Dear all,
Could I ask whether I'm doing the right thing, please. I want to see what areas of the brain are directly correlated with a single-score performance measure (one for each subject). Is it right/valid that entering this score as the co-variate in a 2nd level one-sample t-test, then running the contrast [0 1 ] (where 1 is the covariate column) will show me the regions that directly correlate with that score?
I have then extracted the eigenvariates from this covariate contrast, and indeed they show a correlation with the score (assuming this is valid), however, if I then extract eigenvariates from the same 'sphere' location in the main contrast (active>baseline), the signal change in the same spot does *not* correlate at all with performance score. Hence I'm wondering what the difference is, and which is the method I should use to show neural correlates of the performance measure.
Many thanks for your help,
James Gilleen
Institute of Psychiatry
London, SE5 8AF
tel: 0207 848 0542
|