Spatial Correlations Induced During fcMRI Preprocessing Decomposed into Second-Order Temporal Frequencies Bands

• Activity Number: 354
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Mary Kociuba*+ and Daniel Rowe
• Companies: Marquette University and Marquette University
• Keywords: fMRI ; image processing ; functional connectivity

Abstract

To remove non-biological variability inherent within an acquired functional connectivity MRI (fcMRI) signal, the data is preprocessed before spatial and temporal correlations are estimated. Although, it is well known that preprocessing induces correlations, for example, applying a low pass spatial filter artificially increases spatial correlations among neighboring neural regions. The theoretical spatial correlation matrix can be represented as a linear combination of second order voxel temporal frequencies, such that the spatial correlations arise from overlapping temporal frequency content. This spatial correlation representation provides a means for quantifying the impact of preprocessing on a voxel temporal frequency spectrum, through identifying the temporal frequency bands that contribute significantly to induced correlations of no biological origin.