40 Hz Ultrasound for Health & Longevity

Evidence Review created on 08/10/2026 using AI4L / Opus 5

Also known as: 40 Hz-Pulsed Transcranial Ultrasound Stimulation, Gamma-Frequency Transcranial Ultrasound Stimulation, 40 Hz TUS, 40 Hz tFUS, Gamma-Pulsed Focused Ultrasound, Ultrasonic Gamma Entrainment

Motivation

40 Hz ultrasound is a form of non-invasive brain stimulation. A small device held against the scalp sends sound waves far above the range of human hearing through the skull, switching them on and off forty times a second. The interest lies not in the sound itself but in that rhythm. Brain cells appear to fall into step with it, producing a fast electrical rhythm that is weakened in ageing and in memory disease.

The approach grew out of a decade of work showing that flickering light and pulsing tones delivered at the same forty-beat rhythm can clear sticky protein deposits from the brains of mice. Ultrasound was proposed as a way to reach deeper structures, such as the memory-forming region, which light and sound cannot easily touch. Laboratory work since 2020 has reported both the rhythm change and the protein clearance, while human work with brain ultrasound in general is still at an early stage.

This review examines what is known about 40 Hz ultrasound: how it is thought to work, what has and has not been measured, the reported harms, the protocols used in laboratories and clinics, and how far the evidence extends beyond animals.

Benefits - Risks - Protocol - Conclusion

Grokipedia

Transcranial focused ultrasound

Grokipedia has no article devoted specifically to 40 Hz ultrasound; this is the site’s primary dedicated page for the underlying technique, covering acoustic fundamentals, transducer implementation, therapeutic applications, clinical development and safety. It is useful for understanding the parameter space (carrier frequency, pulse repetition frequency, intensity) within which the 40 Hz variant sits.

Examine

No Examine article exists on 40 Hz ultrasound or on transcranial ultrasound stimulation. Examine covers supplements, nutrition and dietary interventions, and does not cover neuromodulation devices or brain-stimulation procedures.

ConsumerLab

No ConsumerLab article exists on 40 Hz ultrasound. ConsumerLab independently tests dietary supplements and consumer health products for identity, purity and label accuracy, and does not evaluate brain-stimulation devices or clinical procedures.

Systematic Reviews

A real-time PubMed search for systematic reviews and meta-analyses of 40 Hz ultrasound, transcranial ultrasound stimulation and gamma-rhythm stimulation returned the following, selected for relevance, size and recency; none is specific to the 40 Hz pulse rate delivered by ultrasound, because no such body of human trials yet exists.

The Effects and Safety of Gamma Rhythm Stimulation on Cognitive Function in Alzheimer’s Disease: A Systematic Review and Meta-Analysis - Wu et al., 2025

Eight randomised controlled trials (RCTs, studies in which participants are assigned to real or dummy treatment by chance) and 291 participants, pooling brain and sensory routes to 40 Hz stimulation. It reports large cognitive benefits and no excess adverse events versus sham, but the effect sizes are implausibly large for the sample sizes involved and the included trials are small and heterogeneous, so it is best read as a signal rather than an estimate.
Human Studies of Transcranial Ultrasound neuromodulation: A systematic review of effectiveness and safety - Sarica et al., 2022

The reference safety dataset for brain ultrasound in people: 35 studies, 677 participants, with no severe adverse effects and mild symptoms in 3.4% of subjects. It is also the clearest demonstration that stimulation effects vary with sonication parameters rather than being a single fixed phenomenon.
The effectiveness and safety of low-intensity transcranial ultrasound stimulation: A systematic review of human and animal studies - Qin et al., 2024

Eleven randomised human trials and 44 controlled animal studies, covering neurological, psychiatric, pain, sleep and blood-pressure outcomes, with suppression of inflammatory response and facilitation of neurogenesis among the recurring biological findings. Adverse events were confined to reversible headache, nausea and vomiting in a few human subjects.
A systematic review of preclinical and clinical transcranial ultrasound neuromodulation and opportunities for functional connectomics - Pellow et al., 2024

The broadest map of the field to date, 187 studies across animals and humans, with an emphasis on what brain imaging readouts can and cannot tell us about ultrasound’s effects. It is the most useful source for understanding why identical acoustic settings produce excitation in one study and inhibition in another.
Effect of Low Intensity Transcranial Ultrasound Stimulation on Neuromodulation in Animals and Humans: An Updated Systematic Review - Kim et al., 2021

A methodologically explicit review that scored risk of bias in the animal literature and found it low-quality (25–43% on a standard scale), largely because of performance and detection bias. It documents that excitatory and inhibitory effects both occur in animals while only inhibitory effects had then been shown in primates and humans.

Mechanism of Action

40 Hz ultrasound combines two independent ideas: ultrasonic neuromodulation, which is the delivery mechanism, and gamma entrainment, which is the intended biological effect.

The acoustic carrier. A piezoelectric transducer emits a continuous wave at a carrier frequency of roughly 0.25–2 MHz (megahertz, millions of cycles per second). This frequency is chosen because it passes the skull with tolerable attenuation and can be focused to a millimetre-to-centimetre spot at depth. The carrier itself is not the therapeutic signal; it is the vehicle.
The 40 Hz envelope. The carrier is switched on and off, or amplitude-modulated, 40 times per second. This on/off rate is the pulse repetition frequency (PRF, how often the bursts of sound repeat). It is the PRF, not the carrier, that neurons are thought to follow. In the founding protocol the carrier was 2.0 MHz, the PRF 40 Hz, and each pulse lasted 400 microseconds.
How sound moves a neuron. Ultrasound has no direct electrical action. The leading candidate mechanisms are mechanical: acoustic radiation force and cyclical pressure deform the neuronal membrane and open mechanosensitive ion channels (proteins that let charged particles cross the cell membrane when the membrane is stretched), including the TRP (transient receptor potential, a large family of sensor channels) and Piezo1 families and two-pore potassium channels. A competing account, intramembrane cavitation, proposes that the pressure wave causes microscopic expansion and contraction of the fatty bilayer itself, changing its electrical capacitance and thereby firing the cell.
Gamma entrainment. Gamma-band oscillations (fast brain rhythms above roughly 30 cycles per second, generated largely by fast-spiking inhibitory interneurons) are reduced in Alzheimer’s disease. Driving neurons at 40 Hz is intended to restore this rhythm. Rodent recordings show increased spontaneous gamma power, restored coupling between slow theta rhythms and fast gamma rhythms, and increased sharp-wave ripples, the brief high-frequency bursts associated with memory consolidation.
Downstream amyloid handling. Gamma entrainment by any route is reported to recruit microglia, the brain’s resident immune cells, toward amyloid-β (Aβ), and to increase bulk fluid flow through the brain’s glymphatic waste-clearance pathway. Whether ultrasound engages the glymphatic arm as light and sound do has not been shown directly.
Inflammatory signalling. The most recent mechanistic work reports down-regulation of RIPK1 (a protein that switches cells between survival and inflammatory death) and phosphorylated NF-κB (a master switch controlling inflammatory gene expression), with reductions in the inflammatory messengers IL-6, IL-1β and TNF-α. This is a single-laboratory finding and should be treated as provisional.
Competing explanation: the audible click. A 40 Hz pulse train falls inside the range of human hearing and is transmitted through skull bone. Kop et al., 2024 showed that a well-replicated ultrasound effect on the human motor system disappeared once the auditory cue was properly masked and controlled, meaning that the brain was responding to the sound of the stimulation, not the stimulation. Against this, a re-analysis of the largest sham-controlled ultrasound trial in Alzheimer’s disease found no lasting activation of auditory networks and no relationship between auditory measures and cognitive scores (Mitterwallner et al., 2026); this work comes from the group that ran the original trial, which has a direct stake in the result. For 40 Hz specifically, the auditory route is not merely a nuisance: because 40 Hz auditory stimulation is itself an established gamma-entrainment method, an “auditory confound” would still be expected to entrain gamma, making the two explanations unusually hard to separate.

