Hi Vadim
What you enter for your duration depends on how long the "event" is
you are interested in. If you are interested in activity that you
believe will be ongoing throughout your entire 3 sec trial, you are
probably better off entering a duration of 3 sec.
Regarding the rest of your question, SPM uses the event onsets and
durations, along with your basis function(s) (i.e. the canonical hrf),
to build a model of your experiment. This is independent of your TR.
With a 3 second trial length, the peak response to this trial is
likely to occur during a different trial, anyway. This delay is
incorporated in the model if you use the canonical hrf basis set. So,
you don't have to worry, all of the relevant scans will be used to
estimate your model, regardless of what you enter as your event
duration.
Hope this helps,
Jonathan
On Fri, Oct 3, 2008 at 3:58 PM, Vadim Axel <[log in to unmask]> wrote:
> Hi,
>
> My trial length is 3 sec, while I am using short TR of 1.5 sec (partial
> brain scan). In "Specify 1-st level" template I defined Interscan Interval
> as 1.5. Now, if I define durations as "0" and the onsets are the onsets of
> my trials (not TRs), what is going to be my model? Will it take the first of
> two TRs only? Then what is going to occur with the second TR? As far as I
> understand in order to enhance my precision it is better to use both TRs
> for estimation of HDR (since they belong to the same trial). What is the
> correct way to define what I need?
>
> Thanks in advance,
> Vadim
>
|