A browser version of the 40 Hz sound and light stimulation from the MIT GENUS studies. The timing is exactly as published; the only thing I've moved is the sound carrier, down from 10 kHz to a pitch you can actually hear. Each pulse is a whole number of carrier cycles so the click stays clean. You'll want a 120 Hz screen for the light part.
| Parameter | This page | Martorell 2019 (mouse) | Chan 2022 (human device) |
|---|
The studies. Martorell et al. (Cell, 2019) played mice a 1 ms burst of 10 kHz tone every 25 ms, i.e. 40 Hz at a 4% duty cycle, around 60 dB, for an hour a day over a week. Chan et al. (PLOS ONE, 2022) built a home device for people that played a 1 ms rectangular click every 25 ms at 68 to 78 dB, again one hour a day. In both cases 40 Hz was the thing that worked; 8 Hz, 80 Hz and randomised timing didn't do it.
What I changed. Just the pitch of the carrier, and only in tone-burst mode. At 10 kHz a 1 ms gate holds exactly ten cycles. Drop to 600 Hz and a single cycle is already 1.67 ms long, so the gate gets rounded to the nearest whole cycle (minimum one). Every pulse starts on the 25 ms grid at zero phase and ends on a zero crossing, which means no DC thump and every pulse comes out identical. The 40 Hz rate itself never changes. Click mode skips the carrier entirely and plays the 1 ms rectangular pulse from the human device instead.
Does the pitch matter? Honestly, nobody knows for sure. The mouse study never tried a different carrier, and the human trial didn't use 10 kHz at all, it used a broadband click. What we do know from EEG and MEG work is that the 40 Hz auditory steady-state response in people gets stronger as the carrier drops (Ross, Draganova, Picton and Pantev, 2004, found roughly three times the amplitude at 250 Hz versus 4 kHz), and most older adults can barely hear 10 kHz anyway. What both papers actually pin the effect on is a sharp-onset pulse repeating at 40 Hz, and that part is untouched here whatever pitch you pick.
Safety. The studies ran at 60 to 78 dB SPL, which is about normal conversation level. Set your speakers or headphones so it's clearly audible but comfortable, and bear in mind that an hour of loud headphones can damage your hearing. Flashing light can trigger seizures in people with photosensitive epilepsy. Don't use the light if you've ever had a seizure, and stop if you start to feel off. None of this is medical advice and this page isn't a medical device.
Light. Martorell 2019 flickered a light at 40 Hz with 12.5 ms on and 12.5 ms off, and in the combined condition each flash lined up with a tone onset. Chan 2022 used a white LED panel giving roughly 390 to 400 lux at the eyes. Here the whole screen flashes on the same 25 ms grid as the sound. Red is the default; white, which is what the studies used, is in the menu. 25 ms is exactly 3 frames at 120 Hz, so the flicker is clean on a 120 Hz display and ragged on a 60 Hz one. The page measures your refresh rate and tells you which you've got. A laptop screen at arm's length gives you a lot less light than the study's panel, so sit close and turn the brightness all the way up.
Sources: Martorell et al. 2019, Chan et al. 2022