Historical Context & Evolution

Original use of the technology. Ultrasound entered medicine as an imaging tool in the 1940s and 1950s and, in high-intensity focused form, as a way to destroy tissue thermally. Its use to modulate rather than destroy neural tissue dates to work by Fry and colleagues in the 1950s, was largely dormant for decades, and was revived from about 2008 onward as transducers, skull-aberration correction and neuronavigation matured.
Origin of the 40 Hz idea. The gamma-entrainment line began separately, with optogenetic (light-triggered firing of genetically modified neurons) and flickering-light experiments in Alzheimer’s mouse models reported by Li-Huei Tsai’s group at MIT in 2016, which found roughly halved amyloid plaque load in the visual cortex after 40 Hz light exposure, and frequency specificity: 20 Hz and 80 Hz did not reproduce it. That group’s members founded Cognito Therapeutics to commercialise light-and-sound gamma stimulation, a financial interest that runs through much of the subsequent gamma literature and is disclosed in those papers.
Why ultrasound was proposed. Flickering light reaches the visual cortex; pulsing tones reach the auditory cortex. Neither reaches the hippocampus or deep cortical structures where Alzheimer’s pathology is most consequential. Ultrasound can be focused at depth through the intact skull, so pulsing it at 40 Hz was an attempt to obtain the gamma effect anywhere in the brain. Bobola et al., 2020, from the University of Washington, stated this hypothesis explicitly and tested it.
What the early findings actually showed. The 2020 report found microglia colocalised with plaque in 36.0% of cases after ultrasound versus 14.2% under sham, and a plaque burden reduction of about 47% after five daily one-hour sessions. The 2021 Korean replication (Park et al., 2021) found reduced insoluble Aβ42 in prefrontal regions, fewer hippocampal plaques, increased spontaneous gamma power and no increase in microbleeding, but also found microglial process length and number decreased rather than increased. These are opposite-direction microglial findings from two competent laboratories, and the discrepancy has not been resolved; it plausibly reflects the two-week versus five-day exposure, the anaesthetised versus sedated preparation, or different microglial states at different timepoints.
The critique, and its own evidentiary status. The claim that ultrasound neuromodulation is “just the sound of the machine” is frequently made and is now supported by direct experimental evidence in the motor system. It has not, however, been tested at 40 Hz, nor in any amyloid or rodent-pathology paradigm, where the outcome is a tissue measurement rather than a subjective or motor readout that could plausibly be produced by hearing a click. The critique is strong where it was tested and unproven where it was not.
Where opinion stands and what changed it. Enthusiasm for gamma stimulation in general cooled when human trials produced weaker and less consistent results than the mouse work, including a light-therapy trial that failed to reduce brain amyloid in patients. It has partially recovered with reports of preserved function over six months of daily audiovisual stimulation and, more recently, biomarker changes over two years in a five-patient open-label extension (Chan et al., 2025). The ultrasound branch has not yet been tested in humans at all, so no consensus about it exists to be revised.

Expected Benefits

Every benefit below rests on animal work or on inference from other 40 Hz stimulation routes. No human study has yet delivered ultrasound pulsed at 40 Hz and measured a clinical outcome. For a reader accustomed to weighing interventions on human trial data, that absence is the single most important fact in this review, and it is the reason no benefit is graded High or Medium.

Low 🟩

Reduction of Amyloid-β Plaque Burden

Repeated daily sessions of 40 Hz-pulsed ultrasound reduce amyloid plaque load in transgenic mouse models of Alzheimer’s disease, chiefly 5×FAD animals (a strain engineered to carry five Alzheimer’s-linked mutations and develop plaques early). The proposed mechanism is gamma-driven recruitment of microglia to plaques together with enhanced bulk clearance. The finding has been reproduced by at least four independent groups in the United States, South Korea and China using different transducers, exposure durations and mouse lines, which is unusually consistent for a preclinical effect. It has never been measured in a human brain, and amyloid reduction has repeatedly failed to translate into proportionate cognitive benefit for other interventions, so plaque change should be read as a mechanistic marker rather than as a health outcome.

Magnitude: Approximately 47% reduction in plaque burden after five daily one-hour sessions in 5×FAD mice; reduced insoluble Aβ42 in prefrontal cortex and fewer hippocampal plaques after two weeks of two-hour daily sessions.

Microglial Engagement with Amyloid Plaques ⚠️ Conflicted

Acute exposure changes the behaviour of microglia, the brain’s resident immune cells, around amyloid deposits. The two best-characterised studies disagree on the direction: the founding work reported a large increase in microglia physically colocalised with plaque one hour after sonication, whereas the independent replication reported decreased microglial process length and number after two weeks. Both are consistent with microglial state change but not with a single simple story, and the discrepancy may reflect acute activation followed by resolution, or different anaesthetic and exposure conditions. Because microglial activation is beneficial when transient and harmful when sustained, this ambiguity matters more than it might appear.

Magnitude: 36.0 ± 4.6% of microglia colocalised with plaque after ultrasound versus 14.2 ± 2.6% under sham in the acute experiment; reduced microglial process length and number in the two-week chronic experiment.

Entrainment of Gamma-Band Brain Rhythms

Ultrasound pulsed at 40 Hz increases power in the gamma band and restores the coupling between slow and fast rhythms that is disrupted in Alzheimer’s models. This is the intended proximal effect and the one most directly tied to the mechanism. Rodent hippocampal recordings show the enhancement builds during stimulation and persists for around five days afterwards, along with an increase in sharp-wave ripples, the brief high-frequency bursts tied to memory consolidation. A recent parameter comparison found that 40 Hz amplitude-modulated ultrasound was outperformed by 5 Hz modulation for driving hippocampal network coupling, which suggests 40 Hz is not automatically the optimum for ultrasound as it appears to be for light.

Magnitude: Increased spontaneous gamma power with normalised cross-frequency coupling in treated animals; enhancement persisting approximately 5 days after the final session.

Improved Memory Performance in Disease Models

Treated animals perform better on standard rodent memory tests than sham-treated littermates. The proposed mechanism is restoration of hippocampal network timing plus reduced amyloid and inflammatory load. Effects have been shown in Y-maze and Morris water maze paradigms and, in the most recent work, using a wearable transducer on freely moving animals, which removes anaesthesia as an explanation. Rodent maze performance is a weak proxy for human cognition, and the studies are small, unblinded to a variable degree, and conducted by groups with an interest in the technique’s success.

Magnitude: Optimal cognitive outcome at an acoustic intensity of 2.14 W/cm² over a 14-day regimen; memory improvement observed at both 40 Hz and 200 Hz pulse rates in a separate hippocampal study.

