One-trial object recognition in rats and mice: Methodological and theoretical issues