Dear SPM list,
I am using SPM for analysing an fMRI study, with 8 runs or sessions
per subject. Each run has five conditions (4 higher level conditions
and a fixation condition) experienced once every run.
At the beggining of each task block there is a period of 12 seconds
where visual instructions were given to the subjects. I was planning
to discard these volumes from the statistical analysis.
My concern is that to remove these volumes from the time series will
invalidate the assumptions of SPM statistics. It is to be
expected that there is temporal autocorrelation within the time
series as the TR was 4 seconds.
Will removing these three volumes at the beginnig of each block
(during instructions) contradict the delayed boxcar reponse
function? Would it be more appropriate to leave these volumes in
(i.e., retain the integrity of the time series) and develop a user
defined response function. If so, how would you do this?
Laslty is it necessary to have a complete or uninterrupted
time series for a session in SPM stats?
Any help on these issues would be much appreciated.
Thanks.
Kath.
Kathryn Moores
Cognitive Neuroscience Laboratory,
Flinders University of South Australia,
GPO Box 2100, Adelaide 5001
Phone 08 8201 2420 Fax 08 8201 3877
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|