Sébastien joined the Brain-Inspired Computing team at IBM Research Australia in Melbourne at the end of 2019 to work at the interface of computational neuroscience and machine learning, tailored to the medical needs.
From 2016 to 2019, Sébastien was a postdoctoral researcher in the group of Multiscale Brain Modeling at IBM Research in Yorktown Heights, New York, USA. Under the supervision of James Kozloski, he worked on a large-scale brain model incorporating cortico-cortical, thalamo-cortical and striato-pallido-thalamic dynamics to understand and reproduce a cognitive learning task in-silico, using the IBM Model Graph Simulator.
During this time in Yorktown, he also analysed rodent electrophysiology to reveal cortico-striatal cross-frequency coupling alterations associated to mouse models of Huntington’s disease, performed a sensitivity analysis of the connectome harmonics framework, a novel neuro-imaging method based on MRI and dMRI, and co-supervised undergraduate and graduate interns on programming projects and the analysis of brain signals using graph spectral theory.
Towards the end of his assignment, Sébastien enlarged his skillset to machine learning, applying classical classifiers and deep learning methods to discriminate between electro-encephalogram (EEG) traces recorded from several brain regions during trans-cranial magnetic stimulation (TMS).
Prior to IBM, Sébastien did a PhD at the Institute of Systems Neuroscience in Marseille, France, under the supervision of Viktor Jirsa and Christophe Bernard, working on Multiscale Modeling of Epileptic Seizure Dynamics using concepts from dynamical systems theory. His master’s thesis, Computational Agent Modeling of Post-Traumatic Stress Disorder was conducted under the supervision of Jan Treur at the Vrije Universiteit, Amsterdam, The Netherlands, in the Artificial Intelligence department.