The regulation of cognitive enhancement devices: refining Maslen et al.'s model