Filled versus empty intervals in prospective hypnotic time estimation with a real-simulator design.