Dear SPMers We have a weird problem with the paired-t-test. 1) We get the attached error message during estimation when including a subset of our subjects, but not for other subjects. 2) The same contrasts can be used in a 1-sample t-test. 3) The contrasts from the problematic subjects are not different from the "fine" subjects in terms of having significant voxels at the 0.001 threshold (there are some empty maps and some non-empty maps in both subgroups). We would appreciate any help Tali Bitan Haifa University ..... Error running job: Error using ==> spm_spm at 861 Please check your data: There are no significant voxels.i In file "/matlab/spm5/spm_spm.m" (v946), function "spm_spm" at line 861. In file "/matlab/spm5/spm_config_fmri_est.m" (v832), function "run_est" at line 394. -------------------------- Done.