Dear FSL-users,
This message is to inform you that I have found the cause of the strange FEAT results reported earlier.
PROBLEM:
The strange result was that the FEAT difference_contrast was not equal to the difference between activation contrasts at higher level (see earlier post).
CAUSE:
In order to speed up analyses I copied the reg folders for each subject and session in lowest level from an earlier FEAT analysis on the same functional data. However, accidentally, I also copied the reg_standard folders. So the intermediate level FEAT used the incorrect old information from the reg_standard folders, which produced the strange error reported. The error might also be due to the fact that the reg_standard folder was created by of an earlier intermediate level and incorrect FEAT analysis. So an intermediate level FEAT analysis seems not to automatically overwrite the content of reg_standard folders at lower level (which is still quite strange, right?).
SOLUTION:
Removing all reg_standard folders before running the intermediate level FEAT again solved the problems.
Kind regards,
Henk van Steenbergen
|