Contact: matthieu.gilson [at] univ-amu.fr
The study of brain as a complex system is an area where the cross-fertilization of neuroscience and advanced machine learning contributes to the rapid development of both fields. Magnetic resonance imaging (MRI) is a non-invasive technique that allows for accessing both brain anatomy and brain activity, corresponding to structural and functional data (sIRM and fMRI, respectively). A whole line of research has been developed over the past two decades to study the structure-function relationship, aiming to understand the flexible task-specific communication between brain regions. This involved the characterization of spatio-temporal structure of brain activity [Hutchison 2013]. Recent advances in deep learning have led to the emergence of foundation models capable of generating complex data across many contexts, like large language models (LLMs) that can translate concepts across languages. Although such models are still in an early development phase for brain activity, they offer a promising way to explore the brain activity and pathological alterations. The objective is to test and develop generative models in coordination to data augmentation to enhance model-based representation of time series that generalize across subjects and/or conditions (e.g. cognitive tasks).
Refs:The multidisciplinary supervision panel has complementary expertise covering the fields of computational neuroscience, machine learning and medicine:
We are looking for a profile from computational neuroscience, engineering or physics with excellent programming skills in python and experience in autodifferentiation frameworks.
The PhD will be held as a collaboration between Emmanuel Daucé (http://emmanuel.dauce.free.fr/) from the engineering school Centrale Mediterranee and Matthieu Gilson (https://matthieugilson.eu/) from Institut de Neurosciences de la Timone (INT). Centrale Mediterranee (https://www.centrale-mediterranee.fr/) aims to form new generations of innovative engineers and managers, who are open to international cooperation and able to tackle the increasing complexity faced in the present world. This project corresponds to the modeling axis of Centrale Med. INT is a laboratory dedicated to the study of the structure and functioning of the brain and spinal cord. INT conducts basic, translational and clinical research on a major cognitive functions. It hosts the Computational Neuroscience Center (CONECT, https://conect-int.github.io/), an incubator to promote theoretical and computational neuroscience.
Send the following documents to the email contact address. Candidates will be interviewed for 30 minutes, including a presention of themselves and their scientific trajectory (10 minutes).