Dear Experts,
In TBSS analysis and after running the scripts and making the design matrix, I wanted to improve the estimation of the variance that feeds into the final "t" statistic image by using the variance smoothing option -v <std> in Randomise command line.
Actually in randomise_manual (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Randomise/UserGuide) they mentioned that (if you have fewer than 20 subjects (approx. 20 DF), then you will usually see an increase in power by using variance smoothing)
What is the recommended value for smoothing if I have more that 20 subjects. In this case is it advisable to do smoothing?
Thanks
Antoine
|