Does brain training work?

The reviews distinguish learning a game from wider improvement

A rising score tells you that something has been learnt, although it leaves open what that learning consists of. You might come to recognise a puzzle's recurring structure, remember a useful shortcut or simply become quicker at operating its controls, and each of those changes can help with that particular task. The further claim that they improve an unrelated skill needs separate evidence from a test of that skill.

Daniel Simons and colleagues made this distinction in their 2016 review, Do “Brain-Training” Programs Work?, published in Psychological Science in the Public Interest. They found evidence of improvement on trained tasks, less evidence for closely related tasks and little evidence for distantly related tasks or everyday cognitive performance. They also identified weaknesses in study methods. That pattern makes a programme's own score an inadequate test of its wider promises. The Association for Psychological Science's page on the review summarises these conclusions.

A more focused 2016 review by Monica Melby-Lervåg, Thomas Redick and Charles Hulme examined working-memory training, which practises holding and manipulating information. Their meta-analysis of 87 studies found reliable short-term gains on closely related memory measures, without convincing evidence of improvement in reading, arithmetic or reasoning when trained groups were compared with control groups. The paper's abstract states these conclusions. Both reviews concern interventions that were tested, so they cannot establish that every possible future training design will fail.

The distinction matters most when your time is limited. If a game teaches you to solve that game, the learning may still be enjoyable and worthwhile, but an exam candidate needs a further connection between the practice and the questions that will appear on the paper, and someone seeking help with reading needs evidence about reading.

Match your practice to the skill you need

Start by describing a task you would like to perform more reliably in a recognisable setting. “Improve my thinking” is too broad to guide a choice of exercise. “Identify the unsupported claim in a short report” gives you something to practise and a way to notice progress: you can compare your explanation with the text and see whether you separated its evidence from an assumption that the author left unstated. Specific knowledge also affects what you can judge. A clear target reveals the knowledge you need, since judging a report about a subject you understand differs from judging unfamiliar material.

For reading, take a passage similar to the material that causes difficulty, then write its main claim and the evidence supporting it. Compare your account with the passage and mark where you added an assumption. For an aptitude test, practise the question formats it uses and examine your reasoning after each answer. Fresh questions matter because repeating a remembered solution can conceal a weakness in the method that produced it.

A practical extension is to change examples while keeping the intended skill visible. Someone learning to compare proportions could work with a recipe, a discount and a population table, then explain the calculation in each case. This is a proposal for practising the target across contexts, rather than a measured guarantee of transfer. It creates an opportunity to discover whether the rule survives changes in wording and presentation.

If your aim is remembering subject knowledge, select the facts and explanations you will need to retrieve. Close the material, attempt an answer and check what was missing before returning to it later. The choice of content matters as much as the exercise format: remembering a sequence of coloured tiles does not by itself demonstrate that you can explain a course concept.

Check that a product's evidence supports its promise

Read a training claim as a statement about an outcome, a group of people and a comparison. Evidence from children learning a particular exercise cannot automatically answer a claim about adults handling workplace decisions, just as a study showing improvement during training leaves unanswered whether that improvement lasts after practice stops. Those are separate gaps in the evidence. A product should make these boundaries easy to find beside its claim.

The comparison group deserves attention because people can improve simply through becoming familiar with a test or expecting practice to help. Walter Boot and colleagues explain this difficulty in their 2013 paper on placebo effects in psychological research. Even a group doing another activity may have different expectations about improvement, so an active comparison alone does not remove that uncertainty. Their discussion of matching expectations explains why a before-and-after success story cannot establish which ingredient caused the change.

The outcome measured should differ from the training exercise, and the comparison group should receive a credible alternative activity. Read whether participants were assigned to groups randomly, whether the planned outcomes were specified beforehand and whether later testing checked persistence, because a finding becomes more useful when you can tell what was planned, who was included and how long any difference lasted. These checks help you judge how well a claim is supported, although none of them guarantees a useful effect, and when a reported benefit is small its uncertainty and its relevance to your goal deserve particular care.

A citation is useful when you can follow it to the work being described. Scan the abstract for the measured outcome before looking at the promotional summary, then check whether the summary widens that outcome into a promise about daily life. Where the two differ, base your decision on the narrower finding and leave the broader promise unproven.

Use your own time as a practical comparison

You can examine whether an exercise deserves a place in your routine without trying to run a clinical trial. Choose a target task and save some unfamiliar examples for checking it later. During practice, note recurring errors and the approach you tried, since this gives you something more informative than a record of completed sessions.

After a period of practice, return to the saved examples under similar conditions and inspect the same kinds of error. A change in your own results is a useful observation, although it cannot isolate the training from other learning, expectations or ordinary variation. If only the game score improves, consider assigning more of your available practice to the target task and checking that task again.

Enjoyment can justify keeping a puzzle in your day even when wider benefits remain uncertain. The decision becomes clearer when you identify its purpose explicitly, whether that is recreation or preparation for a particular kind of question. For your next session, choose an example of the skill you need and record the reasoning error you want to examine.