 |
 |
On Thu, Jun 7, 2012 at 5:19 AM, Javi Cruz <[log in to unmask]> wrote: > I wonder if the BPM can do a correlation analysis between an image and a > SPM t-contrast. That is, I want to correlate two measures: an average image > lesion map from many MS patients, and secondly, a SPM contrast of DARTEL in > which atrophy in gray matter occurs in the patient group compared to HC's. > No. You can't correlated one image with another using BPM. The only thing you can do with two images is to correlate the voxels in one image with voxels in another image. This will give you the correlation of the image, but won't tell you anything about spatial locations. > > I've tried before to do so subject by subject with two flists: one > corresponding to individual lesion maps, and another with the individual > preprocessed anatomical gray matter scans (smwrc1). I wonder if I can do in > the manner explained above and is methodologically correct. > This will provide you a correlation between lesions and atrophy at each voxel, which is probably what you want to know. When you do the analysis, I would call it a t-test as you are comparing the lesion group to non-lesioned group at each voxel. This analysis is circular though. You have already stated that a subject has a lesion at a location, thus they must have less gray matter at that location. > > Anyone can help me? Thanks >