Search for ways to slow cognitive aging and you land on a pile of devices, each with a confident website. The evidence behind them is wildly uneven. Some of it is a short experiment in a university lab. Some of it is a sham-controlled trial with 670 patients that has not read out. A little of it is a surgical implant whose latest results I cannot find.
Two things to keep straight before we start. First, the medical and consumer lines matter here. A single-channel consumer EEG headset and a physician-prescribed device face very different evidence bars, and none of the six tools below is a defense application. Second, most of the late-stage work targets Alzheimer’s disease, not the ordinary memory slippage of a healthy 70-year-old. Those are different problems, and I think the marketing blurs them more than it should.
Lab results: tACS, tDCS and neurofeedback
1. High-definition tACS. The strongest lab result in older adults comes from Robert Reinhart’s group at Boston University. In a 2022 Nature Neuroscience paper, Shrey Grover and colleagues give healthy adults aged 65 to 88 20 minutes of weak alternating current a day for four days. Low-frequency stimulation over the parietal region improves working memory. High-frequency stimulation over the prefrontal cortex improves long-term memory. Both gains are still visible a month later. 🧠 The catch is size. Oxford’s Masud Husain points out that the gain is roughly three to four extra words from a list of 20, and participants receive stimulation while they do the memory task. Stage: academic randomized study, no commercial product.
2. tDCS. This is the cheapest and most widely copied approach, and the evidence is the least tidy. A 2025 meta-analysis in Alzheimer’s Research & Therapy finds that tDCS raises MMSE scores in older adults with cognitive impairment, but it does not significantly improve memory or executive function. ⚡ In healthy older adults, a separate review pools 13 trials and 647 participants, mostly at 1 to 2 mA over the prefrontal cortex, and calls the results promising. Any consumer headset built on this approach inherits that mixed base.
3. Consumer EEG neurofeedback. A 2021 pilot study in Thailand has older adults play brain games on a single-channel consumer EEG headset for 20 sessions. Resting upper alpha activity improves (P=.04). But the design is pre-post, with no control group, so a practice effect is hard to rule out. 🎮 A systematic review of EEG-based BCI training in older people reaches a similar verdict: promising, thin, and short on well-designed trials.
Where each one stands:
HD-tACS: sham-controlled lab study, effect at one month, no product
tDCS: many small trials, mixed outcomes, no cleared device for this use that I can find
EEG neurofeedback: pilot with no control group, consumer hardware
If a headset maker says its product “builds on” research like this, ask which of those three rows it actually sits in. Which row would you bet on?
Late-stage devices: gamma light and sound, and precuneus TMS
4. Cognito Therapeutics Spectris. The system delivers synchronized 40 Hz light and sound through glasses and headphones, one hour a day at home. The idea traces back to MIT’s Li-Huei Tsai and Ed Boyden. 🚀 In the OVERTURE feasibility study, patients with mild to moderate Alzheimer’s show less white matter loss than sham. A later MRI comparison against matched ADNI controls reports brain volume of +0.58% versus -0.91%. That comparison uses external controls, though, and OVERTURE was not powered to show efficacy. The real test is the HOPE trial: 670 participants, about 70 US sites, sham-controlled, with a primary endpoint combining daily function and cognition. Cognito closed a $105 million Series C in March 2026 and expects topline data before the end of the year. I find no results yet. One more caution: sensory stimulation has not been shown to lower amyloid or tau in people, so the mechanism is still an open question.
5. Sinaptica SinaptiStim. This is personalized rTMS aimed at the precuneus, a hub of the default mode network, calibrated with concurrent TMS and 64-channel EEG plus MRI neuronavigation. Co-founders Emiliano Santarnecchi and Giacomo Koch run the science. 📈 The 52-week Phase 2 results cover 32 patients with mild to moderate Alzheimer’s, who score 1.09 points better on the CDR-SB than sham. The FDA granted Breakthrough Device Designation in 2022, and the system is investigational only. The trial that would support clearance keeps slipping on the calendar, and I find no confirmation that it has started. An earlier rTMS device, NeuroAD, failed its trials and was rejected by the FDA in 2018, which is a useful reminder that small positive trials do not always replicate.
What to watch:
HOPE topline: Cognito says before the end of 2026
HOPE endpoint: daily function plus cognition, a stricter bar than a biomarker
Sinaptica trial start: not confirmed in anything I can find
Amyloid and tau: no human reduction from sensory stimulation so far
Would a clean HOPE result change how you think about non-drug Alzheimer’s care? I think it would, and quickly.
Deep brain stimulation: the invasive bet
6. Fornix DBS. Functional Neuromodulation implants electrodes in the fornix, the fiber bundle that carries output from the hippocampus. 🏥 The Phase 2 ADvance trial enrolls 42 patients with mild Alzheimer’s in an on-versus-off design, and the age analysis finds the most benefit in patients over 65. That result drives the follow-up, ADvance II, a 12-month double-blind study of up to 210 people aged 65 or older. It uses Boston Scientific’s Vercise directional system and carries FDA Breakthrough designation. The trial record lists completion in February 2024. I cannot find published efficacy results, which is odd for a study that finished this long ago. 🔧
The facts that matter:
Population: mild probable Alzheimer’s, age 65 and older
Design: randomized, double-blind, stimulation on versus off
Hardware: Boston Scientific Vercise directional DBS
Status: completed per the trial record, no published efficacy data that I can find
This is a surgical option for diagnosed disease. It is not a wellness product, and nobody should read it as a way to slow ordinary aging. ⚠️
How to read the claims
A 2026 network meta-analysis in Frontiers in Neuroscience finds domain-specific benefits from neuromodulation in Alzheimer’s, with no single technique clearly ahead. That matches what the six tools above suggest. 🔬 Every approach has a signal, and none has a finished case. Every paper, funding round, and trial update we report on ends up summarized and searchable in NeurotechMag Pro, which helps if you want to track HOPE and the others as they move.
If you are evaluating a device or building in this space, run each claim through four questions:
Is the control group sham-stimulated, or just untreated?
Is the population healthy older adults or diagnosed Alzheimer’s?
Is the primary endpoint a clinical scale or a biomarker?
Is there FDA clearance for this specific use? For slowing cognitive aging, I find none among the six.
Most pitches fail at least two of those. 📉 The ones that pass all four will probably be the ones with a pivotal-grade readout behind them, and only Cognito is close to one this year.
Which of these six would you most want to see a clean sham-controlled result from, and what would count as clean enough for you?