Hi,

According to the example given in the 3dANOVA help file, it is possible to run 3dANOVA on data from a single subject "fred" across three levels of a factor. It would appear from the example that there is a single dataset for each factor level (faces, houses and donuts).

I have an analogous situation of one subject's data across 2 levels (easy and hard; each of which is the concatenation of 2 normalised runs) and using either a paired t-test (more appropriate for two levels) or the anova:

3dANOVA -levels 2 -dset 1 nwhere1_12+tlrc -dset 2 nwhere2_12+tlrc -ftr WHERE -mean 1 easy -mean 2 hard -diff 1 2 evh -contr -1 1 hard_easy -bucket anova_where

I get the error message : too few datasets for (ANOVA/t-test).

I do have data from more subjects to include in the analyses, however I was puzzled as to why the single subject approach failed.

Perhaps I misunderstood the example in the help file. Clarification will be greatly appreciated.

Thanks
Sarah