TLE Classification

The study evaluated machine learning models, including support vector machines (SV) and deep learning (DL), to detect brain abnormalities in temporal lobe epilepsy (TLE) using MRI data from the ENIGMA-Epilepsy consortium. The models demonstrated similar or better performance in identifying TLE compared to lateralizing the side of TLE, with structural data showing higher accuracy for diagnosis and diffusion data for lateralization. Overall, SV and DL models exhibited comparable classification accuracies, with SV occasionally outperforming DL, and models for patients with hippocampal sclerosis showed higher accuracy than those for non-lesional patients.