 |
 |
Hi In this case run concat ICA with all 75 data sets and create 5 design matrices using the Glm or Feat GUI. Then test the tXX.txt files (in the report sub directory) against each of your 5 designs with fsl_glm. you then have 5 output files for each component. You need to combine the relevant entries (each output files contains only 15 correct values) into a single file and then can average to look at the group. You will need to do all this on the command line, i.e. forget about the subject/session design in the melodic GUI. hth Christian On 7 Jul 2010, at 16:41, Allison Jack wrote: > First, thanks for your prompt response. However, I could use a little clarification on your suggestions for option II. I have 15 subjects and 5 runs. Are you suggesting that I run temporal concatenation ICA with all subjects and all runs (e.g., inputs would be: subj1-run1; subj1-run2...subj1-run5; subj2-run1...etc.), with a subjects/sessions design matrix specifying which inputs belong to which subjects? > > Thanks again, > Allison