 |
 |
Dear SPM experts, I apologize if I missed this on previous posts to the listserv, but I have a quick question regarding testing the conjunction of the parameter estimates (i.e., condition effects implicitly compared to baseline) at the second level. Assume I've estimated the effect of two conditions, A and B, for every subject at level. Is the following procedure valid? 1. Run one-sample t-tests on the parameter estimates for A and B for the group and compute the spmT images for the positive effect. 2. Compute a minimum statistic image from the spmT images for the one-sample t-tests for A and B (which includes the minimum statistic value at each pixel). 3. Draw inferences from the thresholded minT image. On a related point, I have used Donald McLaren's approach to create binary images for each subject separately, then sum these images to get an image representing the proportion of the group showing the conjunction effect. However, this method does not permit an inferential test. Would an alternative method involve computing minimum contrast images for each subject and then taking those images up to the second level? Thanks for any tips! Bob ------------------------------------------------------------------------------- Bob Spunt Postdoctoral Fellow Social Cognitive and Affective Neuroscience Labs Department of Psychology University of California, Los Angeles