Hi, I had some experience in analyzing task-fMRI data with AFNI or SPM. I am processing some DTI data, and I’ve had some results in white matter tracts. It is said that fMRI could also provide some information in white matter. So I am trying to verify whether these tracts changed during task-fMRI. But I don’t is there any common analyze process for this situation? And what tool is suitable?
In my opinion, I may first pre-process the task-fMRI data as normal process (Realign including slice-timing and head motion corrections, spatial smoothing), then I will use GLM to extract beta weights for each voxel depend on my task design. Finally, I may apply my white matter tract ROIs to the fMRI result (beta imaging) and calculate the mean beta of each tract on each subject, and perform two sample’s t test for these two group (patient and normal).
However, I don’t know is it reasonable. Could you give me some suggestion? Thanks a lot.
The white matter BOLD signal is an interesting emerging topic these days! I think what you proposed is reasonable, just be sure to not use white matter signal as a regressor (or ACompCor components from a white matter mask). Another interesting route is to look at correlations of voxel time courses within a tract, as these are similarly anisotropic like DWI signal (Functional tractography of white matter by high angular resolution functional-correlation imaging (HARFI) - PubMed). In either case, these signals are usually very small and noisy, so even if you post process everything correctly and there is some kind of signal, your acquisition may not be tuned to resolve it (e.g., something like 7T helps). There are still some questions that are not answered that would help our ability to model white matter BOLD, for example if it follows the same canonical HRF as gray matter, which is of course an important consideration for GLMs.
You may also want to only focus on the core of tracts to mitigate noise and uncertainty near the extremity of bundles.