What will fMRIprep do with this if I only specify the participant’s folder?
What is the recommended way of working with this situation?
By design, task ALeft and task ARight can be considered run 1 and run 2. For now, I am processing them separately with the idea of merging them later. Is this a bad idea?
One option would be to treat these two sessions as “sessions”, like this:
sub-00X
├── ses-01
│ ├── anat
│ └── func
└── ses-02
├── anat
└── func
Fmriprep would merge the two T1w images to create a common T1w that would serve as target for the bold registration for the two sessions (more details here: Processing pipeline details — fmriprep version documentation). Thus the bold images from the two sessions would be registered in the same T1w and later would be normalized to the MNI or any template of your choice. The runs task-ALeft and task-ARight would be processed separately but there will be projected to the same T1w space and template space for later analyses. I would not recommend to merge them. You could analyse them together using a common GLM but noting that there are coming from different runs/sessions.
thanks for answering. Just to be clear, fMRIprep will merge the two T1w only if I specify them as ses-01 and ses-02? I would like to understand better how the makeup label impact the fMRIprep pipeline.
I am going to consider the ALeft and ARight as one task at the first level analyses. I wonder how much of a difference I will get if I consider these two as run-01 and run-02 or keep them separate.
In the case of multiple T1w images (across sessions and/or runs), T1w images are merged into a single template image using FreeSurfer’s mri_robust_template.
It does not make a difference within fmriprep processing if the two T1w are within the same session or in separate sessions, they will be treated the same way and the single template resulting from this merge will serve as the target for the functional registration.
I am not expert in fmri analysis and I can’t help you on the second question.
You’ll want to be careful doing this, as there may be scanner drifts between these runs (especially if the runs happened on different days). If you do go down this route, you should include two columns in your first level design matrix, one that contains 1s on every volume that belongs to task Right, and another that contains 1s on every volume that belongs to task Left.