Dear SPM experts,
I’ve checked the spm archives but I couldn’t find an answer to the following
question:
I’m using an event-related design, in which some events depend on the
behavior of the subjects such that there are sometimes only 2 per run.
Since I can not acquire the data in one long run, I would like to
concatenate 5 runs (lets say each contains 200 images) to one long
containing then images 1 to 1000. Right now I’m adding a constant for each
run to correct for session specific effects as well as one extra covariate
modeling the transition from one run to the next (not sure whether this is
really necessary). I would like to highpass-filter the data and use the
default SPM2 value of 128 s. If I apply this filter to the concatenated run,
the “session specific-constants” are filtered too. This is clearly visible
in the design matrix. My first question is: is this a problem??????
In the meanwhile, I’ve changed the spm code a little bit, such that the
filtering is applied for each original run, i.e the filter is applied to
image 1-200 of the concatenated data, then to image 201-400, then to 401-600
and so on. If I check the design matrix the constants look fine. Now
question number 2: is this correct or do I overlook something important
here?
Thanks a lot,
Nici
_________________________________________________________________
Free blogging with MSN Spaces http://spaces.msn.com/?mkt=nl-be
|