What does the random_state parameter do in sklearn's ParameterSampler?

I am trying to implement sklearn's ParameterSampler, but I'm not completely sure what the random_state parameter does.

My guess is that if random_state is set to None, then normal random sampling is used. And if random_state is something other than None, then pseudo random sampling is used?

Also I'm not sure how different int values affect the sampling. For example, is random_state = 1 different form random_state = 2? If yes, how?