Multilayer brain networks can identify the epileptogenic zone and seizure dynamics

Area of Science:

• Neuroscience
• Computational Neuroscience
• Network Science

Background:

• Seizure generation, propagation, and termination involve complex spatiotemporal brain networks.
• Identifying the epileptogenic zone (EZ) is critical for successful epilepsy surgery.
• Understanding large-scale brain interactions, particularly in high-frequency bands, is key to characterizing seizure dynamics.

Purpose of the Study:

• To demonstrate the significance of high-frequency (80-200Hz) large-scale brain interactions for identifying the EZ and understanding seizure evolution.
• To introduce a novel multilayer network model and connectivity measure (mlEVC) for analyzing neural dynamics during seizures.
• To develop and validate an algorithm for EZ identification using consensus hierarchical clustering.

Main Methods:

• Modeling brain connectivity using multilayer networks constructed from stereoelectroencephalography (SEEG) data during seizures.
• Introducing a new measure for temporal network connectivity: multilayer eigenvector centrality (mlEVC).
• Applying consensus hierarchical clustering to identify the EZ as a distinct network cluster during the ictal period.

Main Results:

• The algorithm successfully predicted EZ electrodes in 88% of seizure-free patients post-surgery.
• Significant desynchronization between the EZ and the rest of the brain was observed during early to mid-seizure.
• Aging and epilepsy duration were correlated with increased desynchronization, suggesting neuroplasticity involvement.
• Seizures exhibited diverse network topologies, indicating patient-specific epileptogenic networks.
• Analysis highlighted the necessity of sufficient data for accurate epileptogenic network identification.

Conclusions:

• High-frequency brain network analysis, using multilayer networks and mlEVC, is effective for identifying the epileptogenic zone.
• Early intervention in epilepsy may be beneficial, with network desynchronization potentially indicating disease severity.
• Understanding patient-specific network dynamics is crucial for tailoring epilepsy treatments.
• The findings underscore the importance of comprehensive data collection for mapping epileptogenic networks.