PhD Position in Pretrained Multimodal Neuroimaging Models for Cognition

This PhD position is hosted by the Inria Saclay Center within the MIND team at CEA NeuroSpin on the Paris-Saclay campus, in partnership with the Neurofunctional Imaging Group (GIN) in Bordeaux. The successful candidate will design and benchmark deep generative and self-supervised models using large-scale multimodal neuroimaging databases like the Human Connectome Project and UK Biobank. The work involves building open data infrastructure, predicting individual cognitive outcomes with uncertainty quantification, and evaluating transfer learning strategies. The position expects active collaboration with partner teams, methodological rigor, and publication in top-tier venues.

• Build and release open preprocessing and harmonisation pipelines for large multimodal neuroimaging databases including fMRI, diffusion MRI, and structural MRI.
• Design, train, and benchmark deep generative and self-supervised models of the multimodal brain phenotype to predict cognitive outcomes with calibrated uncertainty.
• Evaluate transfer learning and domain adaptation strategies on small and medium-sized cognitive neuroscience datasets.
• Write scientific literature targeting major venues such as NeurIPS, ICML, MICCAI, OHBM, and journals like Nature Methods or NeuroImage.
• Release models and code openly while interacting closely with the GIN Bordeaux partner team.