Abstracts

Rationale: The success of resection therapies in patients with drug-refractory epilepsy relies on accurately identifying the epileptogenic zone (EZ), which comprises the specific brain regions contributing to seizure generation. The conventional methodology for EZ identification by clinicians is subjective, time-consuming, and lacks integration of the complex temporospatial dynamic processes involved. Development of an automated network-based EZ detection framework has the potential to not only improve surgery outcomes but also to advance next-generation neurostimulation therapies. Currently, most research in the field is constrained to simple linear network analysis or focuses on modeling only one aspect of these complex connectivities. Our study aims to identify the EZ by utilizing various features derived from comprehensive non-linear graphical models of the epileptogenic network, providing a more holistic understanding of the condition.


Methods: We utilize a dataset including five patients with intractable MTLE who have undergone resective or ablative surgeries following stereo-electroencephalography (sEEG) investigation and remained seizure-free more than 1 year post-surgery (Engel I outcome). We create two distinct graphical representations that capture the non-linear spectral and temporal connectivities during pre-ictal and ictal intervals. Various features extracted from this multigraph structure are leveraged by machine learning algorithms tasked with detecting the EZ within the modeled networks. This is formulated as a binary node classification problem, with the nodes corresponding to the recording channels. Within our framework, these nodes are classified as either EZ or non-EZ. Finally, spatial attributes of the electrodes are employed within a graph convolutional filtering process to fine-tune the predicted labels.


Results: The results are obtained using leave-one-out cross-validation, which marks an accuracy of 94.2% and F1-score of 75%. This demonstrates a notable increase compared to other well-known methods, emphasizing the efficacy of our approach. Additionally, we observe a significant decrease in performance when excluding the multigraph-based features or the fine-tuning step, highlighting the importance of utilizing multiple graphs to represent the epileptogenic network and the effectiveness of incorporating the electrodes’ spatial properties in detecting the EZ.


Conclusions: We demonstrate that to comprehensively represent the epileptogenic network, it is necessary to model both temporal and spectral aspects of pre-ictal and ictal intervals. Additionally, utilizing advanced tools in network science and graph signal processing is crucial to obtaining insightful properties of these graphs, enhancing the EZ detection performance.


Funding: Funding: NIH (UH3): Network based neuromodulation for mesial temporal lobe epilepsy