A two-alternative, forced choice method for assessing mouse vision.