Postdoc (M/F) in Physics-Informed Machine Learning for EEG neuroscience data

Organisation/Company

CNRS

Department

Institut Franche-Comté Electronique Mécanique Thermique et Optique – Sciences et Technologies

Research Field

Biological sciences
Computer science
Mathematics

Researcher Profile

First Stage Researcher (R1)

Country

France

Application Deadline

1 Apr 2024 – 23:59 (UTC)

Type of Contract

Temporary

Job Status

Full-time

Hours Per Week

35

Offer Starting Date

1 May 2024

Is the job funded through the EU Research Framework Programme?

Not funded by an EU programme

Is the Job related to staff position within a Research Infrastructure?

No

Offer Description

The Neuro research group (https://neuro-team-femto.github.io ) of the Department of Automation and Robotics, Institut FEMTO-ST in Besançon, is looking for a postdoctoral researcher in the field of learning dynamic systems and physically informed neural networks (PINNs), for application to neuroscience research.

The main task of the postdoctoral fellow will be to develop models for modeling the dynamics of neurophysiological activity in comatose patients, with the aim of establishing physical biomarkers of the probability of awakening.

This mission is based on an already available dataset, collected in collaboration with GHU Paris Psychiatrie et Neurosciences and comprising several hundred coma patients, combining EEG resting-state data on day 4 after the onset of coma, and clinical data on neurological outcome on awakening (https://onlinelibrary.wiley.com/doi/full/10.1002/acn3.52000 ). The prediction of waking state after coma has recently been the subject of intense research activity in machine learning (e.g. https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2… ), but existing methods have not yet explored the characterization of the system as a dynamic system, nor the use of such physical parameters as biomarkers.

Scientifically, the methods explored may draw on the recent state of the art in symbolic regression (e.g. SINDy, Sparse Identification of Nonlinear Dynamics; Brunton et al.), Neural ODEs (Chen et al. NeurIPS 2018) or dynamical system reconstruction using recurrent neural networks (Durstewitz et al. 2023).

The activities of the person recruited will be:
– a research activity in artificial intelligence applied to neurological data, including the development of models in Python, the evaluation of their performance on data, and the analysis of the predictive power of model parameters
– publication and dissemination of scientific results, in the form of international journal articles and conference participation
– scientific mentoring of the team’s youngest researchers (currently 4 PhD students on related subjects)
– participation in the scientific life of the institute (e.g. seminars)

With nearly 750 researchers, FEMTO-ST is one of France’s largest CNRS engineering science laboratories. It is located in Besançon, a regional capital on a human scale, regularly ranked among France’s top cities for its quality of life and student life. The Neuro research group (https://neuro-team-femto.github.io ) of the Department of Automatic Control and Robotics, Institut FEMTO-ST in Besançon, is an interdisciplinary research group at the crossroads of automatic control, dynamic systems and cognitive neuroscience. It is led by JJ Aucouturier (CNRS Research Director), and currently includes 2 permanent staff, 2 post-docs and 4 PhD students. https://neuro-team-femto.github.io/ The post-doc will be co-supervised at FEMTO-ST by Dr JJ Aucouturier and Dr Noura Dridi, and will interact with Sarah Beghanem, M.D., critical care intensivist at GHU Paris Psychiatrie et Neurosciences in Paris.

Requirements

Research Field

Biological sciences

Education Level

PhD or equivalent

Research Field

Computer science

Education Level

PhD or equivalent

Research Field

Mathematics

Education Level

PhD or equivalent

Languages

FRENCH

Level

Basic

Research Field

Biological sciences

Years of Research Experience

None

Research Field

Computer science

Years of Research Experience

None

Research Field

Mathematics

Years of Research Experience

None

Additional Information
Eligibility criteria

The expected skills are those of a high-level researcher in the field of learning dynamic systems, physically informed Neural Networks (PINNs) or system identification in automation:
– a good academic knowledge of the field, evidenced by a PhD thesis in computer science, physics or computational biology/neuroscience
– good programming skills in Python, development and evaluation of machine learning models, as evidenced for example by code-sharing activity on github/lab or other equivalent indicators
– good scientific communication and article writing skills, evidenced by an ongoing publication activity in international journals
– an appetite for biological and neuroscientific data modeling in particular

Website for additional job details

https://emploi.cnrs.fr/Offres/CDD/UMR6174-JEAAUC-009/Default.aspx

Work Location(s)

Number of offers available

1

Company/Institute

Institut Franche-Comté Electronique Mécanique Thermique et Optique – Sciences et Technologies

Country

France

City

BESANCON

Geofield

Where to apply

Website

https://emploi.cnrs.fr/Candidat/Offre/UMR6174-JEAAUC-009/Candidater.aspx

Contact

City

BESANCON

Website

https://www.femto-st.fr

STATUS: EXPIRED