Hello,
I'm having a problem running a flexible factorial with a large dataset.
we have 15 subjects, scanned on 3 occasions, with six resting perfusion
images per session (270 images in total).
i'm running a model with 3 factors
1. subjects
2. session
3. scan
the problem is that while i can get the model to run the contrast manager
won't accept any contrasts. I have changed the settings and found that
1. if i remove one subjects scans it runs fine (14subject*18scans)
2. if i remove the subjects regressors (main effect) from the flexible
factorial design matrix it runs fine.
3. if i run a 18 cell full factorial it runs fine
the only problem is when i run 15 subjects with 18 scans in a flexible
factorial.
i thought perhaps that there was a total limit on the number of columns in
the flexible factorial DM, however if i use only 14 subjects but include an
ANCOVA column (global mean) the total number of columns is the same as with
15 subjects but the model, and contrast definition, runs fine.
does anyone know if there are any potential limits either on the number of
columns, total number of images, or anything that could explain this problem?
cheers
Owen
Dr Owen O'Daly
Post-Doctoral Researcher
Centre for Neuroimaging Sciences
Institute of Psychiatry, KCL
De Crespigny Park
Denmark Hill
SE5 8AF
Tel: 020 3228 3047
|