Computer scientists and mathematicians at the CEU Cardenal Herrera University in Valencia have developed a prediction model that can warn epileptic sufferers of an upcoming seizure with 20 minutes notice.
Being able to predict the onset of a debilitating seizure would significantly improve the quality of life of epileptics. With this goal, scientists from around the world competed in the American Epilepsy Society Seizure Prediction Challenge in 2014. A team from the Universidad Cardenal Herrera CEU (CEU-UCH) reached third place overall, winning them an invitation to the Society’s annual meeting in Seattle. Their prediction model has recently been published in the article Crowdsourcing reproducible seizure forecasting in human and canine epilepsy in the OUP’s Brain.
CEU-UCH researchers Juan Pardo, Javier Muñoz, Francisco Zamora and Paloma Botella developed an algorithm able to interpret the relevant data and predict a seizure up to 20 minutes before it occurs, giving the patient time to take their medication and nip it in the bud. Their model came third out of 502 submissions.
The idea is that a portable device like a mobile phone or a smartwatch reads signals sent from a microchip implanted in the patient’s brain, interpreting the data according to the researchers’ prize-winning algorithm and warning the patient of an upcoming seizure. The benefits of this set-up are two-fold: not only would epilepsy sufferers only have to take their medication in the event of a predicted seizure and not constantly throughout the day, mitigating any adverse effects, but they would also be able to reclaim a convincing semblance of a normal life, only avoiding ‘high-risk’ activities like driving or swimming until the risk of their having a seizure had passed.
Talking to the University back in 2014, Juan Pardo, director of the Department of Physical Sciences, Mathematics and Computing, said that “we are not doctors, but we are experts in massive data processing”, known today as Big Data. “This work shows how computer scientists and mathematicians are needed in many fields to help solve pressing issues, in this instance in neuroscience”.
The data in question was released as part of the competition, and comprised datasets of intracranial EEG recordings from canine and human epilepsy sufferers.
Funding: The authors acknowledge the generous support of the American Epilepsy Society, The Epilepsy Foundation, Kaggle.com (which waived a portion of its normal fee for this competition), and the National Institutes of Health. Data collection, processing, analysis, and manuscript preparation were supported by NeuroVista Inc. and grants NIH-NINDS UH2/UH3 95495 (G.W.), U01-NS 73557 (G.W.), U24-NS063930 (B.L., G.W.), K01 ES025436-01 (J.W.), and R01-NS92882 (G.W.), the Mirowski family foundation, and Mayo Clinic.
Image Source: The image is credited to Worrell et al./Brain.
Original Research: Full open access research for “Crowdsourcing reproducible seizure forecasting in human and canine epilepsy” by Benjamin H. Brinkmann, Joost Wagenaar, Drew Abbot, Phillip Adkins, Simone C. Bosshard, Min Chen, Quang M. Tieng, Jialune He, F. J. Muñoz-Almaraz, Paloma Botella-Rocamora, Juan Pardo, Francisco Zamora-Martinez, Michael Hills, Wei Wu, Iryna Korshunova, Will Cukierski, Charles Vite, Edward E. Patterson, Brian Litt, Gregory A. Worrell in Brain. Published online March 31 2016 doi:10.1093/brain/aww045
Crowdsourcing reproducible seizure forecasting in human and canine epilepsy
Accurate forecasting of epileptic seizures has the potential to transform clinical epilepsy care. However, progress toward reliable seizure forecasting has been hampered by lack of open access to long duration recordings with an adequate number of seizures for investigators to rigorously compare algorithms and results. A seizure forecasting competition was conducted on kaggle.com using open access chronic ambulatory intracranial electroencephalography from five canines with naturally occurring epilepsy and two humans undergoing prolonged wide bandwidth intracranial electroencephalographic monitoring. Data were provided to participants as 10-min interictal and preictal clips, with approximately half of the 60 GB data bundle labelled (interictal/preictal) for algorithm training and half unlabelled for evaluation. The contestants developed custom algorithms and uploaded their classifications (interictal/preictal) for the unknown testing data, and a randomly selected 40% of data segments were scored and results broadcasted on a public leader board. The contest ran from August to November 2014, and 654 participants submitted 17 856 classifications of the unlabelled test data. The top performing entry scored 0.84 area under the classification curve. Following the contest, additional held-out unlabelled data clips were provided to the top 10 participants and they submitted classifications for the new unseen data. The resulting area under the classification curves were well above chance forecasting, but did show a mean 6.54 ± 2.45% (min, max: 0.30, 20.2) decline in performance. The kaggle.com model using open access data and algorithms generated reproducible research that advanced seizure forecasting. The overall performance from multiple contestants on unseen data was better than a random predictor, and demonstrates the feasibility of seizure forecasting in canine and human epilepsy.
“Crowdsourcing reproducible seizure forecasting in human and canine epilepsy” by Benjamin H. Brinkmann, Joost Wagenaar, Drew Abbot, Phillip Adkins, Simone C. Bosshard, Min Chen, Quang M. Tieng, Jialune He, F. J. Muñoz-Almaraz, Paloma Botella-Rocamora, Juan Pardo, Francisco Zamora-Martinez, Michael Hills, Wei Wu, Iryna Korshunova, Will Cukierski, Charles Vite, Edward E. Patterson, Brian Litt, Gregory A. Worrell in Brain. Published online March 31 2016 doi:10.1093/brain/aww045