Dear SPM experts,
I came across this unanswered topic https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=spm;98101cc0.1503
which nicely fits to my question:
I want to insert in my design matrix an unmodulated constant-duration condition regressor and an orthogonal duration-modulating (not amplitude-modulating!) variable-duration condition regressor.
If I try to spm_orth the two condition regressors (one with constant duration, another with exactly the same onset vector, but with variable durations according to reaction times/RT), do I need
- to only orthogonalize the variable duractions vector w.r.t. the fixed durations vector? OR
- to also orthogonalize the onset vectors?
I'm a bit confused as I end with 4 columns in a matrix combining the onset and duration vectors of both regressors and spm_orth then orthogonalizes each colum w. r. t. the preceding ones (i.e., sequentially), which is not exactly what I want to end with in my opinon?
Any help would be really appreciated.
Sandra
|