Learning how to see the invisible - using machine learning to find underlying abnormality patterns in reportedly normal MR brain images from patients with epilepsy

The visual identification of subtle abnormalities in MR brain images that underlie focal epilepsies is a challenging problem. In this study, we used machine learning techniques to uncover patterns of abnormality that exist within reportedly normal brain images from individuals with epilepsy. Our results demonstrated that abnormalities exist in MR images reported to be normal by a human reader, and that these abnormalities exist in a different spatial pattern to that seen in visually apparent cases. We obtained novel insights into why visual assessment may be ineffective in these visually normal cases and provide suggestions on how to improve this situation.