Do Brain Games Improve Memory? What Research Shows

Mostly, brain games make you better at the games themselves. A 2016 review found extensive evidence that brain-training programs improve performance on the tasks people practice. It found less evidence for closely related tasks, and little for unrelated tasks or everyday thinking (Simons et al., 2016). The National Institute on Aging (NIA) says there isn’t enough evidence that commercial brain-training apps have the same effect as the training tested in the large ACTIVE trial (NIA). In ACTIVE itself, the memory training’s benefit had faded by the 10-year mark, while gains from reasoning and speed training held up (Rebok et al., 2014).

Near transfer vs. far transfer

Researchers split the question in two. Near transfer means getting better at the trained task or ones very much like it. Far transfer means improving at different things, such as verbal or nonverbal ability, reading or arithmetic (Melby-Lervåg et al., 2016; Gobet & Sala, 2023). Better memory in daily life would count as far transfer, and everyday performance is where the evidence is thinnest (Simons et al., 2016).

• In a six-week online study of 11,430 people, participants trained on reasoning, memory, planning, visuospatial and attention tasks. They improved on every trained task but showed no transfer to untrained tasks, even closely related ones (Owen et al., 2010).
• A meta-analysis of 87 publications on working memory training found short-term, specific effects that didn’t carry over to measures of real-world skills (Melby-Lervåg et al., 2016).
• A 2023 review of the field concluded that the overall far-transfer effect of cognitive training is null (Gobet & Sala, 2023).

The n-back task is a good example of this pattern, and our guide to the n-back test walks through its training studies.

The ACTIVE trial

ACTIVE (Advanced Cognitive Training for Independent and Vital Elderly) is a large randomized controlled trial (NIA). It enrolled 2,832 adults aged 65 to 94 in six US metropolitan areas in 1998 and 1999. They were randomly assigned to memory, reasoning or speed-of-processing training, or to a control group with no contact (Ball et al., 2002). Sessions ran 60 to 75 minutes, twice a week for five to six weeks. Some participants were then randomly picked for booster sessions 11 and 35 months later. The trial was funded by the NIH through the NIA and the National Institute of Nursing Research (NIH).

Speed training targeted visual processing speed, not memory. Participants picked out an object in the center of a screen while also spotting targets around the edges, and the display got shorter and the task harder as they improved (same source).

Follow-up	What researchers found
2 years	Each type of training improved the ability it targeted, but no effect on everyday functioning was detected (Ball et al., 2002)
5 years	Gains in all three abilities were still measurable, and reasoning training was linked to less self-reported difficulty with daily tasks (Willis et al., 2006)
10 years	Reasoning and speed gains held, memory gains did not. All three trained groups reported less difficulty with daily tasks than controls (Rebok et al., 2014)
10 years	Speed training was linked to a 29% lower risk of dementia, a result that only just reached statistical significance. Memory and reasoning training were not (Edwards et al., 2017)
20 years	In Medicare claims, speed training plus boosters was linked to a 25% lower risk of diagnosed dementia. Speed training without boosters, memory training and reasoning training were not (Coe et al., 2026)

The NIA’s director said the 20-year findings suggest moderate cognitive training could delay dementia, while noting that more research is needed (NIH). The result comes with caveats:

• The 20-year analysis covered the 2,021 participants in traditional Medicare and relied on diagnoses recorded in claims. Its authors note that people with better access to care, family members who notice decline, or more education are more likely to be diagnosed, and that selection bias after randomization is possible (Coe et al., 2026).
• One author disclosed consulting for, and owning stock in, the company that markets the speed-training software (same source).
• A letter published in June 2026 reanalyzed the paper’s published numbers. After a correction for multiple comparisons, no single training arm stood apart from the control group. The overall pattern looked more like a shared protective effect across all three types of training, concentrated in participants aged 65 to 69 (Dickens, 2026). The letter is newer but works only from summary statistics. The original paper, which used participant-level data, is the stronger evidence, though the debate shows the result isn’t settled.

For memory specifically, ACTIVE’s memory training lost its effect by 10 years and wasn’t linked to fewer dementia diagnoses at 20 (Rebok et al., 2014; Coe et al., 2026).

What scientists’ consensus statements say

In October 2014, the Stanford Center on Longevity and the Max Planck Institute for Human Development published a statement from cognitive psychologists and neuroscientists. It said brain-game claims are often exaggerated and sometimes misleading, and that there is little evidence that playing brain games improves broad underlying cognitive abilities (Stanford Center on Longevity). More than 70 scientists signed it. Several months later, 133 scientists and practitioners replied with a letter arguing that research shows brain training has broad benefits (Simons et al., 2016).

To settle the dispute, Simons and colleagues reviewed every peer-reviewed intervention study cited by leading brain-training companies. Many had major flaws in design or analysis, and none followed all the best practices the authors set out (same source).

A 2017 review by the National Academies of Sciences, Engineering, and Medicine found “encouraging but inconclusive” evidence that cognitive training could help prevent or delay Alzheimer’s, mild cognitive impairment or age-related cognitive decline. It reached the same verdict on blood pressure management and physical activity (NIA).

The FTC’s 2016 case against Lumosity

In January 2016, Lumos Labs, the company behind Lumosity, agreed to pay $2 million to settle Federal Trade Commission charges of deceptive advertising. The FTC said the company claimed its games would improve performance at work, at school and in sports, and protect against age-related decline, mild cognitive impairment, dementia and Alzheimer’s disease, all without scientific evidence to back those claims. A $50 million judgment was suspended because of the company’s financial condition. The order requires competent and reliable scientific evidence before any similar claims in the future (FTC). The FTC’s consumer protection director summed it up: the company “simply did not have the science to back up its ads” (same source).

What the National Institute on Aging says

The NIA says cognitive training “appears to have benefits for maintaining cognitive health in older adults,” and it is funding a large trial of whether speed training can reduce cognitive impairment and dementia. It also tells readers to beware of claims that computer and online games can improve memory and thinking (NIA). NIA’s consumer pages on this topic were last reviewed in 2023 and 2024, before the 20-year ACTIVE results came out in February 2026 (NIA; NIH).

What the evidence does support

The NIA says research suggests these steps may lower the risk of cognitive decline, though for several of them it says more research is needed (NIA):

• Manage blood pressure. In the SPRINT MIND study, people 50 and older who lowered their systolic blood pressure below 120 mmHg had a lower risk of mild cognitive impairment over five years of treatment.
• Move more. Federal guidelines recommend at least 150 minutes of physical activity a week. Studies linking exercise and brain health are encouraging, but more research is needed.
• Learn something new and demanding. In one study, older adults who learned quilting or digital photography improved their memory more than those who only socialized or did less demanding activities.
• Stay connected. In a trial of almost 200 adults aged 75 and older, early results showed that regular internet calls could help lower the risk of cognitive decline.
• Look after the basics: sleep seven to nine hours, treat hearing or vision loss, manage conditions like diabetes and depression, and ask your health care provider whether any of your medicines affect memory.

The NIA adds that researchers can’t say for certain that healthy lifestyle changes will protect against dementia, but that taking these steps may help lower your risk (NIA).