Dear FSL community,
I am running a higher level analysis in FEAT using a modal hippocampal mask derived from a FIRST segmentation and it has been running for about 3 days now. There are 25 subjects, each with 2 runs. A whole brain analysis on these subjects typically takes less than a day.
One thing that I have done different with this analysis is that during first level processing I only registered the EPI images to the BETed 3D structural. I then used the FIRST derived FLIRT matrices to register the EPIs and 3Ds to MNI. So all of the appropriate files are in the reg directory (including the inverse transforms).
Would this cause FEAT/FLAME to take so long?
The first few temporary stats directories were produced relatively quickly, however, they have been coming out slower and slower. Yesterday, there were only 12 produced. It is at #77 right now, and I believe it has to go until 181.
I have restarted the analysis a couple of times only to encounter the same "problem."
Any idea?
Thanks,
Carlos
|