Figure 3.

A summary of research analysis methods applied to resting-state functional MRI. (a) With seed-based functional connectivity analysis, a voxel or region is predefined and correlations are estimated between the selected “seed” and the remaining brain voxels. (b) An illustration of regional homogeneity (ReHo). (c) An illustration of amplitude of low-frequency fluctuations (ALFF). (d) Principal component analysis (PCA) transforms the original data into a new coordinate system where orthogonal variables are identified while retaining most of their variance. (e) Independent component analysis (ICA) is useful for searching a set of underlying sources of resting-state signals that are maximally independent of each other which can explain the resting-state patterns. (f) Graph theory views ROIs as nodes and correlations between them as the connectivity of the edges and then computes the connectional features of the graph.