Can't warp masks to individual space

Submitted by aperry on Fri, 07/26/2013 - 10:06

 Hello,

I am performing resting-state analysis on a elderly population. So I have to make sure the masks (ROI, Brain mask etc) fit to each subject. I don't want to use DARTEL, but rather have a subject-specific template.

However, I am having trouble warping the masks into the subject space. The masks either cut out part of the cortex of the functional image, or are deformed. Because of this, I will not get the full number of parcellations as defined in my ROI mask. If I calculate in MNI space without warping, the default masks are too big for the subjects brain, so the ROI mask includes non-brain space.

The coregistration between the T1 and FunImg are fine, as is the normalisation (see attached). I have attached screenshots of the parameters I used (according to the original space template), the T1 image after segmentation (I presume in MNI space), the mean functional image (wmeanaimg etc), and both the ROI mask and Brain mask after being warped.

I presume that I can always extract the ROI time series myself, If I can warp the ROI image into individual space (using FSL or ANTS), and then use the rest_ExtractROITC_gui script?

Thanks,

Alistair 
 

AttachmentSize
Image icon wmeanaimg001.png106.55 KB
Image icon t1mni.png129.39 KB
Image icon ROI.png75.11 KB
Image icon brainmask.png69.2 KB
Image icon 5184A.png117.69 KB

Submitted by YAN Chao-Gan on Thu, 08/01/2013 - 01:52

Hi Alistair,

Actually, DARTEL is using group-specific template. DPARSFA provide efficient way to warp your masks from MNI back to original space.

If you want to use your own procedure, please make sure the affine matrix is the same for your funcitonal data and warpped masks, then you can use REST Extract ROI Signals.

Best,

Chao-Gan