Both low and high increases in use were negatively but weakly associated with adolescent cognitive performance two years later.