I extracted the beta weights for this voxel and it is negative for all three tasks, but when I average them in the same way that the -acontra command does, I get the same result as the AFNI result suggests. So it is a negative beta issue it appears. I'm just stumped as to which 'effect' is more 'true'. The raw activity plot or the beta plot? The Raw activity plot shows this region is more active in my memory conditions. The Beta effect shows that this region has a steeper negative slope for memory. I guess they aren't totally inconsistent..in this case a negative slope = more activity is all. Which effect would you report though...both? and then explain why the two appear to be contradictory bur in essence are not?

N

P.S. to answer the other post. My baseline is not a 'rest/fixation' baseline. It is a non-memory verbal control task..so subjects are doing something during it.