Dorian -
I would argue that contrasting trial-related responses with different
trial durations is valid, provided that one is clear about the
interpretation: that any resulting differences in the parameter
estimates reflect differences in activity per unit time (not per trial).
Here is a snippet from
http://imaging.mrc-cbu.cam.ac.uk/imaging/DesignEfficiency that might help:
"Whether these are better modelled as events, or as epochs of different
durations (e.g, with duration equal to the RT for each trial), is
debatable. For example, if the stimulus duration were constant and only
RTs varied, then the activity in V1 may not be expected to vary with RT,
so an "event model" might fit better (and in this case, the parameter
estimate can be interpreted as the BOLD response /per trial/). For
activity in premotor cortex on the other hand, greater activity might be
expected for trials with longer RTs, so a "varying-epoch model" might
fit better (and in this case, the parameter estimate can be interpreted
as the BOLD response /per unit time/). So the answer depends on your
assumptions about the duration of neural activity in the particular
brain regions of interest."
Rik
Dorian P. wrote:
> Hi SPM list,
>
> Does it make sense to contrast conditions with different trial
> duration? I.e., one with 2s presentation of the stimulus and the other
> of 5s (convolved with 2s and 5s HRF, respectively).
>
> I imagine not, because the beta values will only tell if the HRF fits
> more or less well the signal and cannot tell if condition A produces
> more activity then condition B. But what is the right way to think
> about it?
>
> Dorian
>
--
-------------------------------------------------------
Dr Richard Henson
MRC Cognition & Brain Sciences Unit
15 Chaucer Road
Cambridge
CB2 7EF, UK
Office: +44 (0)1223 355 294 x522
Mob: +44 (0)794 1377 345
Fax: +44 (0)1223 359 062
http://www.mrc-cbu.cam.ac.uk/people/rik.henson/personal
-------------------------------------------------------
|