Invited Talk by Jonas Richiardi : Learning from neuroimaging and brain genomic data with brain graphs
Learning from neuroimaging and brain genomic data with brain graphs
Modelling interaction between different parts of the brain as a graph has enabled systems-level insights into brain network topology and organisation. The graph view is also a powerful way of teasing out relationships between very different imaging and biological modalities.
Recently, there has been interest in appropriate methods for statistical learning with brain graphs, in particular with applications to prediction for individual subjects in clinical neuroscience (diagnosis and prognosis) and cognitive neuroscience ("brain reading"). In this talk, I will first show a basic pipeline to extract brain graphs from resting-state fMRI data, and a few ways to represent the graphs via embedding so the problem can be cast as a machine learning task. I will then show applications of the approach to diagnosis problems in clinical neuroscience, with results in multiple sclerosis and psychotic symptoms. In the second part of the talk, I will show how the brain graph representation can be used for gene expression data, and present a statistical approach to testing the coherence of spatial relationships between gene expression and fMRI data.
Bio
Jonas Richiardi (
homepage) is currently a post-doctoral fellow funded by the EU under his Marie Curie International Outgoing Fellowship project "Modelling and Inference on brain Networks for Diagnosis" (#299500), jointly affiliated to the FINDlab at Stanford University and LabNIC at the University of Geneva.
Between 2009 and 2012, he was a post-doctoral research fellow in the Medical Image Processing Lab, a joint position between the Ecole Polytechnique Fédérale de Lausanne (EPFL), Institute of Bioengineering, and the University of Geneva's Faculty of Medicine, Radiology and Medical Informatics Department.
