On July 20, 2026, a batch of research came out of the Alzheimer’s Association International Conference (AAIC) in London that puzzle enthusiasts and worried adult children have been forwarding to each other ever since. The headline claim: a randomized trial called ENACT found that a specific brain training program moved a blood biomarker tied to Alzheimer’s disease, in the direction you’d want.<\/p>\n\n\n\n
That is a real finding worth taking seriously. It is also, as of this writing, a conference presentation rather than a published, peer-reviewed paper, it comes from a study with 53 participants, and the software involved was made by the same company that funded and ran the trial. None of that means the result is wrong. It means the honest answer to “does brain training prevent Alzheimer’s” is still not proven, and anyone telling you otherwise is skipping steps. Here is what actually happened, and what it does and doesn’t mean if your own brain training consists of a sudoku grid and a cup of coffee.<\/p>\n\n\n\n
ENACT stands for Elucidating the Necessary Active Components of Training. It was presented at AAIC 2026 in London, with earlier detail released via GlobeNewswire on July 13, 2026 and a follow-up release on July 20 covering additional AAIC findings (GlobeNewswire, July 13<\/a>; GlobeNewswire, July 20<\/a>).<\/p>\n\n\n\n The design was a randomized controlled trial, which matters because it means participants were not choosing their own group. Researchers randomly assigned 53 adults to one of two arms: one group did 20 hours of training on BrainHQ, a cognitive training platform made by Posit Science that focuses on speed-of-processing exercises, and the other group played conventional computer games, including word search puzzles, as an active control. Blood was drawn before and after the 20 hours of training to measure a biomarker linked to Alzheimer’s pathology. According to Digital Journal’s coverage, researchers believe this is the first study to directly examine the relationship between brain training and beta-amyloid biomarkers (Digital Journal<\/a>).<\/p>\n\n\n\n The biomarker in question is the Abeta42\/Abeta40 ratio, a blood measure of beta-amyloid that researchers use as an indicator of amyloid buildup in the brain, one of the pathological hallmarks of Alzheimer’s disease. A lower ratio is generally associated with more amyloid accumulation and higher risk. The GlobeNewswire release reports that the group doing BrainHQ speed-of-processing training showed a significant improvement in this ratio compared to the word search control group after the 20-hour training period.<\/p>\n\n\n\n Two things worth flagging on the numbers. First, the coverage does not report a precise effect size in plain percentage terms, just that the change was statistically significant in a group of 53 people, split across two arms. That’s a small sample for a biomarker study, which is exactly why the releases themselves note that larger studies will be needed to confirm the effect. Second, the results reportedly differed by sex, which the release ties to a broader pattern in Alzheimer’s research where men and women respond differently to some interventions. Posit Science CEO Dr. Henry Mahncke was quoted saying the results “will be surprising” to many and that this is the first study to link brain training directly to beta-amyloid levels.<\/p>\n\n\n\n The rule for any study like this: check who ran it, check whether it’s been peer reviewed, and check what the measured outcome actually proves before repeating the headline.<\/p>\n\n\n\n On funding and involvement, the training software is BrainHQ, made by Posit Science, and the ENACT results were announced through press releases distributed by Posit Science, with its CEO quoted directly in the coverage. That doesn’t invalidate the trial. Randomization and blood biomarkers are objective enough that a company can’t simply will a result into existence. But a company with a commercial product being deeply involved in generating and publicizing the evidence for that product is a conflict of interest readers deserve to know about, and it’s a reason to wait for independent replication before treating this as settled.<\/p>\n\n\n\n On peer review, what’s public right now is a conference presentation and press releases, not a published paper in a peer-reviewed journal. The July 13 GlobeNewswire release notes the results are to appear as a conference abstract in Alzheimer’s & Dementia, which is a step toward publication but is not the same as a full peer-reviewed paper with methods, statistics, and independent expert review available for scrutiny. Conference abstracts get revised, sometimes substantially, before anything resembling a final paper appears, and sometimes they don’t turn into a paper at all.<\/p>\n\n\n\n And on the biomarker itself: a shift in a blood ratio linked to amyloid is not the same thing as a lower rate of Alzheimer’s diagnoses. Amyloid biomarkers are a proxy, a plausible one, but a proxy. History in Alzheimer’s research is full of interventions that moved a biomarker in the right direction without translating into fewer people getting the disease. Nobody in this trial was followed for years to see who did or didn’t develop dementia. That’s a different, much longer and more expensive kind of study.