I have been a CNRS Research Engineer since 2006. At present, I am the technical coordinator of the NIT (Neuroinfomatics and Information Technology) core facility at INT. I’m also a member of two research teams at INT : MECA and BANCO

My research focuses on developing and applying machine learning tools for new problems encountered in neuroscience. For this, I collaborate with the Qarma team of the Laboratoire d’Informatique Fondamentale de Marseille. At the moment, my main interest is to understand, model and exploit inter-individual differences measured through multi-modal MRI to better understand brain functions, using a generic framework called inter-subject learning. For this, I am promoting the development of open source code and the sharing of open data.

Recent highlights :
 2020 : Akrem Sellami’s paper on deep learning fusion of multi-modal MRI data to improve predictive models of individual differences has been accepted to IJCNN !! Find the paper here
 2019 : our ShareElec project has been funded as part of the Open Science program of the French ANR ; we aim at contributing to the standardization of the organization of neurophysiological data and metadata ; stay tuned !
 2019 : Qi Wang’s paper has been accepted to NeuroImage : Inter-subject pattern analysis : A straightforward and powerful scheme for group-level MVPA. You can get the pdf here.
 2019 : I co-organized (and hosted at INT) the INCF-sponsored GEANT workshop, dedicated to data management practices in neuroscience
 2018 : our multi-modal MRI data set InterTVA is online on the OpenNeuro portal ; it is aimed at studying inter-individual differences in voice perception and benchmarking multi-view machine learning algorithms
 2018 : Qi Wang’s paper accepted to PRNI workshop : Population Averaging of Neuroimaging Data Using Lp Distance-based Optimal Transport
 Two papers that include new applications of inter-subject learning : the first one (2017) to quantify the power of EEG features in newborns to predict language acquisition (Enhanced Neonatal Brain Responses To Sung Streams Predict Vocabulary Outcomes By Age 18 Months) and the second one (2018) to demonstrate structure-function relationships (Anatomo-functional correspondence in the superior temporal sulcus)
 2017 : yet another demonstration of the usefuleness of our graph-kernel approach to design predictive models, this time for diffusion-MRI : Learning from Diffusion-Weighted Magnetic Resonance Images using graph kernels
 2016 : a new paper using our graph kernel has been accepted to Medical Image Analysis : Structural graph-based morphometry : A multiscale searchlight framework based on sulcal pits
 2015 : I have, finally, defended my PhD thesis in machine learning applied to neuroimaging : A multi-source perspective on inter-subject learning. Contributions to neuroimaging.
 2015 : our ANR project LIVES (Learning with Interacting ViEwS) has been funded ; I am the coordinator of the INT team and we will design and apply multi-modal machine learning methods on MRI data (anatomical, structural, resting-state, task-functional)
 2014 : our graph kernel paper has been accepted in Plos ONE : Graph-based inter-subject pattern analysis of fMRI data ; the source code and dataset are available here
 2013 : release of Vobi One, the first open source software suite dedicated to the processing of functional optical imaging ; the paper is available here
 2013 : co-PI of the PEPS BMI CAMPA (CAractérisation Multimodale des déficits sensori-moteurs Post-AVC) project ;
 2011 : organizer and chair of the MLNI workshop, which gathered leading researchers in machine learning and neuroimaging ;
 2010-2012 : PI of the Neuro-IC GRABBR (GRAph-Based Brain Reading) project ;
 2009 : publication of our PNAS paper on ICA and fMRI, which includes some the work I did during my six years at Princeton University.