Thanks Stephen, much appreciated. The one good thing is we always sampled
the HR at the same relative point in time, so any biases should not be too
significant. I.e. even though TR differed, the HRF sampling should be
similar:
Conditions a, d: Stimulus 750ms, scan occurred at 375ms + 5-7 seconds
later
Conditions b, c, e, f: Stimulus 2250ms, scan occurred at 1125ms + 5-7
seconds later
The reason I was checking in about this is because I observed activation
in two of the tasks, then, in a third condition combining the two tasks,
deactivation of the same (mesial temporal) structures. FWE significant
results at whole brain level. After this post i contrasted beta images
from the baseline conditions and confirmed they did not differ... so I'm
stuck with exciting results that I'm having trouble explaining!
Best
Christopher
On 4/26/11 10:21 AM, "Stephen J. Fromm" <[log in to unmask]> wrote:
>If you're taking a difference, like "d vs f", and TR is completely
>confounded with d and f, then I think you're in trouble, unless you can
>find a good reason why the confounding wouldn't bias your results.
>
|