Personalized Morphological Brain Network Analysis for Temporal Lobe Epilepsy: Combining Cortical Thickness and Volume Features

Temporal lobe epilepsy (TLE) is the most common form of drug-resistant epilepsy and is characterized as a brain network disorder. Identification of reliable biomarkers is crucial for accurate diagnosis and effective treatment. T1-weighted magnetic resonance imaging (MRI) is widely used to assess TLE, with cortical thickness and volume atrophy in widespread brain regions being the main morphological changes. However, focusing only on morphological features overlooks the interactions between different brain regions. This study used the Morphological Inverse Divergence (MIND) method to construct personalized brain network by integrating cortical thickness and volume, and explored TLE biomarkers by combining network and morphological features. The results showed that network-based biomarkers improved diagnostic accuracy by more than 10% compared to morphological biomarkers alone, and their combination provided better classification performance. These findings offer new insights into TLE pathology and have important implications for personalized clinical management.