Dear John,
thanks very much for the quick response. However, I would prefer to use extracted values from the ROIs - these individual GM volumetric indices would allow me to compute more complex prediction models, since I'm planning to create predictors that differentiate between patients and controls using the data from the first sample - next, these predictors could be evaluated using the data from the second sample. My initial question might have been to unspecific - sorry.
Best regards,
Ben
You can model scanner effects as a covariate of no interest. In SPM, it just means another column of ones and zeros in the design matrix for the case on 2 scanners. For more than scanners, you'd need additional columns.
Best regards,
-John
On 4 November 2013 08:11, Ben Becker <[log in to unmask]> wrote:
Dear SPMers,
I used VBM8 to compare grey matter volumens in a priori specified ROIs in a group of patients vs controls. Next I analyzed data from an independent sample (again: patients & controls, but from a different scanner) to replicate the initial findings.
For further analyses I extracted ind. grey matter volumes from the ROIs were I find group differences using marsbar. To increase power I would like to pool the extracted data from both samples. However, both samples were acquired on different Scanners (Philips 1.5 vs Siemens 3T Trio) using different sequences. Are the extracted values study- & scanner-specific or do they represent absolute scanner-unspecific indices of grey matter volumes?
Thanks in advance & best regards
Ben
|