I posted a question a couple of weeks ago re the use of a fourier set in a
sparse design er-fMRI experiment, but didn't receive a reply.
I have two conditions/event types (control and task) delivered in randomised
order that I have modelled with non-windowed fourier sets without HRFs, and
have also entered the realignment parameters for each subject as covariates
in the design matrix. Although I have the resulting F-contrasts for each
condition, I would now like to compare them in a manner analogous to a
subtraction design. I understand that this is not straightforward given that
I have used fourier sets.
I can think of possibly two ways of doing this in SPM99 involving exclusive
masking or orthogonalising the contrasts to see if the task condition
accounts for more variability than the control condition, but am not
entirely sure how to proceed or how best to interpret the results of these
approaches. Could anyone provide me with some information as to the relative
merits of each approach and how best to interpret the results?
thanks,
Greig
--
Centre for Magnetic Resonance
The University of Queensland
Brisbane, QLD 4072
AUSTRALIA
Tel: +61 (0) 7 3365 4250
Fax: +61 (0) 7 3365 3833
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|