In favor of behavior: on the importance of experimental paradigms in testing predictions from Gray's revised reinforcement sensitivity theory