Hi all,
I'm assissting with an aging study and a good handful of their scans are coming up with watershed errors. My assumption was errors with the skull strip leading to empty brain volumes. I have read through all the lists and tried the following:
- Manually corgistering to MNI to improve Talairich alignment; Same error - Using a range of -wsthresh (from 30-55) with the -no-wsgcaatlas -wsthresh command in recon all; these all had very bad skull strips but avoided watershed error. They all crashed later on at register_mri: find_optimal_transform - Rescaling the voxel intesity ranges to be closer to "successful" scans (watershed error again)
I'm not entirely sure what to try next. I'm using Magnetic Transfer Images that I've multiplied the voxel intesities by 1000 (as normally 0.6-1.5) in order to freesurefer (which works well in most subjects). Other images I have available are T1, T1w, PDw,MTw. I'm avoiding the weighted scans due to coil induced bias. I have tried the T1 with identical results.
WM values=~1800-1900 GM Values= 900 CSF = 50 Out of brain ~300
Cheers for any help you can provide,
- Chris