Computational modeling of seizure spread on a cortical surface

Area of Science:

• Computational neuroscience
• Epilepsy research
• Medical imaging and modeling

Background:

• Neural field models offer insights into large-scale seizure dynamics but lack patient-specific personalization.
• A link between cortical seizure spread and theta-alpha activity (TAA) patterns in electrographic signals has been suggested but not patient-specifically demonstrated.

Purpose of the Study:

• To computationally link seizure propagation across a patient-specific cortical surface with observed TAA patterns.
• To demonstrate the utility of patient-specific cortical geometry in personalized computational epilepsy modeling.

Main Methods:

• Simulated seizure dynamics using The Virtual Brain platform on a patient's realistic cortical geometry.
• Compared simulation results with intracranial electrographic signals, including TAA patterns.
• Performed simulations on surrogate surfaces to assess the impact of patient-specific geometry.

Main Results:

• Simulated electrographic signals qualitatively matched patient-recorded signals.
• The best quantitative fit was achieved using the patient's actual cortical surface geometry.
• The study successfully linked simulated seizure spread to a specific TAA pattern instance.

Conclusions:

• Patient-specific cortical geometry is crucial for accurate personalized computational epilepsy models.
• The Virtual Brain platform can be utilized for personalized modeling, improving understanding of seizure dynamics.
• This approach highlights the importance of integrating patient-specific anatomical data into computational models for clinical relevance.