N*N coherence computation and test

Many thanks Francois, I tried with the real data I want to test, no error occurred this time. Sorry for another question: When I computed the N*N coherence at the source level, should the source be rectified before the coherence computation?