via hackster.io

Roughly 1% of the population lives with anxiety around their next seizure event. Anti-convulsant medication does not help at least 20% of those afflicted, but also carries negative side effects in high doses. This fear stands in the way of performing everyday tasks like driving a car or swimming. There exist devices that can detect that a seizure is occurring, but we aim to build a wearable EEG device that can predict that a seizure is about to occur so a caregiver and/or the user can pause what they are doing and take appropriate preventative action.

Five years ago, a Kaggle contest framed the challenge of developing a machine learning algorithm to help predict seizure events. Contestants were provided 106 GB of training data in.mat files. Each training sample represents 10 minutes of raw intracranial EEG readings. For 5 dogs, data included 16 channel readings sampled at 400 Hz. For 2 human patients, data was sampled at 5000 Hz over 24 channels.

Read More.

Each Friday is PiDay here at Adafruit! Be sure to check out our posts, tutorials and new Raspberry Pi related products. Adafruit has the largest and best selection of Raspberry Pi accessories and all the code & tutorials to get you up and running in no time!