Things I’ve learned, published for the public benefit
Hope This Helps header image

Does listening to a 40 Hz tone “clean up” the brain in Alzheimer’s patients?

In 2012, I made a Web-based tone generator with the goal of helping tinnitus patients determine the frequency of their tinnitus to better target therapy. Since then, I have heard from people using my generator to teach physics, practice violin, drive away carpenter bees, tune DIY speakers, analyze room acoustics, calibrate vintage synthesizers, cause mischief in class with frequencies the teacher can’t hear, and even open a portal to Sedona, AZ. Far be it from me to take away from all these worthwhile applications, but last week, I got a message from Dennis Tuffin (of Devon, England), describing a new use for my generator which may very well trump everything else:

For the past 7 weeks I have been using your tone-generator for a purpose I wouldn’t think you had envisaged but about which I am sure you will be interested.

I have been following up on some research which my daughters had done about the treatment of Alzheimers by using a 40Hz flickering light source or alternatively a 40Hz sound source. There is sparse info on the net about these experiments though there is a recent piece about it. [here Dennis is referring to this paywalled article]

So I have been trying the sound therapy on my wife who is in the advanced stages of Alzheimer’s and to my surprise after 8 days she started to show small signs of being more mentally alert than before. So I have continued to use your tone generator using a 40Hz sine wave for about an hour each day. (I’ve recently started to do it twice a day for slightly shorter sessions). I found it necessary to connect external speakers to my laptop in order to pick up such a low note and to run it at a level of between 46-54 decibels so that she hears it wherever she is in the room. (Dementia sufferers get very fidgety!). So now 7 weeks on the improvement in her awareness has continued to the point where she is starting to be able to put a few words together and to respond to questions neither of which she has been able to do for nearly a year. Her odd physical habits have not been changed so far but she is definitely walking better and not shuffling her feet as she used to. Surprisingly, she is also sleeping better and not suffering as much with the sleep apnoea problem that she’s always had.

Photo of laptop and a small speaker on a kitchen countertop

The setup used by Dennis. The black box on the left is the external speaker.

Of course I expect there to be a limit to this progress as in the 8 years since my wife was first diagnosed her brain will have shrunk considerably so I do not expect her memory to return but on the other hand my wife’s quality of life has been improved.

To date I have not gone public on this and only close family have known but by the end of another week when it will be 8 weeks since we started I think I would like to spread the word and hopefully prompt a few professionals to do more proper research.

