%   cfg.correcttail = correct p-values or alpha-values when doing a two-sided test, 'alpha','prob' or 'no' (default = 'no')

Let's just talk about the prob alpha no 

in case no 

in fact when doing alpha=0.05 and no correcting,  you are effectivly tested with alpha 0.1

in case alpha 

when doing alpha=0.05 and correcttail='alpha'; you are effectivly tested with alpha=0.025 one side 

incase prob (factor of two)

Effectively, this means multiplying the p-values (in stat.prob, stat.posclusters.prob and stat.negclusters.prob) with a factor of two.