Spectris AD fails to slow decline in a large Alzheimer’s trial

A brain-stimulating headset built around flashing light and rhythmic sound has fallen short in a closely watched Alzheimer’s disease trial. After 12 months, people assigned to the active device did not do better cognitively than those who used a sham setup, despite the idea that driving gamma-band activity might support memory networks and preserve function.

The device, Spectris AD, pairs goggles with headphones and is worn at home for an hour each day. Its design rests on a plausible biological premise: flickering visual input and synchronized auditory cues may encourage gamma-wave activity, a pattern linked in earlier research to attention, memory processing and network coordination. In neurodegeneration, that concept has attracted interest because circuit dysfunction appears early, before widespread neuronal loss leaves too little salvageable tissue to influence.

The new trial delivered a blunt answer. Investigators randomised 673 people with Alzheimer’s disease to active or sham treatment, then followed them for a year. On a 156-point scale measuring cognition and ability to perform daily activities, the active-device group declined by 8.97 points on average, while the sham group declined by 8.82 points. Those nearly overlapping results leave no persuasive signal of overall efficacy.

A full-year exposure did not separate active treatment from sham

The absence of benefit matters because this was not a brief proof-of-concept experiment. Participants used the headset for 60 minutes a day across 12 months, a long enough interval to detect at least a durable symptomatic effect if one existed. Instead, the decline ran almost in parallel in both arms, suggesting that the headset did not alter the underlying trajectory of established disease in the enrolled population.

One sentence captures the central result: the active treatment arm declined just as much as sham.

Cognito Therapeutics said the mild cognitive impairment subgroup showed a stronger signal, with active-device users experiencing 41.6% slower decline than sham-device users, a difference of about 3 points on average. That finding may be biologically interesting, but it sits in tension with the broader null outcome in the overall cohort. In Alzheimer’s research, subgroup effects can reflect true treatment sensitivity, yet they can also emerge from chance, baseline imbalance or post hoc slicing of a negative study.

Eric McDade, a neurologist at Washington University School of Medicine in St. Louis, was blunt in his assessment of the overall dataset. He called the trial negative and said any next step should be a follow-up study focused on the mild population. His view reflects a familiar problem in device development: when the primary population fails, any claim of benefit in a narrower subgroup has to earn its way through a prospectively designed confirmatory study rather than optimism alone.

Why gamma-wave modulation still attracts attention

The scientific appeal of gamma-wave modulation has never rested only on symptom scores. In preclinical work, rhythmic sensory stimulation has been linked to changes in neuronal firing patterns and brain activity states associated with cognition. The theory is attractive because it tries to influence network function without invasive procedures or systemic drug exposure, which makes it appealing for people who may already be taking multiple medications or are intolerant of additional pharmacologic burden.

Yet Alzheimer’s disease is not a single-circuit disorder. By the time a patient has clear cognitive and functional impairment, synaptic loss, tau pathology, amyloid burden and neuroinflammation may all be interacting across multiple brain regions. A headset that nudges oscillatory activity may be too modest a tool for a disease process that has already become structurally entrenched. The lack of separation between active and sham treatment in this trial is consistent with that possibility.

Rob Howard, a psychiatrist at University College London who specializes in dementia, said the data were exploratory rather than persuasive proof of efficacy. He added that if this had been a drug, such a clear failure would likely stop a company from pursuing regulatory approval. His criticism goes to the heart of the matter: a safe intervention is not automatically a useful one, and low toxicity cannot compensate for absent clinical benefit.

Earlier disappointment weakens the case for approval

This is not the first time Spectris AD has disappointed researchers. A smaller trial in 2024 also failed to slow cognitive and functional decline in Alzheimer’s disease. Two negative studies in succession make the central question less about whether the device is safe and more about whether the underlying mechanism can produce a meaningful clinical effect in humans with established neurodegeneration.

Christian Howell, Cognito’s chief executive, said the company still sees beneficial signs in patients with mild cognitive impairment and plans to file for US Food and Drug Administration approval next year. That stance suggests the company is betting that the mild subgroup contains the clearest responders and that regulatory reviewers will look favorably on a more selective claim. The data presented so far do not settle that argument; they sharpen it.

For clinicians, the practical consequence is straightforward. A headset that promises to slow Alzheimer’s disease has not done so in the broad population tested, and the strongest signal now sits in a narrower subgroup that remains exploratory. Patients with dementia and their families are often drawn to non-drug technologies that appear gentle and mechanistically elegant, but the evidence still has to show that those features translate into real-world cognitive preservation.

What the trial leaves unresolved

The present trial leaves several questions open, but not in the abstract sense of “more research is needed.” The more immediate issue is whether any future study can identify a mild-stage population, a dosing strategy or a biomarker-defined responder group in which gamma stimulation produces an effect large enough to matter at the bedside. Without that, the headset remains a biologically interesting device with no demonstrated clinical advantage over sham treatment.

The study also reinforces a broader truth in dementia therapeutics: plausibility cannot substitute for outcomes. Neural entrainment may change brain rhythms, but rhythm change alone does not guarantee better memory, better day-to-day function or slower decline. For now, the headline finding is simple and unforgiving — the active headset did not beat sham treatment in the overall Alzheimer’s cohort, and the field is left waiting for a better-designed test of whether any subgroup can truly benefit.


Reporting by Dr. A. Manasa, MBBS for Medicine World Foundation, with additional data provided by British Journal of Cancer.
PRIMARY CLINICAL EVIDENCE & REGISTRY DATA