Dear Amy,
Almost all of these questions are answered in the FSL course practical on FEAT:
http://www.fmrib.ox.ac.uk/fslcourse/
I advise you to do the first practical exercise there (with the AV data) as this will
teach you how to set these parameters up (when coupled with reading the FEAT
webpage documentation).
All the best,
Mark
On 21 Dec 2010, at 00:56, Amy Zheng wrote:
> Hi all,
>
> I have a block design with 4 different conditions, and I'm trying to figure out how to properly set up the model. It starts with an ON period of 25s then alternates with OFF 25s, etc; so, Condition A, Rest, Condition B, Rest, Condition C, Rest, Condition D, Rest, total lasting about 200s.
>
> I'm running the first-level analysis and in the full model setup I put 4 EVs, with basic Square shape and Gamma Convolution.
> For the 1st EV:
> Skip (s)=0
> Off (s)=25
> On (s)=25
> Phase (s)=25
> Stop after (s)=50? (I'm tempted to put 50s, as in 25s On+25s Off for each condition, or should I leave it as -1?)
>
> For the 2nd EV:
> Skip (s)=50?
> Off (s)=25
> On (s)=25
> Phase (s)=25?
> Stop after (s)=50?
>
> For the 3rd EV:
> Skip (s)=100?
> Off (s)=25
> On (s)=25
> Phase (s)=25?
> Stop after (s)=50?
>
> For the 4th EV:
> Skip (s)=150?
> Off (s)=25
> On (s)=25
> Phase (s)=25?
> Stop after (s)=50?
>
> ALSO, should I change the default "Stddev and Mean lag" values under the Gamma Convolution option, which are 3, and 6 s, respectively? Another question is: how does one determine what high pass filter value to use?
>
> Thanks,
> Amy
>
|