The human brain is the most variable of all primates. Understanding this normal variability should allow us the better detect abnormal variation in cases of brain disorders. Our aim is to analyse, describe and understand the patterns of structural/functional diversity, to produce statistical models of this variability, and to link it to genetic variability, brain development and evolution. Mental disorders are a major source of years of disability in our societies. The aim of the project we propose is to better understand the normal variability of the human brain, which should facilitate the detection of pathological deviations. It is then a project related to health and of public interest. We will use the phenotypes already computed by UK biobank, and re-process all the data to complete a larger set of neuroanatomical phenotypes such as regional volume, cortical thickness and grey/white matter contrast. In particular, we will use our in-house tools to compute a series of novel global and local folding measurements. We will use multivariate regression to produce a statistical model of neuroanatomical variability, taking into account covariates such as age and sex. Our code will be made available on GitHub (no individual-level data will be shared). We aim at including the full cohort, currently N~5k subjects, N~95k in the 6-7 years to come as data becomes available.
