 |
 |
Yes, you need to add an intercept to your model. Aside from the first level analysis (analyzing the BOLD time series data) you must add the intercept (column of 1s) in the higher level analyses. It will be in the model, but you don't include it in the contrast. So, [0 1] and [0 -1] would test the positive and negative trends. It might be a good idea to quickly review linear regression. Jeanette On Thu, Oct 14, 2010 at 8:27 AM, Hui Jing Yu <[log in to unmask]> wrote: > Thanks! > > I just want to make sure my design/contrast file is setup correctly. > > So is it okay in this case that I only have one EV for my clinical measure? > NumWaves 1 > poscorr 1 > negcor -1 > > or do I need to have another EV with all ones for my single group > correlation (column 1 for group, column 2 for clinical measure)? > NumWaves2 > poscorr 1 1 > negcorr 1 -1 >