Dear SPMers,
I am trying to specify a 2nd level factorial design with 40 subjects using
SPM5. The factors are group (2 levels between) and stimulus type (3 levels
within).
Everything is working fine except when a subject factor is being introduced
(I wanted to make the design look like an SPM2 Repeated Measures Design; the
design matrix has 2 x 3 + 40 = 46 columns). The problem arises (a) when
specifying contrasts using the contrast manager for group differences and
(b) when specifying contrasts for the interaction between the two factors.
In both cases, the error shown is 'Invalid contrast'. Interestingly, the
contrast for the stimulus type factor can be specified without problems. I
tried to specify those contrasts using the spm_FcUtil command which made the
estimation possible, but the F-values seem far too high when comparing the
SPM5 results to the results of a corresponding SPM2 Repeated Measures
Design, but the pattern of effects and the locations of most of the local
maxima are the same.
Could anybody tell me if this design is legal at all and if yes, possibly
give me a hint about what I am doing wrong?
Thanks and Greetings from Salzburg
Johannes Klackl
|