Open Innovation Advances the Prediction and Detection of Epileptic Seizures
The wisdom of the crowds has once again triumphed. This time by developing new computer algorithms to detect, predict and help prevent epileptic seizures.
National Institute of Neurological Disorders and Stroke (NINDS), United States
Tens of millions of people around the world are affected by epilepsy, a disorder caused by abnormal electrical activity in the brain. Although therapeutic interventions can do much to alleviate the condition, millions of people still suffer from seizures, the onsets of which cannot be detected in advance. Being alerted to when a seizure is about to occur can save lives and improve quality of life. If a seizure is detected it can be prevented with neurostimulation.
Leveraging the Expertise of the Crowd Outside of Neurology
The American Epilepsy Society (AES), the National Institutes of Health’s National Institute of Neurological Disorders and Stroke (NINDS) and the Epilepsy Foundation sponsored an open innovation challenge called the 'Seizure Detection and Prediction Challenge'. It asked the crowd to devise algorithms that would more accurately detect and predict seizure activity.
So, instead of the conventional route of approaching scientists in their labs to come up with innovations, the contest published huge volumes of data for the crowd to get to work on.
The challenge attracted an enormous response, with 504 teams competing in two challenges. They were for seizure detection and seizure prediction. Among those taking part from countries all over the world were people with diverse expertise in industry, academia and data analytics.
Open Innovation Approach Yields Dramatic Benefits
Brian Litt, M.D., co-chair of competition, sang the praises of the benefits of open innovation for these specific challenges. "This is an incredibly dynamic and powerful way to do community research and bring new minds into the challenge of epilepsy.
Before this, after 15 years of research, it was only last year that investigators were able to publish seizure prediction better than random. As a result of the contests, in just a few months, teams with no expertise in epilepsy were consistently getting performance as high as 84%.”
The prediction challenge was won by a team of engineers who decided to combine their intellectual resources during the course of the contest. They were able to forecast abnormal electricity in intracranial recordings of humans and dogs with epilepsy with 82% accuracy. For their efforts, the team shared $20,000 in prize money.
The detection challenge was won by Australian software engineer Michael Hills who picked up a check for $8,000. In this phase, contestants had to analyse EEG recordings from dogs with epilepsy and individuals with medication-resistant seizures.
Data Available for All
In keeping with the open innovation spirit of the contest, the data sets are available for researchers to use freely, and contest participants are openly sharing their software and strategies.
The long-term benefit will be the development of novel detection devices that employ these new algorithms.
Next Story »