Determination of antidromic excitation by the collision test: Problems of interpretation