gamma entrainment · light and sound · nothing here puts current into a body
Photosensitive epilepsy. Flicker in this range can trigger seizures in susceptible people. If you
have epilepsy or a family history of seizures, use the auditory method only — it does not flicker.
The visual methods stay locked until you confirm.
Session
No session selected — pick one from the library below.
Confirm before any visual method. I do not have epilepsy or a family history of seizures, and
I understand the visual channel may be approximate on this display.
measuring display…
idle
Audio is sample-accurate — it runs on the audio clock, so the beat frequency is
exact at any Hz. Video is bounded by your refresh rate: a 60Hz display cannot place 40Hz pulses on equal
frame boundaries, a 120Hz display can. Diffuse room light, never a strobe aimed at the eyes. Comfortable
volume. The clinical protocol for 40Hz is one hour per day, daily.
The library
Every session carries its evidence grade in the data, not in the marketing.
strong = trials in progress ·
moderate = replicated, entrainment verified ·
promising = small studies ·
weak = inconsistent, listed honestly ·
traditional = no clinical evidence at all.
Gamma / Cognition
The 40Hz research line. The strongest evidence in the whole field.
GENUS 40strong
60 min · 40Hz peak · combined
The clinical protocol: 40Hz, one hour, daily. Phase 2A over 3 months in mild probable Alzheimer's showed lesser ventricular dilation and hippocampal atrophy, increased default-mode connectivity and better face-name delayed recall vs control. 670-participant pivotal trial (HOPE, NCT05637801) has completed enrolment.
The full protocol is an hour. Shorter runs have not been tested for the same outcomes.
GENUS 40 · shortpromising
15 min · 40Hz peak · auditory
40Hz auditory stimuli produced the highest EEG response and increased regional cerebral blood flow in healthy participants. Fifteen minutes is a practical dose, not the trial dose.
GENUS 40 · vibrotactilepromising
30 min · 40Hz peak · auditory
A FIFTH delivery route. Kim et al. (Scientific Reports, July 2026) compared 40Hz vibrotactile stimulation delivered by a GLOVE against combined visual+auditory in 15 healthy participants, and found the glove evoked 40Hz EEG responses in central, frontal and — to a lesser extent — occipital cortex. That matters because it reaches the gamma response with NO light at all, which is the whole problem for photosensitive people.
This session plays the 40Hz audio reference. The browser cannot deliver true 40Hz vibrotactile — the Vibration API is a coarse on/off pattern, not a waveform, and phones vary wildly. The real route is built hardware; see the build guide. We are not going to pretend a phone buzz is the study.
Pain
Alpha entrainment. Dose-response has been shown.
Alpha 10 · painmoderate
20 min · 10Hz peak · combined
Four minutes of 10Hz sensory stimulation entrains alpha and decreases pain, and the degree of frontal alpha power increase CORRELATES with the pain reduction — a dose-response relationship. A 2025 randomised crossover trial (Journal of Pain) ran two weeks of active vs sham pre-sleep stimulation at home in fibromyalgia: alpha power was enhanced under active vs sham, with self-reported improvement in pain and sleep.
Sleep
Descending ramps that mirror the natural descent into sleep.
Descent 8→2promising
35 min · 8Hz peak · auditory
Audio-visual stimulation descending from 8Hz to 2Hz achieved entrainment and improved both pain and insomnia symptoms in chronic-pain participants. The ramp mirrors the natural alpha→theta→delta descent rather than holding one frequency.
Audio-only by default so you can run it with your eyes shut and let it end without you.
Pre-sleep Alphamoderate
20 min · 10Hz peak · auditory
The pre-sleep arm of the 2025 fibromyalgia crossover trial — 10Hz delivered at home before sleep, verified by enhanced alpha spectral power against sham.
Pink noise · consolidationpromising
45 min · 1Hz peak · auditory
The best-supported idea in sleep audio is not a tone at a frequency — it is sound delivered in PHASE with your own slow oscillations during deep sleep. A 2026 paper in npj Science of Learning covers slow-oscillation and spindle stimulation effects on physiology and memory; a 2026 Neuroscientist review ("Echoes of Pink Noise") proposes the mechanism for enhancing sleep-dependent memory consolidation.
HONEST LIMIT: real closed-loop stimulation needs live EEG to find the up-phase of your slow waves. Open-loop pink noise — what a browser can play — is the weaker cousin, and we grade it as what it is.
Calm / Anxiety
Theta and delta. The largest anxiety effect in the binaural literature.
Theta 6 · calmmoderate
20 min · 6Hz peak · auditory
Theta/delta binaural beats carry the largest anxiety effect in the meta-analytic literature: Hedges' g = 0.69 (medium-to-large) across five effect sizes in four studies, total N=159. Small N — directional. Overall binaural-beat effect across cognition, anxiety and pain: g = 0.45.
Before a Proceduremoderate
30 min · 6Hz peak · binaural
A 2025 systematic review and meta-analysis of 15 RCTs (>1,000 patients) found perioperative binaural-beat audio significantly reduced anxiety, postoperative pain, systolic blood pressure and heart rate — and outperformed non-binaural "placebo" music head-to-head. A separate 2025 review of 9 dental trials found significant reduction in dental anxiety.
