I am a mouse
February 6, 2017 6:10 PM   Subscribe

How do I buy/set up a thing that will flash LEDs at me at 40 hz frequency?

The idea is here. Apparently one mouse study showed very impressive results for using light to reverse amaloid plaques and Alzheimer's symptoms. (Giddy Radiolab coverage for more.)

I'm not banking on this working, obv, but I don't see a downside to trying and I basically live in terror of Alzheimer's. So I want to try.

I looked this up online, a few people have tried to do this, but the instructions I'm finding feel very complicated, involve electrical diagrams I don't understand and references to arduinos and other shit I know nothing about. Can you explain this to me like I'm someone with early dementia, which I kind of think I am?

Ideally, I would just buy something off the shelf. When I heard the paper's author interviewed, I think on that same Radiolab (who can remember anything? Not me!) she mentioned that she bought Christmas lights that she could adjust to flash at that frequency. I would ideally like that solution, but will work with whatever I can.

Second best option: how to I sign up to be part of human trials for this therapy?

posted by latkes to Health & Fitness (23 answers total) 8 users marked this as a favorite
If you can't find anything off the shelf, maybe you could get in touch with a Maker Space in your area, and find someone who could build it for you. Even if you had to reimburse them for the parts, it should be fairly cheap. I know that in the past I have been in search of interesting projects to work on, and I'd have jumped at something like that, with a chance to practice my skills and help someone at the same time.
posted by primethyme at 6:43 PM on February 6 [1 favorite]

Cool idea. There are several maker spaces in my region (SF Bay Area). I guess I could just... walk in and ask people? Feels weird but maybe worth a try?
posted by latkes at 7:03 PM on February 6

An Arduino is a great choice for prototyping something like this, even if they're quite unfamiliar. With an Arduino, a USB cable, a couple LEDs and resistors, a bit of wire, and a breadboard to plug stuff in, you'd be set. Just copy and paste some Arduino code someone more familiar writes for you, push "upload", done. The code would essentially say, "40 times a second, toggle the connected LEDs on and off".

Which is to say, for someone with a bit of electronics experience, this would be pretty straightforward.
posted by silentbicycle at 7:05 PM on February 6 [1 favorite]

I could probably do this, and I kind of want to try, for my own purposes as well as yours. If you're interested, let me know, and I might have time to work on it tomorrow.

If it works out, I'd be happy to send you a working model - it would be pretty cheap to make. You can let me know if there are particular wavelengths of LED you think would work, or just white light (cool/warm/whatever) (I haven't read the whole article yet). I also will need to know if you think a particular brightness is called for.

I'm pretty sure I can get this to work with just some LEDs, a 555 timer chip, some resistors, and a 5v or 9v power supply - no Arduino required.

This won't necessarily be as bright as a room light - please let me know what kind of setting you're thinking of trying.
posted by amtho at 7:14 PM on February 6 [5 favorites]

This is a dead simple thing to whip up with a 555 timer IC, a couple resistors, a cap, and the LEDs. Plus a bread board to plug everything into and a wall-wart transformer to power it. It used to be a very common first IC project.

There are kits available if you can solder.

Here's another kit using discrete transistors instead of the 555.

The problem with flasher kits is you need someway of dialing in the 40 Hz value. This is easy with a oscilloscope but hard otherwise. I bet there is an app for it but I couldn't find one.

The only down side to the 555 is it can only drive 200mAs but you can use the output of the timer circuit to drive a power transistor instead of the LED to flash a bank of LEDs of pretty much any size you want with the right transistor. A 12V LED portable work light might be a good thing to hack up for this purpose.

The bicycle flasher circuit at the bottom of this page for example does this and could be completely implemented on a breadboard (no soldering just plug things in). Leave off the transistor and LEDs on the bottom half (everything after the 160 resistor) and using the values c1=0.1 µF R1=3.3 KΩ R2=180 KΩ will give you a 39.7192 Hz flash1 and a 50.5% on time. Power with either the suggested C/D batteries or a 6V wall-wart (plug in transformer).