The science so far

  • It has been known since at least the 1980s that cognitive activity triggers brainwaves (wave-like patterns of activation) at a frequency of 40 Hz in humans and other mammals.
  • In 1991, researchers from the NYU Medical Center discovered that Alzheimer’s patients have reduced 40 Hz brainwaves compared with healthy people. (paywalled paper)
  • In 2016, MIT’s Alzheimer’s group did experiments on transgenic mice with early Alzheimer’s disease and found that exposing them to a light flickering at a frequency of 40 Hz (40 times a second) for 1 hour a day for 7 days causes an almost 60% reduction in β-amyloid plaques, which are a molecular hallmark of Alzheimer’s. Flickering at 20 Hz and 80 Hz did not have the same effect. An important qualification here is that the effect was limited to the visual cortex, which is not significantly affected in human Alzheimer’s patients. Here’s an accessibly written report in The Atlantic and here’s the original paper (published in Nature) if you’re strong in science-speak. MIT also made a video about the findings.
  • In March 2016, scientists at the University of Toronto published the results of a small, placebo-controlled pilot study (paywalled paper), in which they exposed 20 Alzheimer’s patients to a 40 Hz sound. After six 30-minute sessions (done twice a week), the patients’ average score on the 30-point SLUMS scale improved by 4 points, while the placebo group did not improve. It should be noted that the “dosage” of the treatment was rather low, which may explain the modest results.
  • In January 2017, Cognito Therapeutics, a company formed by some of the members of the MIT team, started conducting preliminary trials to assess the safety of exposing AD patients to simultaneous flickering lights, an audio tone, and vibrations – all at 40 hertz.
  • In January 2018, the New Scientist reported (paywalled article) that the same MIT team achieved even better results by playing mice a 40 Hz sound. β-amyloid plaques shrank by about 50% in the auditory cortex and – crucially – in the hippocampus, perhaps because the two areas are close to each other. This would be a very important discovery, because the hippocampus is the region of the brain which is involved in forming memories. It is the hippocampus that suffers the most damage in human Alzheimer’s patients. According to the magazine, these results were presented at the Society for Neuroscience conference in Washington in November 2017. However, the published paper described a significantly different protocol (see below), so it is likely that the New Scientist didn’t get the details right.
  • In July 2018, the International Journal of Alzheimer’s Disease published the results of a pilot study in which 6 human patients were exposed to a 40-hertz flickering light bulb for 2 hours a day for 10 days. The therapy was administered in a home setting by the patients’ caregivers. No difference in β-amyloid plaque was found after therapy. If there was an effect, it must have been smaller than 20%, which is not comparable to the 50% reduction seen in mice.
  • In March 2019, Cell published another (paywalled) paper about another study done by the MIT group. Here’s a NYT article about it. Here are the main points:
    • After mice were exposed (for 7 days, 1 hour a day) to a series of clicks repeating at a frequency of 40 Hz, the amount of amyloid plaque in their auditory cortex and their hippocampus was reduced by about 40%. The mice also did better on several tasks involving the use of memory.
    • The clicks were 10 kHz waves, 1 millisecond long, repeated 40 times a second (each cycle had a 1 ms tone followed by 24 ms of silence). In communication with me, one of the paper’s authors said that pure 40 Hz tones were not used because mice cannot hear tones of such a low frequency.
    • When this auditory treatment was combined with light pulses at 40 Hz, microglia (“brain cleaner cells”) started clustering around amyloid plaque, and the reductions in plaque extended to parts of the prefrontal cortex (an area related to functions like attention and short-term memory). This effect was not observed with either audio or light treatment alone.
  • In 2018, Cognito Therapeutics started three clinical trials of their device (called “GammaSense”) that combines visual and auditory stimulation. The device is essentially LED goggles with over-the-ear headphones.
  • [NEW] In 2019, Dr Tsai’s MIT group started two small trials of another device – a white screen illuminated with LEDs flashing at 40 Hz combined with a soundbar emitting a 40 Hz “buzzing” or “humming” sound. Here is a summary of Dr Tsai’s presentation on it and here are the pages for the two trials on
  • [NEW] In April 2021 at the AD/PD conference, Cognito presented preliminary results of the GammaSense trials, most notably, the Overture trial (news story from, which is designed to evaluate efficacy. 46 patients with Alzheimer’s Disease received a daily 60-minute session with the GammaSense device for a period of 6 months. The control group comprised 28 patients who received sham treatment (i.e. a placebo-like device that isn’t supposed to do anything). The study seems to tick all the boxes for good scientific research: it’s randomized and multi-center. Although it’s not strictly double-blind, the subjects, their caregivers and the people who rated the patients’ cognitive functions were blinded (i.e. they did not know who has received real treatment and who the sham). Back when this trial was announced, I wrote that it was rather small and the only way it could show anything is if the device had a large effect. Well, the results are decidedly mixed, and I can pretty confidently say that GammaSense didn’t have the sort of huge effect everyone was hoping for. There was no significant difference between GammaSense and placebo on 3 out of 5 cognitive tests used in the study, and GammaSense came out slightly ahead on two of them. Brain tissue loss also seemed to be slower in the treatment group, though that doesn’t have to mean anything, considering that the control group was older than the treatment group. I would definitely like to see a proper, peer-reviewed paper rather than just a presentation – someone should go over Cognito’s work and make sure the results are statistically valid. The researchers collected amyloid PET scans, CSF, and blood samples, but have yet to analyze them. I do like the fact that cognitive function was assessed, not just amyloid deposits – we already have experimental drugs that remove β-amyloid, but don’t do anything when it comes to actual dementia. Too bad the results have fallen short.
  • Cognito is not giving up and has announced another, larger study, which is to start in the second half of 2021.

Further reading/listening

Update from Dennis (March 2020):

I’ve gotten in touch with Dennis Tuffin to ask about his long-term experiences with 40 Hz therapy. Here is his response:

I reached a point when I knew that my wife’s condition was worsening so after [about 8 months] I gave up using the sound treatment. (…) My wife was already losing verbal capacity and was decidedly becoming slower in her movements, so it’s not the case that stopping the treatment caused these things as they were already happening. I was enthusiastic about it, because there appeared to be a noticeable improvement in her capabilities for that 8 or 9 month period, so noticeable that it was remarked on by almost everyone that she knew. So I am still sure that it helped, if only for a limited period and I think that it’s also likely that if it had been applied from the beginning of her diagnosis it may have had an even longer lasting effect. I’d still say give it a try – it costs nothing.

Dennis also told me that he had recently learned that his wife has vascular dementia in addition to Alzheimer’s Disease. I mention this, because that fact might have some bearing on the effectiveness of sound therapy.

