Dear Chris,
> Here are examples of the results:
> http://www.stanford.edu/~cdwhite/roi-histograms.jpg
Thanks for posting the URL, my machine has been getting bloated with large
email attachments of late. Ah GIMP, redhat.
The histograms look rather odd. In general they look normal +/- skew.
There seems to be a large variation in the number of voxels within the ROI
between subjects, and in some subjects, there are a lot of zeros - presumably
meaning that some of the ROI voxels were estimated to be outside the brain
using whatever thresholding/masking you had in the analysis. Your last
subject seems very different from the others, with a large shift to more
positive Z scores, but smaller N.
How did you determine the ROIs? Why is there such a difference in number of
voxels between subjects? Is the ROI near the edge of the brain - and nearer
the edge in some subjects than others?
I suppose you might also consider using the con*.img activation images instead
of the Z scores to give you a measure more directly related to average signal
change.
Regards,
Matthew
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|