Speculative 🟨

Reduction of Neuroinflammatory Signalling

The most recent mechanistic work reports that 40 Hz ultrasound down-regulates RIPK1, phosphorylated NF-κB and markers of necroptosis (a form of inflammatory cell death), with reductions in the inflammatory messengers IL-6, IL-1β and TNF-α, alongside changes in innate-immune gene expression. If real, this would be a more plausible route to clinical benefit than plaque clearance, because chronic neuroinflammation tracks cognitive decline more tightly than plaque count does. The basis is a single laboratory’s RNA-sequencing and protein work in female mice only, with no replication, so it remains mechanistic evidence rather than a demonstrated benefit.

Enhanced Glymphatic Waste Clearance

Gamma stimulation delivered as light and sound has been shown to increase cerebrospinal fluid influx and interstitial fluid efflux, with amyloid removal abolished when glymphatic clearance is blocked. Whether ultrasound engages the same pathway is untested; the mechanical action of ultrasound on perivascular spaces could plausibly add to or interfere with it. The basis here is entirely inference from a different stimulation route.

Preservation of Cognitive Function in Healthy Ageing

The longevity-relevant question is not whether the technique treats dementia but whether earlier use preserves function in people who are still cognitively intact. Gamma power declines with normal ageing, and the deep-targeting advantage of ultrasound would in principle allow the hippocampus to be reached before atrophy sets in. No study of any kind has examined 40 Hz ultrasound in cognitively healthy people; a trial in preclinical Alzheimer’s disease is only now enrolling. The basis is mechanistic reasoning alone.

Sleep Quality and Circadian Consolidation

Gamma-frequency stimulation delivered as light and sound has been associated with reduced night-time activity and improved daily functioning in dementia patients, and disrupted sleep is both a risk factor for and a consequence of neurodegeneration. A registered trial is now testing hippocampal ultrasound explicitly against sleep endpoints. No sleep data from 40 Hz ultrasound exist yet, so this rests on the shared mechanism and on a single ongoing study.

Benefit-Modifying Factors

APOE4 carrier status: APOE4 is a variant of the apolipoprotein E gene that raises Alzheimer’s risk and is associated with earlier and denser amyloid deposition and greater vascular fragility. Carriers plausibly have more substrate for a plaque-directed intervention to act on, but also more amyloid-laden vessels, and by analogy with anti-amyloid drug therapy they are the group in whom clearance-promoting interventions carry the highest vascular risk. No 40 Hz ultrasound study has stratified by genotype.
Skull thickness and skull density ratio: The single largest determinant of how much acoustic energy reaches the target. Thick or dense temporal bone attenuates and defocuses the beam, and inter-individual variation in transmitted intensity across a population can exceed the difference between an active and a sham exposure. This is a physical, individually measurable modifier that has no analogue in drug therapy and is why serious protocols use imaging-based acoustic simulation.
Baseline gamma power and entrainment capacity: The two-year audiovisual gamma extension study found that the patients who benefited were those who retained strong measurable entrainment on electroencephalography (EEG, a recording of the brain’s electrical rhythms from scalp electrodes), while those who lost it did not benefit. Baseline and on-treatment EEG gamma response is therefore the most direct candidate predictor of response.
Baseline amyloid and tau burden: Plasma p-tau217 (a phosphorylated form of the tau protein measurable in blood that tracks Alzheimer’s pathology) and amyloid imaging define whether there is pathology to modify at all. An intervention whose primary preclinical effect is plaque clearance has, by construction, nothing to clear in a person without amyloid, whatever else it may do to brain rhythms.
Sex: The most detailed mechanistic study to date was conducted exclusively in female mice, chosen because female 5×FAD animals show more aggressive pathology. Human Alzheimer’s disease is more prevalent in women, and sex differences in microglial phenotype and inflammatory signalling are well documented, so the direction and size of any effect may differ by sex. There are no human data and no male-versus-female comparison within the ultrasound literature.
Pre-existing health conditions: Cerebral small-vessel disease (damage to the brain’s smallest blood vessels, seen as white-matter changes on a scan), prior stroke and untreated sleep apnoea all independently degrade gamma-band activity and cognition, and would be expected to blunt any rhythm-based intervention. Conversely, conditions that thin or breach the skull (prior craniotomy, burr holes, cranial implants) alter acoustic transmission unpredictably and can locally concentrate energy.
Age: The largest sham-controlled trial of any ultrasound neuromodulation in Alzheimer’s disease found the cognitive benefit concentrated in patients aged 70 or younger, with the interaction between treatment and age significant while the main effect was not (Matt et al., 2025). That trial used single ultrashort pulses at 5 Hz rather than 40 Hz gamma pulsing, but it is the only human age-stratification available in this technology class, and it points toward earlier rather than later application. Skull density also rises with age, reducing transmitted energy in exactly the population most likely to seek the intervention.

Potential Risks & Side Effects

The safety record summarised here comes from low-intensity transcranial ultrasound in general, since no adverse-event data exist for the 40 Hz variant specifically. Most published human exposures are single sessions of minutes; the 40 Hz protocols proposed on the basis of animal work involve one to two hours daily for weeks, an exposure regime for which no human safety data exist at all.

Medium 🟥 🟥

Transient Headache, Scalp Discomfort and Neck Pain

The most consistently reported adverse effects of transcranial ultrasound are mild and self-limiting: headache, scalp heating or tingling, neck pain and muscle twitching. The proposed mechanisms are direct thermal deposition at the skin–skull interface, pressure from the transducer and coupling apparatus, and incidental stimulation of scalp nerves. The evidence base is the pooled human literature, in which no severe adverse effect has been reported across hundreds of participants. Symptoms resolved without intervention in every reported case, but the exposures involved were far shorter than the daily hour-long regimens implied by the animal protocols.

Magnitude: Mild symptoms in 3.4% of participants (14 of 425) across 35 human transcranial ultrasound studies, with no severe adverse events.

Audible 40 Hz Tone and Off-Target Auditory Co-Stimulation ⚠️ Conflicted

A pulse train at 40 Hz produces an audible buzz conducted through skull bone, which is both an unwanted stimulus in its own right and a potential explanation for any observed effect. The mechanism is bone- and air-conducted sound reaching the cochlea and auditory cortex. Four controlled experiments across three institutions showed that a well-established ultrasound effect on the motor system was entirely attributable to this auditory route rather than to direct neuromodulation. A re-analysis of the largest sham-controlled ultrasound trial in Alzheimer’s disease reached the opposite conclusion for long-term effects, finding no lasting auditory-network activation, though that analysis was performed by the group that ran the original trial. Practically, the risk is a combination of tinnitus-like discomfort and a substantial chance that the intervention’s apparent benefits are attributable to something other than the mechanism claimed.

Magnitude: In four controlled human experiments the entire motor-inhibition effect of pulsed ultrasound was accounted for by the auditory confound; no lasting auditory-network change was detected in a 60-patient crossover trial.

Low 🟥

Nausea, Vomiting and Dizziness

Reversible gastrointestinal and vestibular symptoms have been reported in a small number of human participants. The likely mechanism is incidental stimulation of vestibular structures or brainstem regions when the acoustic beam is imprecisely targeted, aggravated by skull-induced beam distortion. The evidence base is the pooled randomised human literature, where these were the only adverse events recorded beyond headache. All reported cases were transient and required no treatment.

Magnitude: Reversible headache, nausea and vomiting reported in a small number of subjects across 11 randomised human trials; no adverse effects reported in 44 controlled animal studies.

