Every few months a new gadget promises to keep your brain young. Caps, headbands, light helmets, apps. š§ Each one comes with a study attached, and that is where the trouble starts, because āa studyā can mean a 5-person case series, a company press release, or a randomized trial that ran for 20 years. Those are not the same thing, and they should not get the same headline.
I sort each tool below by how far it has actually traveled: lab demo, early clinical trial, late-stage trial, or cleared for use. One caveat first. Nearly all of the clinical data here comes from people with mild cognitive impairment or Alzheimerās disease, which is a medical use. As far as I can find, none of it is cleared to slow normal aging in healthy people. ā ļø That gap between treating a disease and keeping a healthy brain sharp is exactly the one marketing likes to blur.
1. Speed training: the only tool with 20-year data
The strongest result in this whole space is not a stimulator. It is the ACTIVE trial, a randomized study of more than 2,800 adults aged 65 and older, enrolled in 1998 and 1999. Participants got memory training, reasoning training, speed-of-processing training, or nothing. The training was about ten sessions over roughly six weeks. š
The 20-year follow-up reported by UAB found that people who did speed training and came back for booster sessions had about a 25% lower risk of a dementia diagnosis. I think that is a remarkable result. I also think it needs a careful reading. š
Design: randomized, with a no-contact control group, and dementia diagnoses pulled from Medicare claims covering 1999 to 2019.
The numbers: 105 of 264 people (40%) in the speed-plus-booster group were diagnosed, versus 239 of 491 (49%) in the control arm.
What did not work: memory and reasoning training showed no statistically significant drop in risk, with or without boosters.
The catch: the booster comparison is a subgroup of people who showed up for the extra sessions, and claims-based diagnoses are blunter than clinical exams.
Stage: an updated version of the same drill is sold as the Double Decision exercise in the BrainHQ app, so this one is consumer software today, not a lab demo.
The funny part is that the brain games most people actually buy look like the memory drills that did nothing. Fast, adaptive visual attention is the thing that moved the needle.
2. tDCS and tACS: real signals, small trials
Electrical stimulation through the scalp is the most hyped category here, and the evidence is the most uneven. š The best-known work comes from Robert Reinhartās lab at Boston University. A 2022 Nature Neuroscience paper gave adults aged 65 to 88 20 minutes of stimulation on four consecutive days, and the memory gains lasted about a month. Low-frequency stimulation over the parietal region helped working memory, while high-frequency stimulation over the prefrontal cortex helped long-term memory. A 2019 study from the same field reported that older adults (60 to 76) matched younger adults on a working-memory task while the stimulation was on.
Plain tDCS looks less convincing. A 2024 meta-analysis in Frontiers in Aging Neuroscience pools tDCS paired with cognitive training and finds a standardized mean difference of 0.35 for working memory, rising to 0.72 with ten or more sessions. An earlier analysis of 266 participants found tDCS barely beat sham when added to working-memory training (g = 0.07). Both can be true, which tells you the protocol matters more than the brand name. š§Ŗ
Stage: lab demos and small early trials, mostly in healthy older adults.
Outcome measured: performance on memory tasks, not dementia onset.
Open question: how long the effect lasts past a month, and whether it transfers to daily life.
If youāre building here: dose and personalization are the whole game. A single fixed montage for every older brain is probably the wrong product.
3. rTMS: the best Alzheimerās data so far, from one site
Transcranial magnetic stimulation has been FDA-cleared for depression since 2008, so the hardware is mature. The Alzheimerās version is not. The candidate to watch is SinaptiStim from Sinaptica Therapeutics, which uses EEG-personalized, neuronavigated rTMS aimed at the precuneus, part of the default mode network. š§²
The companyās 52-week Phase 2 trial enrolled 48 patients with mild-to-moderate Alzheimerās at a single center, randomized and sham-controlled. NeurologyLive reports a 44% slowing on the CDR-SB scale (a 1.36-point decline versus 2.45 for sham). Co-founder Giacomo Koch of the University of Ferrara leads the science. The device has FDA Breakthrough Device Designation (2022), and a larger trial is the next step. David Knopman of the Mayo Clinic called the results strong enough to justify a bigger study, which is a fair summary and a lot more measured than the word ābreakthroughā on a website. š
Stage: early clinical trial, with a Phase 3 planned. I find no readout from it yet.
Sample size: 48 people, one site. Treat 44% as a hypothesis.
History: an earlier rTMS device for Alzheimerās, NeuroAD, failed its trials and was rejected by the FDA in 2018.
Setting: clinic-based sessions. This is not something you strap on at home.
Do you trust a 48-person result enough to follow the company, or do you wait for the replication? I lean toward waiting, but I keep checking.
4. 40 Hz light and sound, and near-infrared light
This is where the Alzheimerās money is going. Cognito Therapeutics builds Spectris AD, a physician-prescribed, at-home system that delivers synchronized light and sound at 40 Hz through glasses and headphones. The idea comes from animal work suggesting gamma-frequency stimulation can reduce amyloid plaques, and the human question is whether it slows decline. š”
The early human data is the small OVERTURE feasibility study, which was not powered to show efficacy. A post hoc analysis, reported by the company, claims a 76% reduction in cognitive decline over six months. I would not lean on a post hoc slice of a small study. The real test is HOPE (NCT05637801): a sham-controlled, late-stage trial with 670 patients at more than 60 U.S. sites, MMSE scores of 15 to 28, 12 months of daily home use, and a composite primary endpoint of ADCS-ADL and MMSE. Cognito closed a $105 million Series C in March 2026 to support the readout and regulatory filing, and topline data was promised before the end of 2026. I find no published topline as of today. For the full research, funding, and trial picture behind this story, NeurotechMag Pro has you covered. š°
Near-infrared photobiomodulation sits a step behind. A 2017 case series used 810 nm light pulsed at 10 Hz in five patients with dementia and reported gains on MMSE and ADAS-cog, but the authors themselves call for larger controlled trials. A pilot sham-controlled trial in mild cognitive impairment (NCT05563298) is posted as a preprint, so it is not peer reviewed yet. š
Spectris AD: late-stage trial, investigational, with Breakthrough Device Designation and no reported ARIA events.
Photobiomodulation: early clinical trials in dementia and MCI, with a positive pooled signal across 13 mostly small studies.
Consumer note: near-infrared helmets are sold as wellness gear. That is a retail fact, not a clinical one.
5. Neurofeedback, and what would change my mind
The sixth tool is EEG neurofeedback, where you watch your own brain activity and learn to shift it. The results are the most mixed of the six. šÆ A small placebo-controlled study of 17 older adults found that sensorimotor rhythm training improved visual working memory. But a double-blind study of 20 older adults found that gamma and beta training raised brain activity within sessions with no effect on cognition, and a recent randomized trial of motor-imagery neurofeedback saw no cognitive gains beyond classical stimulation. Older brains produce noisier signals, and protocols built on young adults do not always transfer. š
Stage: early trials and pilots, with consumer EEG headbands already on the market.
Pattern: people learn to move the signal, but the cognitive payoff is inconsistent.
Bottom line: promising enough to study, not strong enough to buy on faith.
Here is what would change my mind. A clean HOPE readout would make 40 Hz stimulation the first at-home neurotech with late-stage evidence in Alzheimerās. A multi-site rTMS replication would move magnetic stimulation from interesting to credible. A properly blinded photobiomodulation trial with hundreds of people would settle a decade of case reports. Until then, the only thing with two decades of data is a speed drill you can run on a laptop. Which of these would you bet on if the next readout lands in your favor?