<html dir="ltr">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" id="owaParaStyle"></style>
</head>
<body fpstyle="1" ocsi="0">
<div style="direction: ltr;font-family: Tahoma;color: #000000;font-size: 10pt;">
<p class="MsoNormal">After much anticipation the <b>Melbourne-University AES-MathWorks-NIH Seizure Prediction Challenge</b> has launched on Kaggle.com!<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Enter for your chance to win part of the <b>US$20,000</b> prize pool and test your data science skills against the one-of-a-kind long-term human intracranial EEG database from the world-first human clinical trial of the NeuroVista Seizure
Advisory System that was co-ordinated by the University of Melbourne. This device was implanted in the heads of epilepsy patients to record brain activity over a period of 6 months to 3 years. Typical recordings of intracranial EEG in humans only last up to
two weeks and do not provide enough data to allow accurate evaluation of seizure prediction algorithms because often only a handful of seizures can be collected over two weeks. The durations of data in the NeuroVista dataset overcome this problem. <o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Analysis of the human NeuroVista dataset has indicated that seizure prediction in humans is in fact possible, however, improvements can still be achieved depending on the patient. This contest seeks to find improved methods by contributing
data from 3 patients whose seizures are difficult to predict.<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">In 2014, our contest partners from the Mayo Clinic and University of Pennsylvania ran a seizure prediction contest on Kaggle.com involving long-term data from dogs, that were also implanted with the NeuroVista device, and short-term human
data. The contest revealed several novel and existing approaches that performed well and now we want to know how well they can perform on long-term data from humans. Can you help us find out, or can you come up with even better algorithms?<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"><o:p>Neural net/deep learning approaches are strongly encouraged, although any approach is welcome.</o:p></p>
<p class="MsoNormal"><o:p><br>
</o:p></p>
<p class="MsoNormal">Everything you need to get started is on the contest web page:<o:p></o:p></p>
<p class="MsoPlainText"><a href="https://www.kaggle.com/c/melbourne-university-seizure-prediction">https://www.kaggle.com/c/melbourne-university-seizure-prediction</a><o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Be sure to get started soon as the <b>contest ends on November 21</b> and the winners will be announced at the American Epilepsy Society Annual Meeting on December 5.
<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Good luck!<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">On behalf of the organising team<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">University of Melbourne: Levin Kuhlmann, Mark Cook, David Grayden, Dean Freestone, Philippa Karoly<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">University of Pennsylvania: Brian Litt<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">Mayo Clinic: Greg Worrell, Ben Brinkmann<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">Alliance for Epilepsy Research: Susan Arthurs<o:p></o:p></p>
<p class="MsoNormal">And our co-sponsors <o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">American Epilepsy Society<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">MathWorks<o:p></o:p></p>
<p class="MsoNormal" style="text-indent:36.0pt">National Institutes of Health<o:p></o:p></p>
<p class="MsoNormal">And Kaggle.com<o:p></o:p></p>
</div>
</body>
</html>