<\/p>\n\n\n\n ENACT isn’t happening in a vacuum. It builds on a longer research thread, most notably the ACTIVE study (Advanced Cognitive Training for Independent and Vital Elderly), which enrolled 2,802 older adults back in 1998 and 1999 and has been tracking a large share of them ever since. A February 2026 analysis of that cohort, published in Alzheimer’s & Dementia: Translational Research & Clinical Interventions, found that participants who did five to six weeks of speed-of-processing training, plus a handful of booster sessions at 11 and 35 months, had a 25 percent lower incidence of dementia measured 20 years later, compared to a no-training control group (ScienceDaily<\/a>). Notably, people who did the same training but skipped the booster sessions showed no measurable benefit, which suggests the reinforcement mattered as much as the initial training.<\/p>\n\n\n\n That ACTIVE finding is a genuinely strong piece of evidence: a large sample, two decades of follow-up, an actual dementia diagnosis as the outcome rather than a biomarker proxy, and it’s published, not just presented. ENACT is testing something related, speed-of-processing training on the same general family of software, but with a much smaller group, a much shorter follow-up, and a biological marker rather than a clinical outcome. Taken together they tell a coherent story about the same idea. Taken separately, the smaller study is nowhere near as solid as the older one.<\/p>\n\n\n\n Here’s the honest version, not the inflated one. Sudoku, nonograms, crosswords, and the kind of puzzle you do on this site are not the same intervention as BrainHQ’s speed-of-processing exercises. BrainHQ is a specific, adaptive computer program designed to train visual processing speed under time pressure, with reinforcement scheduling built in. A sudoku puzzle trains logical deduction, pattern recognition, and working memory, at whatever pace you choose. They are cousins, not twins.<\/p>\n\n\n\n So no, doing sudoku every morning does not mean you are replicating the ENACT trial or the ACTIVE study results. Nobody has run a randomized trial testing sudoku against a biomarker or a 20-year dementia outcome the way these studies did for BrainHQ’s speed training.<\/p>\n\n\n\n What sudoku and the tested programs do share is the underlying use-it principle: mentally demanding activity, done consistently, appears to matter more for brain health than passive activity. And puzzles have one advantage no clinical software has: they cost nothing, most people already enjoy them, and there’s no reason to wait for more research before doing something you like that keeps your mind engaged. Just don’t tell yourself it’s proven to prevent Alzheimer’s, because as of today, that claim isn’t supported for sudoku, and it isn’t fully proven yet even for the software these researchers tested. If you want the fuller picture on what’s actually been shown for sudoku specifically, including where the evidence is thin, see our deeper look at sudoku and brain health research<\/a> and our piece on whether sudoku makes you smarter<\/a>. And if you’re playing for an older parent or want a gentler setup, our large-print online sudoku for seniors<\/a> is built for exactly that.<\/p>\n\n\n\n Not proven. What the AAIC 2026 research showed is that one specific brain training program, tested on 53 people over 20 hours, moved a blood biomarker of beta-amyloid in a favorable direction, and that a separate, older, larger study found a 25 percent lower dementia rate two decades after a short course of similar training. Biomarker changes and reduced incidence in a specific trial are meaningful signals, not proof that brain training prevents the disease in general. The ENACT results also haven’t cleared peer review yet, just a conference presentation.<\/p>\n<\/div><\/div>\n\n\n\n In the loose, everyday sense, yes, sudoku is mentally demanding and engages working memory and logical reasoning. In the strict sense used in these trials, no. The tested programs are specific software with adaptive difficulty and set training protocols, and sudoku hasn’t been run through the same kind of trial. Sudoku shares the general principle of active mental engagement, but it isn’t a substitute for the specific interventions being studied.<\/p>\n<\/div><\/div>\n\n\n\n Across the research, the interventions with the most consistent support are aerobic exercise, sufficient and regular sleep, and staying socially connected. Cognitive activity, which includes both structured brain training and everyday mentally engaging habits like puzzles, reading, and learning new skills, functions as one pillar among several rather than a standalone fix. The strongest approach on current evidence combines all of them rather than leaning on any single one.<\/p>\n<\/div><\/div>\n\n\n\n <\/p>\n\n\n\n <\/p>\n\n\n\n Sources<\/strong><\/p>\n\n\n\n A randomized trial called ENACT, presented at AAIC 2026 in London, found that 20 hours of speed-of-processing training shifted a blood biomarker tied to Alzheimer’s disease in 53 participants. Here is what the study actually did, why it isn’t proof of prevention yet, and what it means if your own brain training is a sudoku grid.<\/p>\n","protected":false},"author":1,"featured_media":997,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[178],"tags":[],"class_list":["post-989","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-sudoku-health-wellness"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\tWhat moved, and by how much<\/h2>\n\n\n\n
The caveats<\/h2>\n\n\n\n
How this fits the older evidence<\/h2>\n\n\n\n
What it means for your daily puzzle habit<\/h2>\n\n\n\n
FAQ<\/h2>\n\n\n\n
Does brain training prevent Alzheimer’s?<\/h3>
Do puzzles like sudoku count as brain training?<\/h3>
What actually has the strongest evidence for brain health?<\/h3>
\n