Expert in: Unsupervised classification
Professeur agrégé, Professeur associé
- Unsupervised classification
- Nonparametric statistics
- Signal processing
- Image processing
My main interest is to characterize the anatomo-functional architecture of individual brains using neuroimaging data, and in particular using resting-state fMRI. I am also interested in examining how brain connectivity can be used as a biomarker of neurodegenerative diseases such as Alzheimer's disease. These questions raise considerable methodological challenges, which feed the technical aspects of my work. To explore the resting-state networks in fMRI, I use some unsupervised pattern recognition techniques, i.e. various types of clustering and component analysis. To deal with the statistics associated with a stochastic clustering process, I have been working on non-parametric statistical methods, in particular based on the bootstrap. Besides the exploration of real data, my research also includes the development of fully synthetic neuroimaging databases which cover many aspects of the data-generating process, from neural activity and physiological noise to the physics of image acquisition, to provide a test bed for the evaluation and validation of neuroimaging analysis methods.