Mood Deterioration and Transient Anxiety

Mood worsening, anxiety and sleepiness appear in the adverse-event lists of human ultrasound neuromodulation studies. Mechanistically, ultrasound aimed at frontal or deep limbic targets is explicitly being developed to change mood, so mood change in the unwanted direction is a plausible on-target effect rather than an incidental one. In the largest sham-controlled trial in Alzheimer’s disease, depression was the most commonly reported symptom, but it occurred at similar rates during real and dummy stimulation, which argues that much of it reflects the underlying condition and study burden rather than the ultrasound.

Magnitude: Depression was the most frequent reported symptom in a 60-patient crossover trial, occurring at comparable rates in the active and sham conditions; mood deterioration and anxiety appear among the 3.4% mild-symptom total in the pooled human dataset.

Tissue Heating and Mechanical Bioeffects

Ultrasound deposits heat, preferentially at the skull where absorption is highest, and exerts mechanical stress that at sufficient pressure can nucleate gas bubbles. The mechanism is straightforward physics: acoustic absorption converted to heat, plus negative pressure peaks driving cavitation. The international reporting and safety consensus documents — issued by a consortium that counts ultrasound device manufacturers among its members, and so not a disinterested author of the ceilings it sets — exist precisely because published protocols vary by orders of magnitude in intensity and in duty cycle (the fraction of each second during which sound is actually being emitted), and because the 40 Hz protocols used in rodents employed intensities well above diagnostic imaging norms. No thermal or cavitation injury has been documented in a human neuromodulation study, and one 40 Hz rodent study specifically checked for and found no increase in microbleeding.

Magnitude: International consensus treats a temperature rise of up to 2 °C, or an absolute tissue temperature of 39 °C, as non-significant risk, with the maximum permitted exposure time falling from 80 minutes at that band of the thermal index (a number printed by the machine that estimates how much a given setting will heat tissue) to 10 seconds at the highest; the founding 40 Hz rodent protocol used a spatial-peak pulse-average intensity of 190 W/cm².

Speculative 🟨

Seizure Provocation

Rhythmic 40 Hz stimulation is, by design, an attempt to impose synchrony on cortical networks, and imposed synchrony is the defining feature of a seizure. Photic stimulation in the gamma range is a recognised trigger in photosensitive epilepsy, and ultrasound has induced afterdischarges (bursts of abnormal electrical activity that outlast the stimulus that provoked them) in animal preparations at high exposure. Against this, the pooled human ultrasound literature reports no seizures, and the gamma-stimulation meta-analysis found sensory 40 Hz stimulation safe even in participants with epilepsy. The basis for concern is mechanistic and analogical rather than observed.

Sustained Microglial Activation

The intended mechanism is microglial recruitment, but chronically activated microglia drive the neuroinflammation that contributes to neurodegeneration. Whether daily stimulation over months produces a self-limiting activation-and-resolution cycle or a persistently primed inflammatory state is unknown, and the two principal animal studies point in opposite directions on microglial morphology. No study has followed microglial phenotype beyond a few weeks, and none has done so in humans.

Unintended Blood–Brain Barrier Permeability

Focused ultrasound is used deliberately, with injected microbubble contrast agents, to open the blood–brain barrier for drug delivery. Without microbubbles the pressures used for neuromodulation are far below the threshold for barrier opening, but the margin narrows with high duty cycles, skull-induced standing waves and any circulating contrast agent. The concern is theoretical for the parameters described in the literature and would become concrete only with equipment operating outside characterised limits.

Cumulative Effects of Multi-Year Daily Exposure

The protocols implied by animal work involve one to two hours of daily sonication, potentially for years if used as a preventive measure. Nothing is known about cumulative acoustic exposure of the human brain on that timescale; the longest human safety data in this technology class cover a two-week course. The basis is the complete absence of long-term data rather than any observed signal.

Risk-Modifying Factors

APOE4 genotype: Carriers of this apolipoprotein E variant have more amyloid deposited in vessel walls, and are the group in whom amyloid-clearing therapies most often produce brain swelling and microbleeds. Any intervention that mobilises vascular amyloid should be assumed to carry more risk in carriers until shown otherwise.
Anticoagulant and antiplatelet use with baseline microbleed count: Anticoagulants and antiplatelet agents (blood thinners, which slow clot formation) raise bleeding risk everywhere, including inside the skull. Gradient-echo or susceptibility-weighted magnetic resonance imaging (MRI, a scan that shows brain structure without radiation) reveals pre-existing microbleeds. A high microbleed count, particularly combined with anticoagulation, is the standard exclusion criterion in amyloid-directed and in ultrasound blood-brain-barrier trials and is the most relevant baseline biomarker for physical risk here.
Skull integrity and implanted hardware: Craniotomy flaps, burr holes, cranial plates, shunt hardware, cochlear implants and deep-brain-stimulation leads all distort the acoustic field. Metal and air interfaces can reflect and concentrate energy, converting a nominally safe exposure into a locally intense one. This is a hard physical contraindication rather than a graded risk.
Seizure threshold and epilepsy history: Prior seizures, epileptogenic lesions, and medications or conditions that lower seizure threshold plausibly increase the risk of the one serious theoretical harm of rhythmic gamma-frequency stimulation. Ultrasound neuromodulation trials routinely exclude participants with these features, which is also why the reassuring safety record cannot be extended to them.
Sex: The mechanistic work reporting inflammatory pathway suppression was conducted exclusively in female animals, and microglial responses differ by sex in both rodents and humans. There is no basis for asserting a sex difference in risk, and equally none for asserting equivalence; the honest position is that the risk profile has been characterised in one sex only at the mechanistic level and not at all clinically.
Age: Older skulls are denser and thicker, so a fixed transducer output delivers less energy to the brain and more heat to the bone, shifting the risk from central to local. Older brains also have more amyloid-laden vasculature and more baseline microbleeds. The one human age-stratified dataset in this technology class found benefit concentrated below age 70 without a corresponding safety signal above it.

