Measuring the digital brain

Before the advent of magnetic resonance imaging (MRI), the shapes of brain convolutions were inaccessible to living patients, except in the very specific context of neurosurgery. As a result, they did not arouse much interest in the medical world. The 3D anatomical images now routinely produced by MRI on millions of individuals have made it possible to realize the considerable variability of these patterns from one individual to another. This variability has long been considered a nuisance in the world of functional brain mapping, but algorithmic advances aimed at quantifying it are now turning the tables. These patterns, which emerge during brain development, provide information about the underlying architecture and are likely to retain traces of disruptive events. The first part of this talk will describe a research program aimed at modeling and quantifying cortical wrinkling so as to exploit it to produce biomarkers.
But measuring the digital brain from MRI goes far beyond quantifying macroscopic shapes. Imaging the molecular agitation of water molecules, known as diffusion MRI, gives us access to the organization of the fibers that make up the brain's wiring. The term "connectome" is now used to describe the network of elementary brain entities linked by bundles of fibers. The connectome of the human brain is in fact little-known, as it is virtually inaccessible by dissection. The second part of the presentation will describe an approach for modeling all the major bundles that make up the brain's skeleton, and for quantifying any anomalies in these bundles.