Does brain training work?
Brain training reliably improves the task you practise. It does not reliably make you smarter, and no strong study has shown it prevents memory decline or raises general intelligence.
What the evidence shows
The best-known test of this question ran in 2010, when researchers put more than eleven thousand people through six weeks of online cognitive training and compared them with a control group. Everyone who trained improved, but only at the specific tasks they practised. None of the gains carried over to other mental tasks, even ones that used very similar skills.
A 2016 review by a panel of cognitive scientists reached the same conclusion from the wider literature. People get better at the trained task. Evidence that this turns into better memory, attention or reasoning in daily life is thin, and evidence that it raises general intelligence is thinner still.
Practice does not transfer much
Psychologists call this the transfer problem. That repeating a task makes you faster at that task, known as a practice effect, is not in question. What is in question is “far transfer”: whether practice on one task changes anything outside it. Most of the available evidence says it mostly does not, or does so only a little.
The same pattern shows up in working memory training specifically. People improve on the trained working-memory task, but a meta-analysis of the training studies found little evidence that the improvement spreads to unrelated measures, such as reasoning tests.
What a short daily task is still good for
None of this makes a short daily task pointless. It is a low-stakes way to practise attention, inhibition or working memory under a little time pressure, and tracking your own scores over weeks is a reasonable way to see whether you are getting better at that specific task. What it will not do is diagnose a problem, measure your intelligence, or replace sleep, exercise or medical advice.