Key Interactions & Contraindications

Ultrasound contrast agents (Definity, Optison, Lumason): Absolute contraindication within the elimination window. Circulating microbubbles convert neuromodulation-level pressures into cavitation nuclei and can open the blood-brain barrier. Mitigation: no sonication within 24 hours of any contrast-enhanced study.
Anti-amyloid monoclonal antibodies (lecanemab, donanemab): Caution bordering on contraindication. These are laboratory-made antibodies, given by infusion, that strip amyloid out of the brain. Both carry a documented risk of amyloid-related imaging abnormalities (ARIA, brain swelling or small bleeds visible on scans), and adding a second amyloid-mobilising stimulus without data is not supportable. Mitigation: no concurrent use; where either drug is under way, MRI surveillance per the drug’s own schedule and no ultrasound while ARIA is present.
Anticoagulants and antiplatelet agents (warfarin, apixaban, rivaroxaban, clopidogrel, aspirin): Caution. The clinical consequence of concern is intracranial haemorrhage should any mechanical bioeffect occur. Mitigation: baseline susceptibility-weighted MRI, exclusion above a defined microbleed count, and lower exposure limits.
Drugs that lower the seizure threshold (bupropion, tramadol, clozapine, high-dose theophylline): Caution. The consequence is seizure provocation by rhythmic gamma-frequency drive. Mitigation: dose stability for at least four weeks before starting, and no sleep deprivation or alcohol withdrawal around sessions.
Sedatives, anaesthetics and hypnotics (propofol, ketamine, midazolam, zolpidem): Monitor. Hypnotics are sleep-inducing medicines, and this group as a whole calms brain activity. These agents directly suppress or reshape gamma-band activity, so they do not create danger so much as abolish the intended effect and any ability to detect it. Mitigation: sessions scheduled away from sedative dosing, with medication state recorded alongside any EEG measurement.
Cholinesterase inhibitors and memantine (donepezil, rivastigmine, galantamine, memantine): Monitor; potentially additive. Cholinesterase inhibitors are standard dementia drugs that raise levels of acetylcholine, a signalling chemical involved in memory; cholinergic tone modulates gamma generation, so these may amplify entrainment. There is no evidence of harm from the combination and it is the expected background therapy in any dementia population. Mitigation: a regimen held stable for at least four weeks before and during any assessment period, so that changes can be attributed.
Over-the-counter anticholinergics and sedating antihistamines (diphenhydramine, doxylamine, dimenhydrinate): Monitor. These blunt cholinergic transmission and measurably degrade cognition and gamma activity, working directly against the intervention. Mitigation: substitution of a non-sedating antihistamine (loratadine, cetirizine), and none on session days.
Alcohol and caffeine: Monitor. Alcohol suppresses gamma-band power and fragments sleep; caffeine raises cortical excitability and gamma power, potentially confounding both response and any EEG readout. Mitigation: intake standardised before sessions rather than eliminated abruptly.
Supplements affecting cortical excitability or cholinergic tone (huperzine A, alpha-GPC, citicoline, high-dose Ginkgo biloba extract, nicotine-containing products): Caution; additive with the intervention’s intended cholinergic and excitatory direction, and therefore also additive with its seizure-threshold concern. Mitigation: stimulating cognitive-enhancement supplements withheld on session days, with any regimen fixed during monitoring periods.
Supplements affecting bleeding risk (high-dose fish oil, vitamin E above 400 IU daily, Ginkgo biloba, nattokinase): Caution; additive with anticoagulants on the haemorrhage-risk axis described above. Mitigation: these counted alongside prescription anticoagulants when eligibility is assessed.
Other neuromodulation (transcranial magnetic stimulation, transcranial direct or alternating current stimulation, 40 Hz light and sound devices, deep brain stimulation): Caution. Stacking synchronising interventions compounds the seizure concern and makes attribution impossible; implanted stimulators are a physical contraindication. Mitigation: one modality at a time, with a washout of at least two weeks.
Populations who should avoid this intervention: anyone with an implanted cranial or intracranial device (deep-brain-stimulation leads, cochlear implants, ventriculoperitoneal shunts, cranial plates); skull defect, craniectomy or burr hole at the intended acoustic window; intracranial haemorrhage within 90 days; ischaemic stroke within 90 days; active intracranial neoplasm (a tumour inside the skull); seizure within the past 12 months or a diagnosis of epilepsy; amyloid-related imaging abnormalities on MRI within the past 3 months; pregnancy; children and adolescents; and anyone who has received an ultrasound contrast agent within 24 hours.

Risk Mitigation Strategies

Imaging-based acoustic targeting before first exposure: Careful protocols acquire a computed tomography or ultrashort-echo-time MRI of the skull and run acoustic simulation to estimate the pressure and temperature actually delivered at the target. This mitigates both under-dosing through skull attenuation and the local over-exposure that causes thermal injury, and it is the difference between a characterised exposure and an unknown one.
Baseline susceptibility-weighted MRI with a microbleed threshold: Screening for pre-existing cerebral microbleeds and superficial siderosis (iron staining left on the surface of the brain by earlier bleeding) precedes exposure, with exclusion above the conventional threshold of four or more microbleeds used in amyloid-directed trials. This mitigates the intracranial haemorrhage risk that anticoagulation and vascular amyloid create.
Explicit thermal and mechanical exposure ceilings: Temperature rise is held at or below 2 °C, equivalently an absolute tissue temperature of 39 °C, session length is capped accordingly — up to 80 minutes at that band of the thermal index, falling to 10 seconds at the highest — and the mechanical index (a number printed by the machine that estimates how strongly a given setting will stress tissue mechanically) is kept at or below 1.9. This mitigates thermal injury and cavitation, and follows the published international biophysical safety consensus (Aubry et al., 2025); note that the consortium issuing that consensus counts ultrasound device manufacturers among its members, and its authors declare equity, consulting and grant relationships with ultrasound device companies, so it is not a disinterested party.
Duty-cycle and session-duration escalation: Cautious regimens start at 10–15 minutes per session for the first week rather than the one to two hours used in rodent protocols, extending only if scalp temperature, symptom reports and tolerability permit. This mitigates cumulative heating and headache, the two most commonly reported effects.
Active auditory masking and an active sham: Broadband masking noise matched to the 40 Hz pulse envelope is delivered during every session, and an active sham that reproduces the sound and sensation replaces a flip-over sham, in which the transducer is simply turned to face away from the head, for any self-assessment. This mitigates auditory over-stimulation and, more importantly, prevents mistaking a response to sound for a response to ultrasound.
Fixed acoustic coupling protocol: A standardised coupling medium is used, air bubbles are removed from the gel or water path before each session, and transducer contact is re-verified whenever the headset is moved. Air gaps reflect energy back into the scalp and are the commonest cause of local heating and of unexpectedly low delivered dose.
Medication and supplement review against seizure threshold and bleeding risk: Every agent that lowers seizure threshold or increases bleeding risk, including over-the-counter and supplement sources, is listed before starting, with four weeks of stability required. This mitigates the seizure and haemorrhage risks that are otherwise invisible until they occur.
Structured symptom log with stopping rules: Headache, scalp heat, nausea, dizziness, mood change and sleep quality are recorded after each session on a fixed 0–10 scale, with stopping rules for any new focal neurological symptom, any rating above 5 that persists beyond 24 hours, or any suspected seizure. This mitigates the tendency for mild, cumulative adverse effects to be normalised over a long course.
Clinical supervision rather than self-administration: The technique is confined to a registered trial or to the care of a clinician experienced in neuromodulation. Given that no device is cleared for this indication and that dose depends on individual skull anatomy, supervision is the mitigation for the entire class of exposure errors that self-administration invites.

Therapeutic Protocol

There is no established human protocol for 40 Hz ultrasound. The parameters below are those used in the animal studies that generated the evidence and in adjacent human ultrasound work, presented so that any proposed regimen can be judged against them rather than as a regimen to follow.

