Big Data Era in Magnetic Resonance Imaging of the Human Brain

Magnetic resonance imaging (MRI) has opened a non-invasive window into the human brain over the past few decades. MRI studies, specifically functional MRI (fMRI) studies, consist of hundreds of timepoints collected from O(10⁵) to O(10⁶) spatial locations and tens to hundreds of subjects. Thus, even relatively modest studies consist of O(10⁹) data points. So, practitioners in this field have been working with big data well before that term became popular. However, it is only recently, with the advent of faster CPUs and GPU processing, that techniques associated with that term, namely machine learning, have become popular in the field. In this chapter, we review MRI and the types of signals one can measure with it. We then describe some “big data” initiatives in the field. Finally, we review some of the studies that have applied big data approaches to neuroimaging data and discuss some of the issues associated with those types of studies.