None of this has been around long enough to generate good longitudinal data either way, imo. That said, taking a cognitive assessment across a couple domains, playing a series of games based on those very same cognitive tasks, and then retaking the same cognitive assessment seems pretty suspect. I would think results would be biased, similar to retest effects. At the end of the day, improvement is either transferred to a variety of other contexts beyond gaming or not. I would think that the true test. There would be some value in a diagnostic, something that would identify deficits, and I think Lumosity offers such a thing but I don't believe it has been rigorously validated.