Dyslexia is reading disorder - characterized by trouble with reading despite normal intelligence. We used ML methods to find brain regions responsible for this! Check out the paper: "Multi-Parameter Machine Learning Approach to the Neuroanatomical Basis of Developmental Dyslexia" in Human Brain Mapping journal with cooperation with fantastic neurobiologists.
We have a dataset with Magnetic Resonance Images from 236 kids from Poland, France and Germany. Among them 130 were with dyslexia. For each image a brain structure was segmented and measures where computed, such as curvature, area, fold index (work done with FreeSurfer). This created a dataset - which is something we like the most! We run a feature selection on obtained dataset and search for such subset of features which give the lowest classification error with leave-one-out validation. For details please read article :)
We think this is a moonshot! Machine Learning is helping to understand why there is such disorder. On image above there are selected regions that are different for kids with dyslexia.