Hi there,
I have some questions pertaining to the statistical analysis portion of VBM analysis. I am looking at grey matter differences between patients with dementia and a health control group. It is quite established in the literature that there are GM differences between these group and these differences are readily apparent when we visually inspect our anatomical scans.
I have been following the 2015 manual available online and I have successfully produced the VBM images. When I followed the statistical analysis section, I did not get any group effects which is not consistent with what we had expect or the literature. I wanted to double check to ensure that the steps I followed are correct.
(1) I wanted to use total intracranial volume as a covariate. In the model, do you suggest I include TICV as a covariate or as part of the global normalization step?
(2) I calculated TICV following the manual’s suggestion (summing CSF, GM and WM). The resultant volume is in liters. Should I convert these measures in mm3 to be consistent with the voxel dimensions
(3) The type of masking recommended in the manual is absolute masking with a threshold of 0.2. Is this type of masking still okay to be used with my research question?
(4) Design= I had selected a t-test, though would you recommend regression?
Thank you very much in advance for your help!
Best,
Tamara
|