1Common resistors and capacitors have a fairly wide tolerance so if you want exactly 40 hz you'll either have to pay for more expensive 1% resistors and caps or test a bunch of resistors to get one that is closest to the rated value. My understanding is there is some leeway in the 40Hz value anyways. Alternatively use potentiometers for R1 and R2 to dial in the exact resistance.
posted by Mitheral at 7:25 PM on February 6 [1 favorite]

Would a simple web page that blinks a box on your screen at this frequency work?
posted by miyabo at 7:41 PM on February 6 [2 favorites]

I can't find information about the duty cycle (is the light on 50% of the time / off 50%, or on 10% / off 90%), but this article says:
She described the flickering light that kick-starts the process as “very low intensity, very ambient, very soft light.”

“You can hardly see the flicker itself, actually,” she added.
Also, it's only a proof-of-concept, they're not all rushing out to buy 40hz lights, etc., and here's detailed critique. Still, it is super easy to do, so why not?

If anyone can find out the duty cycle used in the study, that would be great.
posted by amtho at 7:50 PM on February 6 [1 favorite]

I am afraid that I have literally no idea what any of the comments above mean! I think the instructions I would need would be very, very basic! Like, I don't know what a breadboard is. I don't think I know what hz are. Etc.

amtho, if you're up for trying this, I would happily pay you (a modest fee plus reimbursement for parts)! I don't have access to read the original article in Nature, but it sounds like they made a strip of LEDs that flicker at this frequency. I don't know if it's a certain color of LED - that wasn't mentioned. I guess I'll take the strip and sit in a dark room and turn it on for an hour a day?

Someone else on the internet posted this thread with some ideas?

Please note to all: I totally, completely understand this is highly unlikely to work - there has only been the one mouse study, etc. But I really don't think this can harm me so hey, why not!
posted by latkes at 7:57 PM on February 6

To the webpage question: Someone set that up on Reddit, but the commentors said it won't work because the monitor isn't compatible with the flicker speed or something else I'm not understanding?
posted by latkes at 7:59 PM on February 6

If you can find more information about it -- like the duty cycle, maybe it's in the $12 version of the article -- I'll be very happy to work on this gratis, and I think I even have the parts already. Let me get it working, then we can figure out shipping, etc. I'm thinking of starting with just some plain individual white LEDs, and maybe mounting the whole thing in a jar.

It sounds like it doesn't have to be very bright, and theoretically you could increase the experienced brightness by locating the light source close to your eyes. More particulars of the experiment would be great.

The critique site I linked to mentions that they were going to publish the flicker protocol on their lab site, so that other labs could reproduce the research, but I was unable to find it there.

I'll make a ~50% duty cycle version, probably, and can modify it if that turns out to be advisable.

This is going to be mildly cool!
posted by amtho at 8:06 PM on February 6 [1 favorite]

Buy an Arduino and some LEDs.

Go to this page, and follow the instructions. (It says to use a resistor, you can ignore this.)

In the code it says "delay(1000);" change both to "delay(12);".

"delay(1000);" delays for 1 second or 1000 milliseconds. 40hz is forty times per second, so 1000/40=25, but you want the LED to be on half the time and off half the time so 25/2=12.5. We'll round 12.5 down to 12.
posted by gregr at 8:15 PM on February 6 [1 favorite]

Regarding the Maker Space, if you wanted to do that, I'd probably ask one of the people who runs it (they probably have a phone number or email address online). They might know of a regular who would be up for this kind of thing. Yeah, it could be awkward, but I don't think you have much to lose :)
posted by primethyme at 8:36 PM on February 6

latkes: " I think the instructions I would need would be very, very basic! Like, I don't know what a breadboard is."

amtho sounds like he has you covered but for anyone following along at home:

A breadboard is the white thing with the rows of holes you see on my first link. Rows of holes are electrically connected to each other (the each numbered five group is connected and the long lengths on each edge are connected, the red and black lines show what is connected to what). This allows you to literally just push the leads of your discrete components into the holes and have the components link up. This allows for rapid prototyping and easy error correction. You can buy printed circuit boards that match the breadboard connection pattern to transfer the components to 1:1 for soldering if one wishes.
posted by Mitheral at 8:48 PM on February 6

Got a yellow LED flashing at 40 hz, so that was easy. It seems that some LEDs can handle flashing that quickly, and some just seem to emit a continuous light, so LED selection matters.

I have some white ones I can try out tomorrow.

