Hello, everyone:
I'm trying to apply voxel-based analysis on FA map using SPM8, but I countered a tough problem in the statistic step. My procedure is as the following.
Firstly, by using DtiStudio, the automated image registration(AIR) realignment is performed to minimize misregistration due to subject motion. Then, tensor calculation produces FA map and B0 image.As a result, FA maps and corresponding
B0s of 4 healthy people and 1 MS patient are aquired.
In SPM8, spacial normalization is implemented on FA maps using the transformation of B0 normalization to EPI template,then use FWHE=[4 4 4] to smooth the FA maps.
At last, I click "Specify 2nd-level", and select two-sample t test, 4 healthy people FA images as the control group, 1 patient FA map as the other group. But when I run the analysis by click "Estimate", there appears an error "Inner matrix dimensions must agreeļ¼"
I wondered if the number of FA maps in the two groups affected the result.
I tried 1 control vs 1 patient, but it warned that "Please check your data: There are no significant voxels"!
I'm so confused about the problem, hoping anyone may help me.
Thanks in advance and really appreciate your help.
Best Regards,
Sophia
|