A quick question as I just start reading about 'randomise.' I just read this in the webpage
http://www.fmrib.ox.ac.uk/fsl/randomise/index.html:
Permutation tests do not easily accommodate correlated
datasets (e.g., temporally smooth timeseries), as null-
hypothesis exchangeability is essential.
Our resting-state datasets have been temporally low-pass filtered with a 0.10 Hz filter. I fear our data will not be 'easily accommodate' in permutation tests.
Am I interpreting this correctly?
Thanks,
- BettyAnn
|