Automatic detection of spatiotemporal propagating patterns in BOLD fMRI of the rats using an ICA based approach

Background and purpose: Low frequency fluctuations (LFFs) in resting state fMRI have been used to map functional connectivity (FC) in humans as well as animal models 1, 2, . Previous work has reported the presence of automatically detectable spatiotemporal patterns in LFFs in humans as well as rats 3, 4 . The iterative automatic detection method proposed by Majeed et al. (referred to as iterative pattern-finding method, or IPFM henceforth for brevity) relies on manual delineation of an ROI for detection of each pattern and requires repetition of the process with random initializations . In this article, we present an independent component analysis (ICA) based approach for detection of propagating spatiotemporal patterns in LFFs. This approach allows detection of multiple propagation patterns using a one-pass analysis, without the need for ROI specification and repetition with random initial conditions.