Interdisciplinary Institute in Artificial Intelligence

DESCRIPTION

In recent years, modeling the brain’s functioning as a network has provided new insights into the development of conditions such as neurodegenerative diseases and disorders of consciousness, particularly at the individual level. These networks consist of a set of brain regions that interact by exchanging information packets and can be classified based on their role within the networks—for example, as hubs or communities. In the case of neurodegenerative or neuropsychiatric diseases, identifying key regions as hubs has clarified the diagnosis and progression of these conditions. Neuroimaging facilities are now available in most hospitals, making it possible to observe brain activity during tasks or at rest. However, this generates a massive amount of data that can be difficult to process and analyze adequately to obtain reproducible results. Robust statistical analyses are needed to achieve these goals, given that patient groups are often heterogeneous, brain structure is well documented, and the data acquisition process has specific physical properties. Most recent studies are based on group analysis, and statistical methods typically do not account for the spatiotemporal dependence within the data. The objective of this project is to develop methods for the robust and statistically consistent estimation of networks using functional data analysis for multivariate datasets such as those observed in neuroimaging. The complementary expertise of the five partners in this project will enable the design of statistical methods tailored to the context of multivariate time series and robust to the specific types of noise observed in brain neuroimaging.

ACTIVITIES

Two PhD students and one postdoc are joining the team.