Low-frequency Functional Connectivity in fMRI - Biomedical Signal Analysis: Contemporary Methods and Applications - page 271

(a) (b)

Figure 9.4
Distance matrices with distances represented by gray values, if N =25clustersis
used for the analysis of subject #1 in the resting state experiment. Distances are
determined based on the correlation method (a) and the prediction error method
(b). The upper and lower rows show the matrices before and after applying the
TMP algorithm, respectively. The dissimilarity matrices were plotted such that the
rows from bottom to top and the columns from left to right correspond to
increasing indices of the PTCs. The block-diagonal structure of the ordered
distance matrices becomes evident. The dark lines represent the cluster borders and
are overlaid on the distance matrices. Small distances are plotted dark, representing
close proximity.

represent a quantitative evaluation of this observation and are shown
in table 9.1. For all applied methods, they range within the interval [0.5,
0.6], showing a fair agreement. It should be noted that the novel TMP