Meeting Banner
Abstract #0235

Component Based Noise Correction for Perfusion fMRI

Restom K, Behzadi Y, Perthen J, Liu T
UCSD

Perfusion based functional magnetic resonance (fMRI) utilizing arterial spin labeling (ASL) techniques has several potential advantages over traditional blood oxygenation level dependent (BOLD) based imaging, such as 1) better localization of neural activity and 2) potentially less inter-subject variability. However, the ASL signal exhibits low signal-to-noise and is often confounded by physiological noise sources. Here we develop and characterize a component based algorithm (CompCor) for the identification and removal of structured noise elements in perfusion based fMRI. We show that CompCor removes respiration induced noise and significantly increases ASL sensitivity.