Post-doc position offer
Topic: detection and anticipation of absence seizures based onbiosignal processing and machine learning
Background. Epilepsy refers to a neurological disorder that affects about 1% of the general population (World Health Organization, Fact Sheet on Epilepsy, Oct. 2012). Among the multiple forms of epilepsy, absence epilepsy (AE) is characterized by absence seizures which consist of a sudden loss of consciousness. Typical seizures usually last for 5 to 15 seconds. They usually stop as suddenly as they start, with patients resuming their normal activities. These short episodes can be frequent and strongly impact the quality of life of patients. In particular, absence seizures strongly impair cognitive functions and learning capacities. Consequences are particularly severe in children (childhood absence epilepsy - CAE) as seizures seriously interfere with their ability to pay attention, to participate in class and ultimately to learn.
Rationale. While AE usually responds very well to treatment, a large fraction (about 30%) remains pharmaco-resistant. In this context, there is a need to develop alternative therapeutic strategies. The first step is to improve automatic detection procedures that strongly help diagnosis by providing quantified information about the seizure pattern and occurrence frequency. A still unsolved and challenging issue is whether absence seizures can be anticipated and even be aborted by external stimulation for instance.
Objective. The general objective is to design personalized methods for detection and anticipation of absence seizures based on Artificial Intelligence techniques combined with biosignal processing. Methods. First, novel detection methods will be developed. They will differ from existing methods by i) an improved computation time allowing for real-time processing of biosignals (EEG, EKG) and ii) an account for the specificity of seizure patterns encountered in atypical forms of AE. Second, the possible anticipation of seizures will be investigated and subsequently, methods will be developed to detect a possible “proictal” state. Data. Patient data will be provided by neurology depts. of Rennes and Paris. Neurologists will be part of the project and will provide expertise on clinical data and assessment of results. Computational models of EEG signals will be used to simulate data that will be complement datasets for training of machine learning algorithms. Supervision. LTSI (Research Lab, Rennes) and Verteego (SME, Nantes, Paris) will co-supervise the research work. LTSI has strong expertise in signals processing and modeling of EEG in the context of epilepsy. Verteego has strong expertise in artificial intelligence algorithms and processing of big data.
Candidate profile. The research project is at the interface between bioengineering (signal processing, computational modeling), data science and neurology (epilepsy). The Post-doc fellow will preferably have a background in biomedical engineering, in computer science or in EE with experience in bio-signals (PhD in signal processing, neuroscience or system biology). Knowledge in electrophysiology and/or EEG analysis and/or pattern recognition would be an asset. The post-doc fellow will join a multidisciplinary team including research scientists in biomedical engineering, signal processing, electrophysiology, and AI (ML, neural nets)
The position will be opened Feb. 1st, 2020. The contract is for 1 year with possible renewal (max. 5 years). The competitive salary will be according to experience (2300 Euros net, minimum). The candidate will also have access to the French system benefits.
Location in the city of Rennes, France. LTSI laboratory, University of Rennes. In addition, the post-doc fellow will have the opportunity to perform visits and to actively collaborate with engineers and researchers of the Verteego company (Nantes, 100 km from Rennes).
Contact(please provide resume, cover letter and email of 2 references)
Fabrice Wendling (DR Inserm, LTSI, France, Email: , ).
Rupert Schiessl (CEO, Verteego, France, Email: )
(c) GdR 720 ISIS - CNRS - 2011-2020.