Hi Belina

Typically identifying the skull tissue is relatively straightforward if normalization has worked correctly so I imagine that the main reason causing the images not to be properly skull-stripped would be if normalization for those subjects has failed. I would check the logs to see if there are any warning on error messages on those subjects when running preprocessing's normalization step, and if there are no indications of errors/warnings I would probably check whether the original structural images for these subjects (before normalization) look relatively close to the templates to begin with (for example in CONN's Setup.Structurals tab using the option 'structural tools -> display anatomical / MNI coregistration (SPM)'), as often manually reorienting these images to bring them closer to the templates before running normalization can help fix this sort of problems.

Hope this helps

Alfonso

Originally posted by Belina Rodrigues:

Hello, 

I’ve run into an issue during preprocessing: while the pipeline (using the direct normalization) worked well for most participants, it seems that for 3 subjects, the skull-stripping step didn’t complete properly, i.e. the skull remains clearly visible in their anatomical images ( unlike the others, where skull-stripping was successful.)

Any idea why this might happen selectively for just a few individuals? Are there known issues that could cause the skull-stripping to fail (e.g., poor image quality, motion artifacts, unusual anatomy)?

Also, do you have any suggestions for how to fix this? 

Thanks

Belina