Algorithms to Detect Seizures: Kaggle Neuroscience Challenge

seizure spike chains


Data-driven insights are changing how we understand the world. Analyzing cell phone patterns can help microlenders assess risk of providing loans to low-income families in developing economies; understanding population-level medical records is revolutionizing healthcare providers’ abilities to predict and treat diseases. Big data is also catalyzing evolutions in neuroscience.

Enter Kaggle, a site where challenges are posted for hackers and datamongers. Programmers battle to develop the best algorithms for cash rewards. EyeWire friends Brian Litt and Greg Worrell, of UPenn and MIT, respectively, have launched a seizure detection challenge with an $8,000 reward. Just two weeks remain before one team reigns supreme. Following the $8K competition, a $20,000 algo challenge for seizure prediction will be launched. Keep an eye on Kaggle for that announcement.

Design an algo; save a life.

Half a million Americans suffer from uncontrollable drug-resistant epilepsy, a disease that cripples lives by causing unrelenting seizures. The best hope of these patients is an implanted neurostimulation device which can detect seizures at their onset and prevent them from spreading – but only if they are caught early enough.

This Kaggle competition challenges participants to algorithmically identify seizures with low error rates. Check it out at



Leave a Reply

Your email address will not be published.