For the OSA longitudinal, the MNI with repeated measures was the "best" analysis. This is also the most logical one - scales for head size. The corrections in order were
clusterm
clustere
tfce
The simple 2-group did nothing, it needed the repeated measures analysis (i.e., including a subject variable). And only the right side showed effects.
Sydney
ID389, Z:\Analysis\Sets\Luke\OSA\
UCLA
baseline, 6 months, 12 months
...