Using fitcdiscr without feature normality

Hi!

I was making a classification task (training a classifier) using fitcdiscr. Is it still possible to use it if my features (the input X) have not a gaussian distribution?

Thanks in advance