Hi,
The vertex analysis is run as a separate test at each vertex (like the voxelwise analyses in images) and it is only in the randomise step that it takes into account the whole set of tests for multiple comparison correction. Therefore, all you need to do to run your test on an ROI is to provide a suitable mask to the randomise step. You can do this by taking the mask image that you would normally use for the vertex analysis (which is the whole boundary) and restricting that to your ROI by multiplying it (in fslmaths) with an ROI mask. If you then feed in this ROI mask then randomise will only use the appropriate portion of the structure.
All the best,
Mark
On 30 Apr 2014, at 12:59, Mauricio Delgado <[log in to unmask]> wrote:
> Dear FSL group,
>
> We would like to perform FIRST Vertex analysis on subregions of the hippocampus and the amygdala rather than the entire structure, and were wondering whether it would be possible to perform the analysis utilizing a region of interest approach? If so, at which step along the processing pipeline would it be possible to specify our ROI's? Also, could we just draw masks in MNI space and apply these during the analysis as ROI's/masks?
>
> Any assistance would be greatly appreciated!
>
> All the best,
> Mauricio
|