Detecting low-frequency functional connectivity in fMRI using a self-organizing map (SOM) algorithm

Low-frequency oscillations (<0.08 Hz) have been detected in functional MRI studies, and appear to be synchronized between functionally related areas. A current challenge is to detect these patterns without using an external reference. Self-organizing maps (SOMs) offer a way to automatically group data without requiring a user-biased reference function or region of interest. Resting state functional MRI data was classified using a self-organizing map (SOM). Functional connectivity between the left and right motor cortices was detected in five subjects, and was comparable to results from a reference-based approach. SOMs are shown to be an attractive option in detecting functional connectivity using a model-free approach. (C) 2003 Wiley-Liss, Inc.