I’d like to perform a group ICA on flortaucipir PET images (spatially normalized to MNI152 space, parcellated using 400-ROI Schaefer atlas) to identify different patterns of spatial tau distribution.
I know that the GIFT toolbox can do this for voxelwise data, but what if my data is parcellated and thereby in a vector format?
You can just process the parcellated data similarly as the original voxel-wise data? It’s just a dimension-reduced one.
Thank you @neo ! However GIFT only accepts nifti files as input. I would like to know about solutions for vector data.
I think you can just save your vector data as nifti format?