Does brain training work? What the evidence actually says

Brain training has been sold on one claim — do these puzzles, become generally smarter — and audited by a large amount of research that mostly does not support it. It is worth knowing both halves before spending a month on any of it.

Near transfer: real, boring, useful

Practise a task and you improve at that task and at tasks that closely resemble it. This part is not controversial. Span tasks improve span; visual search drills improve visual search; a rotation puzzle makes rotation puzzles easier.

It is unglamorous, and it is still worth something. If your work involves holding several things in mind at once, getting reliably better at exactly that is not a trivial outcome.

Far transfer: contested, and mostly small

The claim that trained tasks lift general reasoning, school performance or everyday competence is where the evidence thins out. Meta-analyses repeatedly find small effects that shrink further in studies with active control groups — that is, when the comparison group also does something engaging rather than nothing at all.

There have also been regulatory actions against brain-training marketing in several countries for claims that outran the evidence. That history is a good reason to distrust any product whose main pitch is a bigger number.

What has better support than puzzles

• Sleep. The single largest short-term influence on attention and working memory in any healthy adult.
• Aerobic exercise, with more consistent cognitive findings than most training software.
• Learning something structured and effortful — an instrument, a language, a craft — which keeps demanding new skills instead of the same one.
• Removing interruption from work that needs continuity, which costs nothing and is measurable in a week.

So what is a training exercise good for?

Three honest things. It gives you a repeatable measurement under fixed conditions, which almost nothing else in daily life does. It builds the habit of deliberate, difficult practice. And it makes progress visible, which is what keeps people doing anything at all for more than a week.

What it is not is a shortcut, and an exercise designed to feel good every session is training your patience with a slot machine rather than any ability.

That is why our exercises are finite, scored and comparable only to your own record — and why the week ends in a retest that is allowed to come out lower. See how the training works

How to judge any brain-training product in a minute

• Does it say what it measures, in named abilities, or only give you one number?
• Does it get harder when you succeed, and stay hard?
• Does it compare you to yourself, or to strangers on a leaderboard?
• Does it ever show you a result that is worse than last time? A product that cannot go down is not measuring anything.