Core acoustic parameters: A carrier frequency of 0.25–2.0 MHz, amplitude-modulated or pulsed at a repetition frequency of exactly 40 Hz. The founding rodent protocol from Pierre Mourad’s group at the University of Washington used 2.0 MHz carrier, 40 Hz pulse repetition frequency, 400-microsecond pulses and a spatial-peak pulse-average intensity of 190 W/cm²; the most recent wearable-transducer work from the Shenzhen Institutes of Advanced Technology identified 2.14 W/cm² as optimal over 14 days.
Target and depth: The hippocampus and adjacent medial temporal structures, the regions light- and sound-based gamma stimulation cannot reach. Reaching them is the entire rationale for using ultrasound rather than a flickering lamp, and it is also what makes skull-corrected targeting non-optional.
Session length and course: Rodent work used one to two hours daily for 5 to 14 days. Human ultrasound neuromodulation sessions have almost universally been minutes rather than hours, so the animal-derived duration has no human precedent and should not be assumed transferable.
Competing approach: sensory gamma entrainment: 40 Hz light and sound delivered for one hour daily, developed by Li-Huei Tsai’s group at MIT and commercialised by Cognito Therapeutics, whose founders hold equity in the outcome. It reaches only cortical surface regions but has actual human trial data, including two-year tolerability. Neither approach can currently be described as the default: ultrasound has depth without human evidence, sensory stimulation has human evidence without depth.
Competing approach: transcranial pulse stimulation: Single ultrashort pulses delivered at 1–5 Hz to frontoparietal areas, developed by Roland Beisteiner’s group at the Medical University of Vienna in partnership with a device manufacturer, and tested in the largest sham-controlled ultrasound trial in Alzheimer’s disease. It is not gamma entrainment at all and rests on a different mechanism, but it is the only ultrasound neuromodulation approach with randomised cognitive outcome data, so it forms the practical comparator.
Best time of day: Morning or early afternoon. Gamma-band drive is alerting, the sensory-stimulation literature standardised on daytime sessions, and the one registered sleep-endpoint ultrasound study is examining daytime stimulation with night-time measurement. Evening sessions risk sleep disruption without any offsetting rationale.
Persistence of effect between sessions: Rodent hippocampal recordings show enhanced gamma power and coupling persisting roughly five days after the final session, which is the closest analogue to a half-life for this intervention and suggests that daily dosing may be more than the network requires.
Genetic polymorphisms influencing protocol choice: APOE4 status is the variant most likely to matter, both because carriers have more amyloid to act on and because they have the most vulnerable cerebral vasculature; carriers argue for lower exposure and closer imaging surveillance. No pharmacogenetic variant is relevant, since nothing is metabolised.
Sex-based differences in protocol: The key mechanistic dose-finding study was conducted in female animals only, so the 2.14 W/cm² figure carries an unstated sex qualifier. No human sex-stratified data exist for any ultrasound neuromodulation protocol.
Age-related considerations: Skull thickness and density increase with age, so an identical transducer setting delivers progressively less energy to the brain and more heat to the bone in older users. The only age-stratified human result in this technology class found cognitive benefit concentrated at 70 years and below, which argues for earlier application and for individually simulated rather than fixed output in older users.
Baseline biomarkers influencing response: Baseline EEG gamma power and demonstrated entrainment capacity are the most direct predictors, since the long-term sensory-stimulation data show benefit tracking retained entrainment. Plasma p-tau217 and amyloid status determine whether there is pathology available to modify.
Pre-existing conditions influencing response: Cerebral small-vessel disease, untreated sleep apnoea and prior stroke degrade baseline gamma activity and would be expected to blunt entrainment; treating sleep apnoea first is likely to do more for gamma power than any stimulation added on top of it.

Discontinuation & Cycling

Lifelong versus time-limited use: Every protocol tested to date is a defined course, not an open-ended therapy: five to fourteen days in animals, two weeks in the largest human ultrasound trial, and one to two years in the sensory-gamma extension work. Whether 40 Hz ultrasound would need to be continued indefinitely to sustain any effect is unknown, and the honest framing is that no maintenance schedule has been established because no acute effect has been established in humans.
Withdrawal effects: None reported or mechanistically expected. Nothing accumulates in tissue, no receptor is chronically occupied, and no dependence phenomenon has been described for any transcranial ultrasound protocol. The effect on brain rhythms decays over days rather than producing a rebound.
Tapering: Not applicable. Because there is no withdrawal syndrome, sessions can be stopped abruptly; the practical reason to taper would be to distinguish genuine benefit from expectation, which is better served by a planned break with blinded assessment than by a gradual reduction.
Cycling for maintained efficacy: Plausible but untested. Rodent data showing that enhanced gamma power and coupling persist about five days after the last session imply that daily administration may exceed what the network requires, and that intermittent blocks separated by rest periods would be a reasonable design to test. No cycling schedule has been evaluated in any species, and tolerance to gamma entrainment has neither been demonstrated nor excluded.
Planned interruption as an evaluation tool: A four-week break after a defined course, with the same cognitive and EEG measures repeated at the end of it, is the only practical way for an individual to separate a real effect from the substantial expectation effect that surrounds novel brain-stimulation devices.

Sourcing and Quality

Device availability: No device is cleared or approved anywhere for 40 Hz-pulsed transcranial ultrasound. Research groups use custom or laboratory-grade transducer systems; the human ultrasound neuromodulation devices that exist commercially (transcranial pulse stimulation systems, investigational focused ultrasound platforms from companies including Sanmai and BrainSonix) are not configured for 40 Hz gamma pulsing. There is no legitimate consumer product to source, and any device marketed as one is operating outside characterised parameters.
What to look for in a research or clinical system: Independent acoustic characterisation of the transducer output by hydrophone measurement, not manufacturer specification alone; documented spatial-peak pulse-average and spatial-peak temporal-average intensity, mechanical index and thermal index; and reporting that conforms to the international standardised reporting consensus (Martin et al., 2024), issued by the same manufacturer-inclusive consortium whose members stand to gain from the standards they set. A system that cannot state these numbers cannot be assessed for safety.
Skull-corrected targeting capability: The relevant capability is neuronavigation with individual imaging and acoustic simulation. Without it, delivered intensity varies by a large and unknown factor between individuals, which makes both the dose and the safety margin meaningless.
Coupling and consumables: Acoustic coupling gel or a degassed water path must be free of entrained air. Poor coupling is the commonest source of both local heating and silently absent dose, and it is the one quality variable a user can directly observe.
Avoiding ultrasonic devices sold for other purposes: Physiotherapy ultrasound units, cosmetic and cavitation devices, and hobbyist transducers operate at frequencies, intensities and duty cycles chosen for tissue heating or mechanical disruption, not neuromodulation, and none is designed for transcranial application. Repurposing them is the most likely route by which someone would come to real harm from this intervention.
Clinics offering ultrasound neuromodulation: The distinguishing questions are which protocol a clinic delivers and at what pulse rate. Clinics providing transcranial pulse stimulation are delivering 1–5 Hz single-pulse stimulation, which is a different intervention from 40 Hz gamma pulsing regardless of how it is described in marketing material.

Practical Considerations

Time to effect: Unknown in humans. Rodent studies report acute microglial changes within an hour, gamma-power changes during the session itself, and plaque reduction after five to fourteen daily sessions. The nearest human analogue, the transcranial pulse stimulation trial, measured cognitive change at one week, one month and three months after a two-week course, with effects still present at three months.
Common pitfalls: Confusing 40 Hz gamma pulsing with other ultrasound protocols that share the word “ultrasound”; assuming the flickering-light literature transfers wholesale to ultrasound; ignoring skull-dependent dose variation; failing to mask the audible pulse train and then attributing the result to neuromodulation; and treating amyloid plaque reduction in a mouse as though it were a cognitive outcome in a person.
Regulatory status: No regulatory authority has cleared or approved any device for 40 Hz-pulsed transcranial ultrasound. Low-intensity focused ultrasound devices are in investigational use under research protocols; transcranial pulse stimulation systems carry a European CE mark for Alzheimer’s disease in some jurisdictions but are not United States Food and Drug Administration (FDA) cleared for that use. The 40 Hz sensory-stimulation route, by contrast, has received FDA Breakthrough Device designation, which is a review-pathway designation and not a finding of effectiveness.
Cost and accessibility: Effectively inaccessible outside research. Research-grade focused ultrasound systems with neuronavigation cost in the six figures, clinical transcranial pulse stimulation courses are typically self-funded at a few thousand euros or dollars, and one developer has stated an intention to reach a sub-$500 home device, which remains an intention rather than a product. The imaging and simulation required for safe individual targeting add materially to both cost and complexity. None of the competing approaches is reimbursed by insurers or national health systems anywhere, so no institutional payer currently has a financial stake in favouring one over another; were reimbursement to arrive, the cost gap between a sub-$500 sensory device and six-figure imaging-guided ultrasound would give payers a systematic incentive to favour the cheaper sensory route, and that incentive would be a plausible source of structural bias in guideline formation and research funding.
Participation in research as the realistic route: For a reader who wants exposure to this intervention rather than to the idea of it, enrolment in one of the registered trials listed under Emerging Research is currently the only route that provides characterised acoustic dose, imaging-based targeting and adverse-event monitoring.

