Abstract:
Epilepsy is a common neurological disorder affecting approximately 1% of the global population, leading to recurrent seizures and various neurological, cognitive, and social challenges. The pathophysiology of epilepsy is attributed to an imbalance between neuronal excitation and inhibition, prompting research to focus on the molecular and network-level mechanisms of epileptic seizures.
Recent clinical studies emphasize that epilepsy is a network disorder characterized by abnormalities in the intrinsic structure and dynamic organization of brain networks. Previous research on epilepsy networks has primarily focused on the stationary functional connectivity between brain regions. However, brain networks are inherently transient and dynamically adapting to changing external and internal environments. Stationary brain network analysis methods are insufficient to fully capture the nature of disruptions in epilepsy networks. In this dissertation, magnetoencephalography (MEG) is employed to collect high temporal resolution functional brain data. A novel analytical approach is used to capture transient brain state dynamics. Here, brain states are defined as distinct network patterns that are co-activated in space. Using this method, we compared the brain state dynamics between healthy controls and patients with generalized epilepsy at rest and task. Using machine learning algorithms, we identified nine common brain states shared by both epileptic patients and healthy individuals during rest and task performance. Temporal characteristics of brain states in epileptic patients, such as frequency occurrence (FO) and average duration (AR), mainly showed changes during tasks. Interestingly, our findings indicate that brain state transitions in epilepsy patients are impaired, as evidenced by a decrease in the entropy of transition probability (TP). Moreover, the support vector machine (SVM) classification algorithm with a radial basis function (RBF) kernel, utilizing FO and TP as multivariate features, effectively differentiated between epileptic patients and healthy participants, demonstrating the robustness of FO and TP as biomarkers for brain state dynamics in epilepsy.
Our study suggests that statistical descriptions of changes of brain activity patterns can effectively capture rapid brain state dynamics. Differences in brain dynamics between generalized epilepsy patients and healthy individuals are more easily observed during arithmetic task. In summary, our research provides new insights into the brain dynamics of generalized epilepsy, which may contribute to optimizing the diagnosis and treatment of patients suffering from this condition.
Epilepsy