Segmentation and multi-variate analyses in large-scale datasets
Given the increasingly large number of neuroimaging datasets available, novel developments in image analysis are critical for ascertaining relationships between neuroimaging measures and behaviours. Here, I will discuss three major developments from our group. The first is the development of an image segmentation pipeline devoted to the detailed assessment of noncortical structures. The second is the development of multi-modal integration techniques that permit the analyses over several metrics simultaneously using matrix factorization techniques. The third is the use of machine learning to elucidate prognosis at the individual level.