We are trying to run two subjects through recon-all. The kicker is that these two have very large ventricles and we have been stuck at mris_fix_topology (the script reports about 100-ish defects for each). We use HPC to run the analyses but even with a walltime of 96 hours each, both subjects were unable to be finished within that time frame. My question is that is there any way to improve mris_fix_topology runtime?

Subject 1:

Subject 2: