Exploring limits of neuropsychological screening in ALS: The FAB problem