Hi all,
I made adjusted means images using adjustedmeans-fmri in SPM99 in order to
use them in an ROI analysis. I found that the percent signal change for
the group was not significantly different from 0 even for the contrast
that I used to define the ROI. The ROI was voxels that were significant
in the group level one sample t test using con images.
I verified that there was nothing wrong with my procedure to get
values for the ROI for each person by doing it on the con images - it
yielded the logically required result.
I made both the con images and the adjusted means images both with
and without any high pass filter (when I used one, I used the same one)
and got the same discrepancy. I have always chosen no global scaling in
both cases.
Does anyone know why I might be getting different qualitative results with
con versus adjusted means images? Do they implement default global scaling
or convolving with the boxcar differently?
Thanks,
Joanna
Joanna Salidis, Ph.D
Psychology Department
Stanford University
Jordan Hall
Stanford, CA 94305
Phone: 650-725-0797
Fax: 650-725-5699
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|