Hi, AFNI experts,
Here I have a quick question on ROI analysis. According to my understanding, there are at least threee ways to do the ROI analysis. 1). We can get the ROI coordinates (x,y,z) from the fMRI data analysis, say, a t-test (Reward vs. noreward). Let us have the left side Anterior Insula as an example: Left AntInsula, (-28, 20, -1); 2). Based on the classic literature or meta-analysis. Say I also get the coordinates for Left AntInsula, (-28, 20, -1); 3). We can define a ROI based on anatomy using the Talairach atlas provided with the AFNI package. So this ROI is an AntInsula mask.
For the first one case, my ROI analysis steps include
- a radius a 5-mm radius sphere centered on the peak voxel was used (3dcalc step).
- 3dfractionize, using the template from the contrast, e.g., reward vs. noreward.
- Threshold the output from 3dfractionized step, using the 3dcalc.
- Using 3dDeconlve and extract the signal data.
While for the case 2, we only can use the analysis steps 1, 2, 4 and 5. That is to say, we do not do the threshold step. If this is correct, then with the same coordinates, the signal change from analysis approach 1 and 2 would be different.
While for the case 3, further, there is no step 1. That is to say, we only use the steps 2, 4 and 5. Am I correct?
Any comments for my analysis?