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
...
Repeated measures, MNI, clusterm seem to give reasonable results
Contrasts ("stats")
| Number | Name |
|---|---|
| 1 | baseline mean |
| 2 | 6M mean |
| 3 | 12M mean |
| * 4 | 6M > baseline |
| 5 | 12M > baseline |
| 6 | 12M > 6M |
| 7 | 6Mx12M > baseline |
Right
Left
Female OSA, each time-point: ID212
...