Interaction with Foundational Habits

Sleep: Bidirectional and potentially potentiating. Gamma-band activity and slow-wave sleep both decline with age, and deep sleep is when glymphatic clearance of amyloid is greatest, which is the same pathway 40 Hz stimulation is proposed to recruit. The direction of interaction is therefore expected to be additive rather than antagonistic. Practically, sessions are best scheduled in the morning or early afternoon, since gamma drive is alerting; a registered trial is specifically examining hippocampal ultrasound against sleep quality endpoints, and untreated sleep apnoea should be corrected first because it degrades the baseline rhythm the intervention aims to restore.
Nutrition: Indirect, with no depletion effect. Nothing is ingested and no nutrient is consumed, so the interaction runs entirely through cerebral metabolic state and vascular health. Nutritional patterns that reduce cerebral small-vessel disease and improve glucose regulation, such as Mediterranean-pattern eating, protect the vascular substrate on which any clearance mechanism depends. Alcohol suppresses gamma power and should be minimised around sessions; caffeine raises cortical excitability and gamma power and should be held constant rather than varied, so that it does not confound assessment.
Exercise: Indirect and likely potentiating, with no blunting concern. Unlike interventions that interfere with training adaptation, ultrasound has no known interaction with muscle or systemic recovery. Aerobic exercise raises cerebral blood flow and brain-derived neurotrophic factor and independently improves cognition, and in mouse models physical exercise combined with 40 Hz light flicker reduced amyloid and tau more than either alone, which is the closest available evidence for additivity. Timing is unconstrained; the only practical consideration is that vigorous exercise immediately before a session adds scalp sweat, which degrades acoustic coupling.
Stress management: Indirect and bidirectional. Chronic stress and elevated cortisol impair hippocampal function and degrade the network the intervention targets, so unmanaged stress is expected to blunt any response. Conversely, hour-long daily sessions of sitting still with a headset are themselves a scheduling and adherence burden that some find stressful, which is a real adherence consideration given that the sensory-gamma literature shows benefit tracking sustained daily use. Meditation practices that increase gamma-band activity have been described in long-term practitioners, raising the untested possibility that the two act on the same rhythm from different directions.

Monitoring Protocol & Defining Success

Because no human protocol is established, monitoring serves two purposes: detecting the physical harms that ultrasound can plausibly cause, and establishing whether anything is happening at all. Baseline testing should be completed before the first session and should include structural and susceptibility-weighted brain MRI, a resting EEG with a 40 Hz entrainment challenge, a formal cognitive battery, and the blood panel below, so that any later change has a reference point rather than a recollection.

Biomarker	Optimal Functional Range	Why Measure It?	Context/Notes
Plasma p-tau217	Below assay-specific positivity threshold; stable or falling on repeat	Tracks Alzheimer’s pathology and is the most responsive blood marker to disease-modifying effects	p-tau217 is a phosphorylated form of the tau protein. Fasting not required; use the same assay and laboratory for every draw, as thresholds are not interchangeable. Repeat no more often than 6-monthly
Plasma Aβ42/Aβ40 ratio	Above assay-specific threshold; conventional labs often report no range	Indicates whether amyloid pathology is present and therefore whether a plaque-directed mechanism has any substrate	Aβ is amyloid-beta, the protein fragment that forms plaques. Highly sensitive to pre-analytic handling; draw and process to a fixed protocol. Best paired with p-tau217
Neurofilament light chain (NfL)	Age-adjusted; below the 75th percentile for age	General marker of neuronal injury; a rise would be the earliest signal of harm from an unvalidated exposure	Rises with age and with kidney impairment, so interpret alongside eGFR (estimated glomerular filtration rate, a measure of kidney function). Conventional labs rarely offer a functional range
Glial fibrillary acidic protein (GFAP)	Age-adjusted; stable across repeats	Marker of astrocyte activation; the closest available blood proxy for the neuroinflammatory response the intervention is meant to reduce	Best drawn with NfL from the same sample. A sustained rise argues against the intended anti-inflammatory direction
High-sensitivity C-reactive protein (hs-CRP)	Below 1.0 mg/L	Systemic inflammatory load, which independently degrades cognition and would confound any inflammatory readout	Conventional cardiovascular cut-off is below 3.0 mg/L, materially looser than the functional target. Invalid within 2 weeks of infection or injury
Homocysteine	5–8 µmol/L	Elevated levels accelerate brain atrophy and are a correctable competing driver of cognitive decline	Conventional labs report up to 15 µmol/L as normal, which is far above the functional target. Fasting sample preferred; pair with B12 and folate
Haemoglobin A1c (HbA1c)	4.8–5.4%	Glycaemic control drives small-vessel disease and cognitive decline and is a far better-evidenced lever than any stimulation protocol	HbA1c reflects average glucose over about 3 months. Conventional threshold for concern is 5.7%. No fasting required
APOE genotype	Not a range; result is ε2/ε3/ε4 status	Determines vascular amyloid burden and therefore the risk tier for any amyloid-mobilising intervention	APOE is the apolipoprotein E gene. Tested once, never repeated. Result has insurance and psychological implications that warrant discussion beforehand
Cerebral microbleed count on susceptibility-weighted MRI	Zero to three; four or more is a conventional exclusion	Direct measure of the vascular fragility that determines haemorrhage risk	Requires susceptibility-weighted or gradient-echo sequences; standard clinical MRI protocols often omit them and must be requested explicitly
Resting EEG gamma power and 40 Hz entrainment response	Individually referenced; entrainment response present and stable or increasing	The only direct measure of whether the intended mechanism is engaged in that individual	EEG is electroencephalography, a scalp recording of brain rhythms. Highly sensitive to caffeine, alcohol, sedatives and time of day; standardise all four. Measure at the same clock time each session

Ongoing monitoring should follow a fixed cadence: symptom log and scalp inspection after every session; EEG entrainment response and symptom review at 1 week and 4 weeks; cognitive battery and blood panel at 3 months and 6 months; and susceptibility-weighted MRI at 6 months, or immediately if any new neurological symptom appears. Thereafter every 6–12 months for as long as sessions continue.

Qualitative markers should be recorded alongside the objective measures, because they are what a person actually experiences and because they are also the measures most vulnerable to expectation effects:

Word-finding and name retrieval — the everyday failure most people notice first, and the one most closely tied to the memory networks being targeted.
Sleep quality and night-time restlessness — reported to improve with sensory gamma stimulation and specifically under study for ultrasound.
Sustained attention and mental fatigue late in the day — sensitive to gamma-band function and easy to rate consistently.
Mood and anxiety — both a possible benefit and a documented adverse effect, so it must be tracked in both directions.
Scalp sensation, heat and headache after sessions — the earliest signal of an exposure or coupling problem.

