Re: [Freesurfer] Reliability of labeling procedure in the same scan

2017-09-05 Thread Mike Schmidt
I've run many hundreds of scans repeatedly and found variance between versions of FreeSurfer, but identical output within a version. Something must have changed between runs. Perhaps you could try comparing the log files and see where the deviations begin. Mike Schmidt On Sep 5, 2017 10:54 AM, "

Re: [Freesurfer] Reliability of labeling procedure in the same scan

2017-09-05 Thread Bruce Fischl
no, I don't think so. There is some randomness in the algorithm (e.g. sampling paths through the volume), but I thought we used the same random seed each time to ensure repeatability. Was it on the same machine? What was your command line? And How big were the differences? On Tue, 5 Sep 2017,

Re: [Freesurfer] Reliability of labeling procedure in the same scan

2017-09-05 Thread Phoolan Devi
I mean the hippocampal volume in the aseg... so, it's not normal? Bruce Fischl Sat, 26 Aug 2017 11:24:44 -0700

Re: [Freesurfer] Reliability of labeling procedure in the same scan

2017-08-26 Thread Bruce Fischl
do you mean the hippocampal volume in the aseg or in the subfields? On Sat, 26 Aug 2017, anonymous anonymous wrote: Hi all, Without noticing I run recon-all twice on the same scan of the same subject and we noticed there are some differences, at least in the hippocampal volume estimation. Is