I'm using SPM99 to do parametric analyses of an event-related fMRI
experiment. I want to look at factors that predict the magnitude of the
BOLD response to individual words e.g. frequency, length.
A previous email to the list suggests that SPM can accept multiple
parametric modulators:
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0102&L=spm&D=0&m=4964&P=25502
However, if I specify more than one parameter of these words and estimate
the model, the parametric columns in the design matrix are greyed-out as
"not estimable". I guessed that this was because some of the parameters
were correlated with each other.
I've tried including only columns that are not significantly correlated (by
a pearson correlation), but still can't estimate these columns. I've also
tried using the matlab "orth" function to orthogonalize the matrix of
parameters to no avail.
Any suggestions?
Matt
****************************************************
Matt Davis
MRC Cognition and Brain Sciences Unit
15 Chaucer Road, Cambridge, CB2 2EF
email: [log in to unmask]
tel: 01223 355 294 (#266) NOTE New telephone
Fax: 01223 359 062 extension number
****************************************************
|