|
Project Topic
|
During the last decades, diffusion MRI has been massively used to investigate how alterations in brain connectivity relate to various conditions. However, the mainstream tools were focused on the long- range fiber bundles embedded in deep white matter. The superficial white matter (SWM), namely the outermost layer of the brain's white matter, which contains short-range connections between adjacent areas, was out of reach. Technical advances in diffusion MRI acquisition and analysis now make it possible to study SWM. In the context of the Human Brain Project (HBP), the most comprehensive atlas of short bundles has been established in the adult brain (673 bundles), and is available to the community via EBRAINS. In addition, a method for robustly projecting this atlas onto any adult brain has been designed and is available to the community. This method (named GEOLAB) has been applied to a large open database, UKbiobank (27000 aging subjects), in order to provide the community with normative charts of SWM aging regarding the bundle microstructure. SWM is known to be one of the last brain areas to fully develop and mature. Therefore, the study of SWM is likely to provide biomarkers of choice to shed light on the physiopathology and progression of neurodevelopmental conditions. The goal of this project is to extend the SWM-dedicated normative charts developed for aging in the context of the HBP to the entire spectrum of brain development accessible by MRI. Large open databases will be used to build charts dedicated to babies (dHCP), children (BCP, ABCD, HCP-D) and young adults (HCP-YA). These charts will then be tested through the search of biomarkers for early diagnosis and outcome prediction of several neurodevelopmental disorders: two very high risk factors from the perinatal period (extreme prematurity and fetal alcohol exposure) and three major psychiatric syndromes (Autism Spectrum Disorder ASD, Schizophrenia and Bipolar Disorder). For this purpose, dedicated harmonization strategies will be designed to plunge each clinical dataset into the normative charts. Then, for each pathology, Machine Learning will be applied on large aggregated datasets made up of several smaller clinical studies. Most of the data that will be used to achieve these objectives have already been collected and analyzed for other purposes in primary studies. Therefore, they will not require any curation, which will allow us to quickly apply the open tools of the HBP dedicated to SWM. Three of the open databases (HCP- YA, ABCD and UKbiobank) will allow genetic analysis of the SWM phenotype, which will be used for further interpretation of the biomarkers discovered for each developmental condition. The results of the project (lifetime normative charts, genetics analysis and developmental biomarkers) will be uploaded into EBRAINS for further use.
|
