- Hi there, I find it tedious and error-prone to manually input all the parameters using GUI form of GLM program if you have many subjects, say over 200 in our case. Is there an easier way? It is possible I think, to construct a *.mat and a *.con file and they are all we need to do a "randomise" analysis. Is it possible to construct a *.fsf file from these *.mat and *.con files? I think that using glm (GUI) does have many advantages. Wei Wen