Implications of Matrix Decomposition Methods in Analyzing Imaging Data

• Activity Number: 481
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Imaging
• Author(s): Ani Eloyan*
• Companies: The Johns Hopkins University
• Keywords: ICA ; functional MRI

Abstract

Matrix decomposition methods are often used in analyzing brain imaging data to represent the data parsimoniously, learn about the structure of variability in the data, dimension reduction, etc. In group-level analyses dimension reduction techniques can provide low dimensional biological information that can be used in a secondary analysis to obtain results on group differences. In this talk, a combination of dimension reduction via the Independent Component Analysis and further modeling of the results is discussed to identify differences of brain connectivity in children with autism spectrum disorder with their typically developing peers. The analysis focuses on the motor function and the visual function of the children. The results for a set of 100 children in the study are presented as well as a replication study based on a larger dataset.