Deep Learning in Epilepsy
Epilepsy is a disease involving the spontaneous occurrence of seizure events, which can be dangerous when they occur and are also debilitating to patients' daily lives due to their unpredictability. Developing methods to monitor and study this disease is difficult and time consuming as patients often must be studied individually due the highly variable physiology. The emerging field of Machine Learning (ML) and Deep Learning (DL) based algorithms presents a unique opportunity in this space to develop personalized algorithms tuned to each specific patient after initial recording of some representative data. Here at CID we are working to integrate existing ML and DL techniques with our devices and custom datasets to develop algorithms which can analyze, detect, and predict seizure events in real time with never-before-seen accuracy. We are also working to develop new DL techniques that are optimized for time-series biological data and the types of signals that are often collected in biomedical contexts. All of the code and resources we develop to pursue this goal will be made freely available upon publication to other researchers and companies to aid in the development and deployment of highly successful algorithms for epilepsy patients and others in need of sophisticated personalized signal analysis.