(Don’t) try this at home

Let’s say you are not dissuaded by the lukewarm results from the Cognito trials. You’ve decided that you’ve got nothing to lose and you want to try some kind of do-it-yourself auditory therapy. What kind of tones should you use?

Dennis, the reader from the UK who piqued my interest in this subject, used a pure 40 Hz tone.

According to this AlterNet article (later reprinted by The Salon), a pure tone was used in the preliminary safety study done by Cognito in early 2017.

It appears that the most recently published MIT study on mice used series of clicks (despite previous reports) rather than tones. The New York Times quotes Dr. Tsai, who worked on that study, as saying “your brain seems to be able to perceive clicks more than a tone”, which would seem to indicate a preference that’s not exclusive to mice. However, in response to my inquiry, another co-author of the paper, Ho-Jun Suk, said that 40 Hz pure tones were not used because mice cannot hear tones of such a low frequency.

The New York Times and the Boston Globe published articles about the MIT mice study, including links to audio samples of the stimuli that were used by the researchers. Unfortunately, I have discovered that neither sample represents accurately the audio waves that were played to the mice. The clicks in the published samples are smeared in time (closer to 2 ms) and are not pure 10 kHz tones. Ho-Jun Suk has confirmed that they do not match the source signal. (I don’t know how the NYT and Boston Globe managed to mangle the files so badly, but it’s not because of compression – I tried it on the same encoder and the same parameters that were used by the NYT and it did not distort the signal very much at all.)

This 2021 report from Alzforum uses the words “hum” and “buzz” (interchangeably) to describe the sounds used in the two ongoing MIT trials of the screen+soundbar device. This would seem to indicate that those trials use a pure 40 Hz tone, as a series of clicks doesn’t sound like humming or buzzing. Furthermore, it is reasonable to assume that if the researchers wanted to play a simple series of clicks, they would have used a smaller and more economical speaker, not what is described as a “high-fidelity soundbar”. However, the same article uses the word “buzz” when discussing the 2019 mice study, and we know that one used clicks, not tones, so we should be careful about how much we infer from the phrasing in that particular report.

If you are thinking about using clicks rather than pure tones, I would not recommend using 10 kHz clicks because human ears are not very sensitive to that range of the spectrum. Something like 3 kHz (where human hearing works the best) would probably be more sensible.

It would be very interesting to know what sort of tones are being used in the now ongoing human trials. (If you are in the trials or know anyone who’s in them, please let everyone know in the comments section.)

Technical advice for playing pure 40 Hz tones

A series of clicks is not particularly demanding – it can be played on anything. If you want to try playing a pure 40 Hz tone to someone with Alzheimer’s, here’s some technical advice:

Getting a 40 Hz tone is easy – you can use my frequency generator. (Please note I do not take responsibility for the purity of the produced tone, as it is generated by your Web browser – though I think it should be fine. By the way, I am also not a doctor and I am not giving medical advice or offering any medical product here.)

You will need decent speakers. 40 Hz is a very deep bass tone – the kind of rumbling tone that you feel in your body as much as you hear it. Small speakers, such as laptop speakers or small computer speakers, don’t go that low. If you try anyway, you will either hear nothing, or you will hear mostly – or only – distortion. What is distortion? It’s a higher-pitched, buzzing noise that speakers make when you push them too hard.

Photo of a bookshelf (monitor) speaker

A bookshelf speaker (photo: D. Cedler)

Bookshelf speakers will do 40 Hz, but their output at that frequency will be significantly reduced, so you will need to turn up the volume significantly, and they will produce easily audible distortion. Because the ear is more sensitive to high frequencies, the distortion may be subjectively louder than the fundamental 40 Hz tone (!), and may make the sound harder to tolerate, thus limiting the volume (and possibly the therapeutic effect).

The best solution is a high-quality subwoofer. It won’t be distortion-free, but you can expect the distortion to be 2–3 times quieter than with bookshelf speakers. This will give you as pure a tone as you can get. If you don’t care about playing music, you can get just a subwoofer (without any other speakers) and connect it to your computer or mobile device.

A neat trick to amplify the bass output of any speaker is to place it against as many walls as possible. For the maximum boost, put the speaker(s) on the floor, in a 3-way corner between two walls and the floor – that way, it will be adjacent to three surfaces.

How important is sound quality? It’s hard to say. Dennis seems to have had great results with cheap computer speakers. It is not known to what extent the therapeutic effect depends on volume or the presence of distortion. On the other hand, if you use small speakers, it won’t be obvious whether they’re actually playing 40 Hz or just distortion – so it’s worth getting something bigger just to be on the safe side.

