To follow-up a previous question regarding use of t-stat results from
Randomise (the t-statistic maps uncorrected for multiple comparisons). Say
in TBSS I hypothesize change in the genu and splenium of the CC? Can I
restrict the randomise output to that region? Is the only direct way to do
this to mask for the region of interest? Or is there an approach that would
let ususe the uncorrected t-stat results, but threshold it at a specific
level to approximate multiple correction testing? Our problem is that we
have seen CC specific changes, but whole brain changes run through randomise
from diffusion and perfusion imaging seem to dilute those results.
Suggestions?
Thanks,
Rob
|