The goal of the workshop entitled "Forecasting epileptic seizures" is to develop models based on brain activity monitored by EEG to forecast upcoming periods of heightened seizure risk in patients with epilepsy. The workshop is based on a dataset of 18 patients, each with more than two years of continuous electroencephalographic (EEG) recordings obtained via an intracranially implanted recording device (a sort of pacemaker for the brain). Specifically, participants will have access to these long EEG recordings and will be asked to use input features based on this measurement to forecast seizure probabilities. Before the hand-on experience, the organizers of the workshop (a neurologist and a computational neuroscientist) will offer succinct explanations about the medical context, the origin and nature of the data, and the advantages and disadvantages of seeking deterministic versus probabilistic forecasts for that specific application.

Participants in the workshop will learn about one field of medicine - epilepsy, arguably the most dynamic human disorder - that lends itself to quantitative approaches. From a method standpoint, they will understand the difference between deterministic and probabilistic forecasting for a tangible application. 

• Personal laptop and internet connection
• Coding skills in Python and R - intermediate
• Participants will run Python and R Jupyter notebooks directly on mybinder.org.