I selected hippocampus and amygdala as ROIs to estimated the reliability of a well-controlled multi-site study. The travelling volunteers were scanned in each site with the same type scanners and scan protocol. But, amygdala volume (generated by FreeSurfer) presented a relatively poor ICC scores between sites. So, i want to found the reasons from sites or scanners. According to previous studies, the variation might come from across-site difference and the structure itself. And the major sources of between-site variation might be related to field inhomogeneity and other issues idiosyncratic to individual each scanners.

Due to this reason, i want to check my data to confirm the field inhomogeneity located in amydala area or not. How can i get the difference between the originate image and the images after motion and NU intensity?