Dear SPMers:
I am using High dimensional warping (HDW) to get the jacobian determinants between two T1 volumes (subject's visit1 & visit2). Out of curiosity, I wonder how the image intensity would affect the result of jacobian determinants from this tensor based approach. So I tried to multiple one T1 volume with a random coefficient, for example .*10 with imcalc. And I did a HDW without any manipulation of these T1 volumes. The jacobians are very different with/without such image intensity change. This makes me wonder how the image intensity would affect the jacobians since most our scans would not get the same mean intensity (due to scanner's receiver gain variations). Should we do some intensity normalization before the HDW steps? The previous papers using TBM did not seem to include such step. Is it necessary?
Thanks if anyone could shed some light on this!
Guofan
11.3
Guofan Xu
Department of Medicine
Geriatrics & Gerontology Section
Veterans Administration Hospital
2500 Overlook Terrace
Madison, WI 53705
|