ROI ISC Analysis
The Inter-Subject Correlation dialog can be used to analyze correlations across provided VTC files also for specified volumes-of-interest (VOIs) that can be loaded using the Load VOIs button (see arrow in screenshot below).
In case one has already performed the same analysis at the (whole-brain) voxel level, it might be useful to check the Run only for VOIs option (see rectangle in screenshot above), which will skip the full VTC-level calculation. If this option is off (default), both the voxel-level and ROI-level ISC analysis will be performed.
After clicking the GO button, the obtained ISC results are shown in matrix plot windows, one window per included VOI. The numbers along the x and y axis indicate the index of the functional file as they were entered in the ==Volume time course data ==list of the Inter-Subject Correlation dialog. If enough space inside each pairwise matrix cell, the calculated correlation value is shown on top of the cell's color also indicating roughly the magnitude of the correlation in relation to the color bar shown on the right side.
The screenshot above shows the matrix plot for the first VOI in the example, which was extracted from a spherical region in higher-visual cortex of the right hemisphere. The screenshot below shows the result for the second included VOI, which was specified in a right-hemisphere region of the prefrontal cortex. While pairwise correlation values are, as expected, positive in both cases, the correlation values in the visual cortex are substantially higher than those in prefrontal cortex. This result can be expected since the data used in the example is from a movie watching experiment.
When clicking on a matrix plot, the Matrix Plot Options dialog appears that can be used to load another color look-up table (LUT) file. Furthermore the numerical values of the calculated pairwise correlation matrix can (also) be inspected in a table when clicking the Data Table button; the appearing Matrix Data Table dialog can also be used to save the pairwise correlation values to disk for further processing.
Copyright © 2021 Rainer Goebel. All rights reserved.