entropy production rate (Lynn, 2021).
Results:
The dynamic trajectories in latent space depict the temporal progression of brain states during sessions for two example patients, SUB-HUP094 and SUB-HUP060, who had different surgery outcomes: success and failure, respectively. We also visualizes the latent energy landscape with key points overlaid, corresponding to different seizure states for SUB-HUP060. Interictal sessions for SUB-HUP094 reveal a highly localized cluster, whereas those for SUB-HUP060 display a more diffuse and dispersed cluster. This difference indicates a potential variation in baseline brain state stability between the two patients. Ictal sessions in both patients show similar patterns: preictal points tend to lie near interictal nodes, and postictal points form a pathway back from ictal nodes to the interictal region, suggesting a cyclical transition between seizure states. Distinct trajectories are observed in the ictal sessions of SUB-HUP094, while the ictal sessions of SUB-HUP060 show overlapping trajectories. This overlap in SUB-HUP060 may indicate a less distinct or less localized seizure onset zone (SOZ), potentially contributing to the failure of the surgical intervention. he interictal state is associated with a higher energy level, preserving its stability and thus not prominently represented in the lower-energy valleys, In contrast, both preictal and postictal states have points that fall into the energy valleys.
Conclusions:
From the energy landscape analysis, we find that the interictal state is associated with higher energy levels and greater stability, while preictal and postictal states fall into lower-energy valleys, reflecting a less stable, transitional brain state during seizure activity. Ictal sessions show higher EPR values compared to interictal sessions. The time-dependent fluctuations in EPR can serve as a useful indicator for identifying seizure onset.
Modeling and Analysis Methods:
Methods Development 1
Task-Independent and Resting-State Analysis 2
Physiology, Metabolism and Neurotransmission:
Neurophysiology of Imaging Signals
Keywords:
Computational Neuroscience
Data analysis
ELECTROCORTICOGRAPHY
Epilepsy
Modeling
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?
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:
Provide references using APA citation style.
Guo, K., Zhou, K., Hu, X., Li, Y., Chang, Y., & Wang, X. (2022, June). Orthogonal graph neural networks. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 36, No. 4, pp. 3996-4004).
Lynn, C. W., Cornblath, E. J., Papadopoulos, L., Bertolero, M. A., & Bassett, D. S. (2021). Broken detailed balance and entropy production in the human brain. Proceedings of the National Academy of Sciences, 118(47), e2109889118.
You, Y., Chen, T., Sui, Y., Chen, T., Wang, Z., & Shen, Y. (2020). Graph contrastive learning with augmentations. Advances in neural information processing systems, 33, 5812-5823.