Section: Overall Objectives

Overall Objectives

The research domain of Parietal is the design of statistical analysis and machine learning tools for the analysis of neuroimaging data. Such data have a limited signal-to-noise ratio, and generally display weakly contrasted structures. More importantly, due to the cost and complexity related to acquisition and analysis process, the estimation of reliable models is plagued by the lack of data, leading to little statistical power.

We try to solve these challenges by combining different modalities, by improving the crucial registration steps and by designing tools to assess efficiently the structure of the data. We design adapted learning algorithms to classify and make inference on these images; we also develop a framework to carry out statistical analysis in high-performance frameworks, such as clouds and grids.