Do we really need the salience map to explain oculomotor selection performance? (Yes, we do!)