Dear Jenny,
> I have a question about the number of images SPM will use in one trial (a
> sentence) which is presented for 5.5 seconds. The TR used is 2 seconds. This
> means we have 2 or 2.5 functional images per trial.
> The parameter of duration for each condition in the data analysis is 0
> second, because we consider the trials as events.
> Our question is, if the duration is 0, how many images will SPM use? Only
> the first one or all the images made per trial?
> If we want all images per trial to be used, should we change our duration
> time to 5.5 seconds?
If you are using a canonical HRF to model your data, even with an
event time of 0, the estimated response to an event will last a number
of seconds. SPM will use any scans that overlap with this estimated
response to estimate the response to this trial. The default length
of response is 32 seconds, although most of the response happens in
the first 10 seconds or so.
That being said, if your sentences (visual?) are 5.5 seconds long, you
probably don't want to model these with a duration of 0, because the
neural activity likely extends more than a second. You'll have to
decide over what period of the sentence the process you are interested
in is occurring—it may be that entering the whole 5.5 second duration
is best, if that's how long the sentence is presented for. It's not
so much about whether SPM uses the scans or not (yes/no), but trying
to construct a model that most accurately reflects the underlying
neural activity. If you specify a short duration of neural activity
(i.e., 0) when it is actually longer (i.e., 5.5 seconds), you are
probably going to miss some of the response.
Hope this helps!
Best regards,
Jonathan
--
Dr. Jonathan Peelle
Department of Neurology
University of Pennsylvania
3 West Gates
3400 Spruce Street
Philadelphia, PA 19104
USA
http://jonathanpeelle.net/
|