AI for the personalised treatment of epilepsy

Sandra Acosta

Universitat Pompeu Fabra (UPF), Barcelona, Spain

The response to treatments in epileptic encephalopathies is often very poor or even deleterious. This project proposes the use of brain organoids derived from patients’ induced pluripotent stem cells (iPSC) to evaluate the effect of multiple treatments without risks for the patient. The project aims to bring to market a recently developed deep learning technology to predict the effect of epilepsy treatments based on image analysis of the patient’s organoids. The project propose to test multiple drug treatments alone and in combination using in vitro brain organoids, to help the neurologists choose the most suitable treatment for each patient.

Their methodology was developed by combining image-based organoid analysis and a newly generated deep learning algorithm. To date, they have achieved a proof of concept of the algorithm predicting the wellness of organoids, which is a basic step to evaluate potential toxic effects of the treatments and the disease.