Dear Roy,
Yes, your question is appropriately addressed with a one-sample t-test and a (mean-centered) covariate of memory scores. The t-contrast [0 1] then tests the significance of the regression coefficient corresponding to the slope of the linear fit.
SPM does not give you correlation coefficients, or r-maps. However, the t-test if the slope is equivalent to what some may refer to as 'testing the significance of a correlation'
Best wishes
Martin
On 19 Jul 2012, at 09:51, "Roy Cox" <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Dear spm experts,
I'd like to perform a conceptually very simple analysis in spm, but I'm not sure how to do that.
I have reduced 128 channel EEG data to just one 2D image per subject, where each voxel/electrode value indicates some measure (long story) of activity. I also have a measure of memory performance for each subject, and now I want to calculate the correlation between memory and brain activity at each voxel. So the goal is just a scalp plot of Pearson correlation values (and which ones are significant using random field theory of course).
How can I do this? Intuitively, I'd perform a one-sample t test with memory performance as a covariate, and then just test the relation between the covariate and brain activity. But is this equivalent to the Pearson correlation? Or are there other options?
Best,
Roy
--
Roy Cox, M.Sc. | Brain & Cognition Group | Department of Psychology | University of Amsterdam | Weesperplein 4 | 1018 XA Amsterdam | the Netherlands | room 3.21 | phone: +31 20 525 6847 | email: [log in to unmask]<mailto:[log in to unmask]>
|