Review: 34 Studies Have Tested Brain Training Apps in Older Adults, What They Measured, What They Missed

What the review covered

Roughly 44 million people worldwide have dementia, and projections cited in the paper indicate that by 2050 about 14,298,671 people in the European Union, and 18,846,286 in the broader European region, will live with the condition.^[1] With no known cure, the authors note that 50% to 80% of adults aged 70 or older who score within the normal range on cognitive tests still report perceived cognitive decline.

A team from the Polytechnic Institute of Viana do Castelo (Portugal), the Wroclaw University of Health and Sport Sciences (Poland), the University of Porto, and the Gdansk University of Physical Education and Sport set out to map that landscape. Their scoping review followed PRISMA 2020 guidelines and the PRISMA-ScR checklist, with the protocol pre-registered on the Open Science Framework.^[1]

The search ran across four databases, Web of Science, PsycInfo, Cochrane, and PubMed, and returned 1,733 records. After removing duplicates (n=18), reviews (n=70), and papers on other diseases (n=381), and screening 1,264 titles and abstracts, 40 studies went to full-text assessment and 34 met the inclusion criteria. Eligible studies enrolled cognitively healthy older adults or older adults with cognitive impairment, including mild cognitive impairment (MCI), Alzheimer's disease, and other dementias, exposed to technology-based cognitive training. Two authors independently assessed risk of bias with the RoBANS tool across six domains.^[1]

What it found: 27 apps, wildly different doses

Across the 34 included studies, 27 different apps were used. Brainer, CogniFit, and Cogmed led with 3 published studies each, followed by Road Tour, the iCST app, COGPACK, and GRADIOR with 2 each. Every other app appeared in just one study.

Training "dose" varied enormously. Weekly practice ranged from 20 to 600 minutes, individual sessions lasted anywhere from 10 to 240 minutes, and frequency ran from 2 sessions per week (the overall average was about two) up to daily. Program length spanned one 10-hour intensive week to 12 months.

Samples were equally uneven, ranging from a single participant up to 2,557 subjects, but only 4 of the 34 studies enrolled 100 or more people; the rest averaged roughly 39 participants. Most studies (22, or 65%) included both sexes, yet 11 studies (32%) did not describe the sex of their samples at all. On clinical status, 28 studies involved participants with mild-to-moderate impairment or a diagnosis of Alzheimer's disease, 2 studied adults at risk of dementia, 3 included no impairment, and 1 gave no information.^[1]

Which cognitive functions the apps target

The review's evidence-gap map shows a clear hierarchy. Across all studies, the cognitive functions covered were memory (79.4%), executive functions (76.5%), attention (73.5%), language (35.3%), praxis abilities (26.5%), visuospatial skills (11.8%), and orientation (5.9%). In absolute numbers, memory, executive functions, and attention stood out with 27, 26, and 25 studies, respectively.^[1]

That concentration is double-edged. Memory, executive function, and attention are the domains where decline typically appears first in MCI, so the focus is clinically sensible, but it leaves visuospatial ability and orientation covered by only a handful of the 34 studies, a gap the authors flag for future research.

Where CogniFit sits in this evidence map

CogniFit was one of the three most-studied platforms in the review, tied with Brainer and Cogmed at 3 published studies each. The CogniFit studies mapped by the authors include a randomized controlled trial of computerized cognitive training in 85 older adults at higher dementia risk due to type 2 diabetes^[3] and an early at-home online training feasibility study in older adults.^[4] Per the review's evidence-gap map, the CogniFit interventions covered memory, attention, executive functions, and praxis abilities.

Still, 3 appearances in a 34-study review makes CogniFit one of the most-researched apps here; the review draws no conclusion about which app works best.

Limitations, what this review does not show

The authors are unusually direct about the field's weaknesses, and about their own:

It does not measure efficacy. A scoping review maps what has been studied; it does not pool effect sizes or test whether app-based training improves cognition. The abstract itself concludes that "great diversity in intervention protocols was found" and that exploring efficacy "requires additional scrutiny."

The underlying studies are heterogeneous and often small. With weekly doses spanning 20–600 minutes, durations from 1 week to 12 months, and only 4 studies above 100 participants, results are hard to compare or generalize. Around a third of studies did not even report the sex of their participants, and most did not categorize samples by the specific cognitive impairment measured at baseline.

The search has bounds. Only English-language, peer-reviewed studies were included, and the authors acknowledge that their search strategy "may not have captured all relevant articles" despite snowballing and expert consultation.

Where this fits in the broader evidence

This review complements, rather than replaces, the efficacy literature. A 2014 meta-analysis of computerized cognitive training in healthy older adults, which the authors cite as a "noteworthy path forward", reported modest but statistically discernible gains in cognitively healthy older adults, while showing that design factors such as dose and setting moderated the size of those gains.^[2] Large trials in the review's evidence map, such as a 2,557-participant online randomized controlled trial in healthy older adults,^[5] show that app-based studies can be run at scale. What Silva et al. add is the field-level picture: which of the 27 apps carry published evidence, in which populations, and where the gaps are. Browse the full list of published CogniFit research on our research studies index, or see condition-specific evidence pages such as cognitive training in multiple sclerosis and cognitive training studied after stroke.