I had a few quick questions regarding the output generated by FEAT statistics. I'm currently working with resting-state data and attempting to perform nuisance regression of CSF, WM, Global Signal, motion parameters (standard + extended), and also scrub volumes that exceed a specific threshold of motion using FEAT Statistics. To scrub specific volumes with excessive motion I generated a confound.txt file that includes columns of 0 each with a single 1 indicating the specific volume that needs to be scrubbed. I selected Standard + Extended Motion Parameters to apply the motion parameters generated during FEAT preprocessing. Additionally, I applied CSF, WM, and Global signal nuisance regressors under full model setup by selecting Custom (1 entry per volume) and including three separate .txt files, each including 1 column of average values per volume (for CSF, WM, or Global). Doing so generated the attached design.png and res4d image. Is this the correct way to perform nuisance regression? If so, does the output res4d image look correct? It is very difficult to see the actual image relative to the background. Furthermore, is res4d the right image that I should be using if my goal is to extract the time series of ROIs within this fully processed resting state data? Any help is very much appreciated! Best, Brandon
Dear Brandon,
The reason your residual map is hard to visualize is that typically the data is demeaned when fitting the GLM. So one strategy to visualize your data as a brain could be to save the mean of the data in time before running the denoising and then adding it back to the residuals afterwards for visualization purposes.
Regarding your denoising strategy, it seems fine to me, even if there are some cons of scrubbing (see https://neurostars.org/t/despiking-vs-scrubbing/2157/9 ). What are regressors 3-6 in your design matrix? They look surprisingly regular to me.
Also, usually you might want to also apply a bandpass filter in this process, which you can do in FEAT.
I hope this helps!
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion