 |
 |
Dear JP, Are you interested in within-subject effects / differences between sessions? 1) If you are not, then it should be sufficient to average the different beta images for each of the conditions and forward these into e.g. a one-sample t-test for group inference = two-stage summary statistics approach of the mixed model (estimate betas on individual level, take them to group level). 2) If you are, then you could set up a One-way ANOVA within subject model or Flexible factorial model, with session as within-subject factor. In case of 1) you should make sure that you average across only those voxels for which beta estimates are available for all the sessions (for each of the subjects combine the different mask images and threshold them at n = no. of sessions). Best Helmut