Thanks for your guidance. Bit you said the target affine must be the affine of the searchlight scores. Does AAL.affine work?
On the other hand the searchlight’s scores are in the form of an array. How can I get it’s affine?
Try plotting the resampled AAL image over your searchlight and see if the overlap is good.
You can also try nilearn.image.resample_to_img - Nilearn, using 'nearest' for the interpolation method.
Both resample to image or resample image give the same output. For my case, the number of ROI stays the same i.e. 117 even after resampling. Just as @Steven mentioned , ensure you use “nearest” for the interpolation. Another thing to note, resample img or resample to image expects input image that has been registered to the MNI template[nilearn.image.resample_to_img - Nilearn]. Also note that the affine is the target affine as per the documentation and that works well for me as well. Thanks again @Steven for your response.