Can you use headphones instead? It’s hard to say with certainty, as a 40 Hz tone played through your speakers will not just be heard with your ears – it will also be felt in your whole body. With headphones, the effect is strictly auditory. However, so far I haven’t seen any specific scientific reasons to suggest that this difference is important, and in fact headphones were used in the initial safety studies commissioned by Cognito. If you decide to use headphones, make sure they can do 40 Hz. The earbuds that came with your smartphone are probably not the way to go here. HeadRoom has a database of frequency response graphs for high-quality headphones, so you can check how loud a given model is at 40 Hz. Want a specific recommendation? Get the Koss Porta Pros (, They’ll do the job, they’re the most comfortable headphones I’ve used, and – at $40 – they’re tremendous value.

Call for comments

If you or your loved one has Alzheimer’s disease and you have tried 40 Hz sound therapy, please share your experiences – whether positive or negative – in the comments section below.

152 Comments so far

  • Chris

    Hi, Thanks for this! very interesting articles here.

    Has anyone tried with a Deep Didgeridoo?

    Here’s a recording of one that was crafted to resonate at 40 hz.

    It’s a free track on bandcamp >>

  • Power

    Thanks for a great summary on an important topic. I have used several online 40hz tracks. The two that have worked best for me are:
    and the original clicking track (allegedly) used in the MIT study:
    I’d be interested in your view of whether either of these youtube recordings actually represent a 40 Hz frequency! I use garden variety wired earbuds to listen on my smartphone.
    I use these as a 50-something individual who, like most of us in midlife, has experienced “subjective cognitive decline”, meaning I personally can discern that the old noggin is getting cobwebs, but have not been diagnosed with any official cognitive impairment. Since cognitive diseases including dementia and Alzheimers are progressive, early intervention, even to ‘clean out the cobwebs’ may help forestall or foreclose these ‘clinically official’ diagnoses that can come with time. It may be ‘too little, too late’ to start only once into advanced Alzheimers (imho).
    My results: I listen at night for about an hour before falling asleep (and sometimes fall asleep during). I get almost instant clarity the next day. Quicker thinking, reading, and verbal acuity; a general feeling of ‘hey, this is how my brain function felt about 10 years ago!’ Strangely, if I use it too frequently (say every single night for several weeks on end) it may become counterproductive, and I wake up feeling slightly fatigued with less clarity. So I take a break for a while before starting up again, and again notice almost immediate benefits in cognitive acuity.
    I would love your thoughts on the above two tracks, and on whether you have found better options?
    Thank you.

    • Tomasz P. Szynalski

      I checked the the first YouTube track and it’s not a pure 40 Hz tone – there is some noise and harmonic overtones added, which are quite audible. But it does contain mostly 40 Hz.
      I think it’s possible it simply helps you fall asleep better (and that any other frequency, say 50 Hz, would work just as well) and the cognitive improvements you’ve experienced are just due to better sleep. I know I feel a whole lot sharper when I get a good night’s sleep!

      • Power

        …or, maybe “mostly 40 hz” is good enough to stimulate a biological response. I don’t have problems sleeping or feeling refreshed on waking generally (ie. without listening to the tracks) so that’s not it.
        It is a noticeable difference from just sleep rejuvenation, and most definitely attributable to the listening – that said, I’ll try to find a ‘purer’ 40 hz for the long term. Any thoughts on the usability of the second track? (MIT one)

        • power

          …or will try your tone generator. Thanks so much

        • Tomasz P. Szynalski

          I doubt that a purer 40 Hz tone will make a difference; after all, it will still get mixed with sounds in your environment, so it will never be totally pure.
          About the “MIT” track, I don’t have anything to say other than what I’ve already written in the blog post and the comment section.

    • tom hunt

      you seem to think the mit clicks are the alleged sound recorded by MIT
      i presume the clicks are higher frequencies to mask the lower tone of 40Hz

      anyway, since i sense some skepticism on your part if these were the true clicks used, have you found a tract clicking with 40 Hz that you do trust as being effective and really 40Hz? if so would you send me a copy?

      i listened to the one from MIT for less than a minute, and when i took off the cheapest headphone ever made. i heard a diminishing tone for a minute that did not exist before i turned on the tract with clicking
      thanks for doing all the work

  • Mack Winston

    For clicks you don’t really need to overthink it – just use a square wave rather than a sine wave. A 40 Hz square wave will also have the higher frequency components (a square wave after all is a composite of the fundamental frequency (in this case 40Hz) and each odd harmonic, each one 1/3rd of the amplitude of the previous). So even if you can’t hear a 40Hz pure sine wave well, you’ll hear a 40Hz square wave very easily.

Leave a Reply to Juliet