Gang Chen wrote:
Hi Gang -
I hope this will clear things up. Let me know if you have any other questions.
> Hi Kara,
>
> > I have three runs for each subject:
> > Fixation/Task A
> > Fixation/Task B
> > Task A/Task B
>
> So each run has different set of task assignment?
Yes - and I should have mentioned it is a block design.
> > I have normalized all runs using the average value of the
> > time series as the baseline.
>
> Was the normalization done for each run separately?
Yes - each run was normalized using the average value of the time series as the baseline for that particular run.
> > I would like to concatenate all runs and use
> > 3dDeconvolve, but because I've normalized the data,
> > Task A should be a positive % change value in run 1 and
> > a negative % change value in run 3, which means I would
> > not be able to interpret results from 3dDeconvolve for
> > Task A, right?
>
> I totally get lost here: Why do you have negative regression
> coefficient for task A in run 3?
I'm just talking about the % signal change values that result from normalizing the runs. For voxels specifically associated with Task A - the % signal change would be positive for both run 1 and run 3. But - what about voxels that have increased BOLD signal to Task A, and even larger BOLD signal to task B? For these voxels, the % signal change to Task A would be positive in run 1 (when Task A is compared to baseline), but the % signal change should be negative (though probably small) in run 3 because the baseline for this run was calculated based on the mean of activity during Task A and Task B. The baseline is should be higher for this run, so the same BOLD signal value that was above the baseline in run 1 would fall below the baseline in run 3, and therefore, the % signal change would be negative.
Thanks for your help!
Kara