Hi,
You should not model the baseline condition in FSL.
Hence do not model the fixation condition.
You should model each of the other conditions with
a separate EV, to allow for arbitrary relationships between
their amplitudes (which you do not get with the single
EV approach). It is easiest to set up the two individual
EVs using the 3 column format.
I hope this helps.
All the best,
Mark
On 4 May 2010, at 21:18, Kathryn Cullen wrote:
> Hello, I am struggling to set up my analysis for a face viewing fmri
> study. In this block-design experiment, we had 13 blocks lasting 24
> seconds as follows:
>
> fixation 24
> fear 24
> neutral 24
> fixation 24
> fear 24
> neutral 24
> fixation 24
> fear 24
> neutral 24
> fixation 24
> fear 24
> neutral 24
> fixation 24
>
> I am trying to figure out how to best set up the first level
> analysis. I initially tried to enter each condition as a separate
> EV, assigning 1 to the blocks for that EV and 0 to all others. This
> led to an error, I bleieve because I have 3 items instead of 2. (The
> error goes away when I only include 2 EVs).
>
> I then tried using the 3 column format to incorporate all conditions
> into one EV, in which I assigned fear=1, neutral=-1, and fixation=0,
> similar to that described in the frequently asked questions (http://www.fmrib.ox.ac.uk/Members/joreilly/frequently-asked-questions-ppi
> ). Here is how my 3 column ev looked:
>
> 0 24 0
> 24 24 1
> 48 24 -1
> 72 24 0
> 96 24 1
> 120 24 -1
> 144 24 0
> 168 24 1
> 192 24 -1
> 216 24 0
> 240 24 1
> 264 24 -1
> 288 24 0
>
> The problem is that this makes a wave form where BOLD is highest for
> fear and lowest for neutral, and fixation is in the middle. So I
> haven't taken fixation out of the model, it's being considered a
> middle-of the road activator which I don't want. Could you advise me
> on how to set up my analysis?
>
> Thanks, Katie
>
|