Statistical Analysis of Neuroimaging Data: Past Methods, Current Challenges, and Opportunities

• Activity Number: 358
• Type: Roundtables
• Date/Time: Tuesday, August 5, 2014 : 12:30 PM to 1:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): DuBois Bowman*+
• Companies: Columbia University
• Keywords: Imaging ; high-dimensional data ; prediction ; biomarker ; fMRI

Abstract

There have been major advances in the statistical analysis of neuroimaging data over the last two decades. Yet, due to inherent complexities and the massive dimensionality of the data, many needs remain for novel statistical development. Moreover, there are still enormous opportunities for neuroimaging to affect science and medicine, but these advances depend in part on our ability to establish sound estimation, inference, and prediction frameworks. We will briefly discuss existing approaches to analyzing neuroimaging data and highlight remaining challenges and opportunities. Among other topics, this roundtable will address prediction methods, Bayesian modeling, the search for neuroimaging biomarkers, multimodal imaging, combining imaging with genomics and other high-throughput data, and challenges in resting-state fMRI.