Australasian Science: Australia's authority on science since 1938

Crowdsourced Algorithms Predict Epileptic Seizures

Crowdsourcing of more than 10,000 algorithms worldwide has enabled Melbourne researchers to predict clinically relevant epileptic seizures in a wider range of patients than previously possible.

In 2016 the researchers ran a seizure prediction challenge on the online data science platform kaggle.com. The contest focused on seizure prediction from long-term electrical brain activity recordings obtained in 2013 from a clinical trial of NeuroVista’s implantable Seizure Advisory System. Almost 478 teams developed 10,000 algorithms, and the top algorithms were tested on the patients with the lowest seizure prediction performance based on previous studies.

The results have now been published in Brain (https://goo.gl/DxbX9e). “Our evaluation revealed on average a 90% improvement in seizure prediction performance, compared to previous results,” said Dr Levin Kuhlmann from the University of Melbourne’s Graeme Clarke Institute and St Vincent’s Hospital.

“Epilepsy is highly different among individuals,” Kuhlmann explained. “Results showed different algorithms performed best for different patients, supporting the use of patient-specific algorithms and long-term monitoring.”

The researchers have now developed the online platform epilepsyecosystem.org for algorithm and data sharing to further develop and improve seizure prediction. “Accurate seizure prediction will transform epilepsy management by offering early warnings to patients or triggering interventions,” Kuhlmann said.

“Our results highlight the benefit of crowdsourcing an army of algorithms that can be trained for each patient, and the best algorithm chosen for prospective, real-time seizure prediction. The hope is to make seizures less like earthquakes, which can strike without warning, and more like hurricanes, where you have enough advance warning to seek safety.”