Isaac,
That is very impressive!
I guess it's not surprising at all to see the anatomical structure in the residuals either at the individual or group level. Basically there is some significant amount of signal that still has not been properly accounted for. This inaccuracy would have a huge impact on estimating the spatial correlation (FWHM) that is typically used in correcting for multiple testing.
Gang
Edited 1 time(s). Last edit at 12/17/2015 01:20PM by Gang.