Success, defined honestly for an intervention at this stage, means a measurable and sustained 40 Hz entrainment response on EEG, no adverse trend in neurofilament light chain or microbleed count, and stability rather than decline on the cognitive battery over 12 months. Improvement beyond baseline is not a reasonable expectation from the current evidence, and treating it as the success criterion invites the misattribution that the entire auditory-confound literature warns about.

Emerging Research

The research most likely to change this picture is not more mouse work but the first human exposure to gamma-frequency ultrasound, and the first honest attempts to determine whether ultrasound neuromodulation does anything that its own audible pulse train does not.

Low-intensity focused ultrasound in mild cognitive impairment and mild Alzheimer’s disease: The largest registered trial of low-intensity ultrasound in this population, at the University of California, Los Angeles, with a planned enrolment of 144 and imaging primary endpoints: blood flow through brain tissue and the strength of the connections between brain regions, both measured on MRI. Recruiting since 2022 — NCT05417555.
Personalised transcranial focused ultrasound for mild neurocognitive disorder and healthy ageing: A double-masked crossover Phase 1 study at the Medical University of South Carolina, 25 participants aged 50–85, targeting the hippocampus with MRI-guided navigation, with a face-name associative memory task and EEG as co-primary endpoints. This is the design that comes closest to answering whether deep ultrasound targeting can move human memory circuits, and it includes cognitively healthy older adults rather than patients only — NCT06718140.
Hippocampal ultrasound for cognitive and sleep dysfunction in preclinical Alzheimer’s disease: A Chinese University of Hong Kong study enrolling 20 participants with sleep quality and delayed word recall as primary endpoints, beginning in 2026. It is the first registered attempt to use ultrasound neuromodulation before clinical dementia is present, which is the timepoint that matters most for a longevity-oriented reader — NCT07298876.
A cautionary registry signal: A Phase 1/2 study of ultrasonic neuromodulation for cognitive impairment at the University of Utah, with amyloid positron emission tomography among its primary outcomes, was terminated after enrolling a single participant, the registry recording the reason as an institutional review board administrative closure — NCT06135051. Whatever lay behind that closure, a study that stops at one participant yields no usable data, and it is a reminder that the registry contains abandoned efforts as well as promises.
Commercial development beyond dementia: A 60-participant industry-sponsored trial of transcranial focused ultrasound in Parkinson’s disease, with serious adverse device events as a co-primary endpoint alongside motor scores — NCT07207122. Its safety data will inform the field regardless of its efficacy result, and its sponsor is a device company with a direct financial interest in a favourable outcome.
Research that could weaken the case — the auditory confound: Kop et al., 2024 showed a widely replicated ultrasound effect to be an artefact of the audible pulse train. The decisive future experiment is the same manipulation applied at 40 Hz with a tissue rather than behavioural endpoint. Until it is run, the possibility remains that 40 Hz ultrasound is an expensive way to deliver 40 Hz sound.
Research that could strengthen the case — no long-term auditory signature: Mitterwallner et al., 2026 re-analysed the largest sham-controlled ultrasound trial in Alzheimer’s disease and found no lasting auditory-network activation and no relationship between auditory measures and cognition, arguing that sustained effects are not an auditory artefact. The analysis was performed by the group that generated the original data and holds the reputational stake in it, so independent replication is required before it settles the question.
Research that could reframe the frequency itself: Wang et al., 2026 compared 5, 40 and 80 Hz amplitude-modulated ultrasound and found 5 Hz sinusoidal modulation superior to 40 Hz for driving hippocampal network coupling and spatial memory. If this replicates, the transfer of “40 Hz” from light to ultrasound may turn out to have been an assumption rather than a finding, and the optimal ultrasonic rate may be a theta rather than a gamma frequency.
Multimodal combination: Gao et al., 2026 reported that combining 40 Hz ultrasound with 40 Hz light produced greater gamma power and cross-regional synchronisation than either alone, along with reduced amyloid deposition in mouse models. Combination approaches are the most likely near-term direction if single-modality effects prove too small to detect in humans.
The comparator that will set the bar: Chan et al., 2025 followed five patients through two years of daily 40 Hz audiovisual stimulation, reporting retained entrainment, less decline than matched controls, and falls in plasma p-tau217 in the two patients sampled. It is five patients, open-label, from the group whose members founded the company selling the device, and it is nonetheless the strongest long-term human data any 40 Hz approach has produced. Ultrasound must eventually be measured against it.

Conclusion

40 Hz ultrasound is a proposal, not yet a therapy. It takes a rhythm that flickering light and pulsing tones have been shown to impose on the brain, and delivers it with sound waves that can be focused deep inside the head, where light and tones cannot reach. In animals bred to develop the protein deposits of memory disease, this reliably clears those deposits, restores the fast brain rhythm that ageing erodes, and improves performance on memory tasks. Four separate laboratories have found much the same thing.

Not one person has yet received it. Every claim about what it might do for human health or lifespan is an extrapolation across a species barrier that similar ideas have repeatedly failed to cross. The safety record borrowed from brain ultrasound in general is reassuring but covers minutes of exposure, not the daily hours the animal work implies. A serious body of evidence shows that some effects of pulsed brain ultrasound come from the audible buzz it makes rather than from the sound entering the brain, and at this particular rhythm that confusion is unusually hard to untangle.

Much of the research comes from groups holding patents or company equity in the outcome, on all sides of the question, and so does the expert body that sets the safety ceilings. What exists is a coherent idea about how it would work, with consistent animal support, no human results, and an unresolved argument about whether it works the way it appears to.

Top - Benefits - Risks - Protocol

40 Hz Ultrasound for Health & Longevity

Motivation

Recommended Reading

Grokipedia

Examine

ConsumerLab

Systematic Reviews

Mechanism of Action

Historical Context & Evolution

Expected Benefits

Low 🟩

Reduction of Amyloid-β Plaque Burden

Microglial Engagement with Amyloid Plaques ⚠️ Conflicted

Entrainment of Gamma-Band Brain Rhythms

Improved Memory Performance in Disease Models

Speculative 🟨

Reduction of Neuroinflammatory Signalling

Enhanced Glymphatic Waste Clearance

Preservation of Cognitive Function in Healthy Ageing

Sleep Quality and Circadian Consolidation

Benefit-Modifying Factors

Potential Risks & Side Effects

Medium 🟥 🟥

Transient Headache, Scalp Discomfort and Neck Pain

Audible 40 Hz Tone and Off-Target Auditory Co-Stimulation ⚠️ Conflicted

Low 🟥

Nausea, Vomiting and Dizziness

Mood Deterioration and Transient Anxiety

Tissue Heating and Mechanical Bioeffects

Speculative 🟨

Seizure Provocation

Sustained Microglial Activation

Unintended Blood–Brain Barrier Permeability

Cumulative Effects of Multi-Year Daily Exposure

Risk-Modifying Factors

Key Interactions & Contraindications

Risk Mitigation Strategies

Therapeutic Protocol

Discontinuation & Cycling

Sourcing and Quality

Practical Considerations

Interaction with Foundational Habits

Monitoring Protocol & Defining Success

Emerging Research

Conclusion