UCSF

Imaging biomarkers from magnetic resonance images have provided insights into the progression of multiple sclerosis (MS). As neuroimaging datasets grow in size to accommodate multidimensional association studies, traditional methods for data collection and analysis are too imprecise and inefficient on a large scale. This dissertation addresses the challenges associated with collecting datasets from multiple scanners with non-standardized acquisition protocols, presents software for time and space efficient image processing, and software for collaborative quality control. Finally, a visualization framework is proposed to gain better intuition and understanding of high-dimensional imaging datasets through a web-based interactive data exploration tool. This dissertation lays the groundwork for large, multivariate studies in MS, and translational tools for the use of imaging biomarkers in the MS clinic.