Traumatic avoidance learning: the effects of CS-US interval with a trace-conditioning procedure.