I think it might also be nice to add a second timer so that you can press a button once and the light flickers for an hour or so.
posted by amtho at 9:13 PM on February 6 [2 favorites]

Agree with Arduino and LEDs, and amtho's well on the way, just wanted to add that Adafruit's Circuit Playground might be a good way to go: $20 Arduino with rgb LED's built-in.
posted by cyclicker at 9:33 PM on February 6

Someone sent me the article! Thanks. It has the duty cycle info I wanted, but now I want to know what size room the light was used in (to give an idea of the brightness experienced by the subjects).

Also, I'm not going to stress about getting _exactly_ 40 hz, since the other frequencies they tested were half and double that. I figure 35 - 45 hz is close enough. The device might slightly change frequency over time, also.

FWIW, the non-arduino version components might end up being around $1 or $2. Figuring out which LEDs work might be non-trivial, but I'll see how it goes.

I have no reason to believe this will actually help anyone, and what you do with a flashing light is up to you. I just happen to be right at the point in my electronics learning curve where I can do this, and it's interesting to me.
posted by amtho at 9:43 PM on February 6 [2 favorites]

Ha ha I sent it to you too. I read in a couple places the mice were "inside a box" and exposed to a "strip of lights". I'll have to sit in the dark I think.

I know that no one in this thread is offering medical advice or treatment.
posted by latkes at 10:06 PM on February 6

What makes me think there might be something to this is that apparently, Coherent 40-Hz oscillation characterizes dream state in humans, and the established correlation between Alzheimer's and poor quality and insufficient sleep.

Given that 40Hz connection with REM sleep, I think I would want a 39 - 41 Hz range for my photic stimulation rather than 35 - 45 Hz.
posted by jamjam at 10:24 PM on February 6 [1 favorite]

latkes, I hope this project works for you! Are you in San Francisco or the East Bay? I've been to the TechShop maker space on Howard Street in SF. On the second or third floor, to the left of the kitchen, is a hall with doors to bathrooms and a large bulletin board. On the bulletin board are a bunch of notices, including notices from people like you who want handy folks (which is to say, people not like me) who know about LEDs and arduinos and such. You normally need to pay to use the space, but I bet the staff would be willing add a notice to the board for you. Merely FYI should amtho be unavailable for future projects.
posted by Bella Donna at 11:15 PM on February 6 [1 favorite]

More on gamma oscillations of 40Hz and over (though not on 40Hz specifically):

Why 40Hz?

I have chosen „40Hz“ as the name of this website since it happens to be the nickname for one of the experimental phenomena my group is interested in, namely, fast neuronal oscillations in the so-called gamma-frequency range. Although gamma oscillations can be observed across a broad range of frequencies between 30 and 150Hz, many early studies of gamma activity described them as occurring prominently around 40-50Hz – hence the label „40Hz“.

Gamma oscillations have first been observed by microelectrode recordings in animal studies. These studies have provided evidence for synchronized gamma activity in many neural circuits and across a wide range of species. Gamma oscillations have been observed in all sensory systems, in the motor system and in memory and association structures. The species include primates, carnivores, lagomorphs, rodents, birds, reptiles, amphibia and insects.

What the available studies demonstrate is the occurrence of gamma activity under a wide variety of tasks and paradigms including processing of coherent stimuli, perceptual discrimination, focussed attention, short-term memory, sensorimotor integration, and language processing. Typically, the observed amount of gamma is positively correlated with increased „processing load“ and, thus, with the level of attention, as well as with the difficulty or integrative nature of the processing. Generally, the human data are in good agreement with the animal studies suggesting a role of gamma synchronization in the binding and selection of distributed information. More recently, it has been suggested that gamma-band oscillations may be particularly relevant for communicating novel information in a „bottom-up“ manner to other brain regions.

posted by tenderly at 11:32 PM on February 6 [1 favorite]

So what is the duty cycle???
posted by keeo at 5:16 PM on February 7

Seems to be 50%.
posted by amtho at 10:38 AM on February 8

FWIW, Nature does have video of the mouse being exposed to the light. As they point out in the caption, video does not capture flicker rate accurately. I must say it looks incredibly annoying. But hey, science!
posted by latkes at 6:38 PM on February 8

« Older How will seeing a naturopathic doctor affect my...   |   Am I ruining my headphones? Newer »

You are not logged in, either login or create an account to post comments