Binaural — needs headphones. This is the one clinical use where binaural specifically was tested.
Focus
Beta and SMR. Weakest evidence, highest photic risk — read the warnings.
SMR 14weak
25 min · 14Hz peak · auditory
Cognitive effects are the least consistent finding in the entrainment literature. Binaural beats are not a reliable tool for concentration; they appear to help some people under some conditions. Included because it is asked for, graded honestly.
AUDIO ONLY by default. 14Hz sits inside the 15-25Hz photic danger band once you add light.
Meditation
Theta states. Long tradition, thin clinical evidence.
Theta 4 · deeppromising
30 min · 8Hz peak · auditory
Theta (4-8Hz) is the band of the wake-sleep boundary — daydreaming and meditative states. Entrainment into it is demonstrable; that it produces the *contents* of meditation is not.
Schumann 7.83traditional
30 min · 7.83Hz peak · auditory
The 7.83Hz Schumann resonance is REAL and well documented geophysics — the fundamental of the Earth-ionosphere cavity (Schumann 1952; Rycroft et al. 2000; Harrison 2013 on the Carnegie curve), and it does fall at the theta-alpha boundary of human brainwave frequencies. What has NO clinical evidence is the separate claim that playing a 7.83Hz tone through headphones confers a health benefit. Those are two different propositions and this grade concerns only the second. Any effect from this session is most likely a plain alpha-band effect.
Graded on the audio-playback claim, not on the geophysics. See the Conserved Library paper (knowledge/consciousness/conserved_library_genome_to_cosmos.json, §6) for the continuous-circuit framework from ion channel to ionosphere.
The Chamber
Full-field, enclosed, ~12 minutes. The context IS the intervention.
Chamber · Alphamoderate
12 min · 10Hz peak · combined
The protocol from Cone et al., npj Digital Medicine, 28 March 2026 (n=74): synchronised rhythmic light and audio at 9-11Hz, 11.5 minutes, inside an immersive reflective chamber. Produced substantial acute improvement in anxiety, mood disturbance, flow states and vitality, with STAI state-anxiety reduction reaching magnitudes comparable to established pharmacological and psychotherapeutic interventions requiring far longer treatment.
The study found alpha and theta arms performed EQUIVALENTLY — the immersive context, not the frequency, is a primary active ingredient. Run this fullscreen, in a dark room, with headphones.
Chamber · Thetamoderate
12 min · 6Hz peak · combined
The theta arm (4-7Hz) of the same 2026 chamber study. It performed equivalently to the alpha arm — which is the finding, not a footnote. Offered so you can test the equivalence yourself rather than take it on trust.
Same protocol, different band. If you notice a difference between this and Chamber · Alpha, that is worth recording — the published result says you should not.
Perceptual Training
Strobe used the OTHER way: intermittent occlusion to train the visual system, not to entrain it.
Stroboscopic trainingpromising
15 min · 6Hz peak · flicker
Completely different use of a strobe: not entrainment, but INTERMITTENT OCCLUSION — removing visual information so the system learns to work on less. A 2026 study in Frontiers in Physiology found stroboscopic vision training improved sports-vision and punching performance; a 2024 study in Life tested elite curling athletes; and a 2025 paper in Applied Neuropsychology found repeated training improved ANTICIPATION skill without changing general perceptual-cognitive skills — a useful, specific negative result that keeps the claim honest.
Done with the EYES OPEN and moving, ideally while catching or tracking something. Sitting still and staring at it is not the intervention.
Visionary
Alpha-band flicker with the eyes CLOSED — the Dreamachine effect.
Dreamachinepromising
15 min · 10Hz peak · flicker
Alpha-band flicker viewed through CLOSED eyelids reliably produces geometric form constants — lattices, spirals, tunnels — the effect Brion Gysin and Ian Sommerville built the Dreamachine around in 1959 at 8-13Hz. The phenomenology is well documented and reproducible; the mechanism is understood as driven activity in visual cortex, not as anything mystical.
EYES CLOSED. The light goes through the eyelid — you do not look at it. Still a flicker source, so the photosensitivity warning applies in full.
What the evidence says
The MIT line is GENUS — Gamma ENtrainment Using Sensory stimuli. The hypothesis is that driving
40Hz does downstream biological work rather than just producing a matching EEG trace: increased phagocytic
activity of microglia, augmented vasodilation and transcytosis across the brain endothelium, and glymphatic
clearance — arterial vasomotor pulsation entrained by the envelope of local gamma oscillations, moving
cerebrospinal fluid and clearing amyloid in mouse models.
Whether exactly 40 matters is genuinely unsettled. Recent work drives the 36–44Hz
range to test it, on the reasoning that comparing against frequencies far outside that vicinity reveals
nothing about specificity.
1 · Luminance flicker
The original and most-studied route — and the least comfortable. In the OVERTURE trial, 40Hz flicker
combined with 40Hz-modulated 10kHz clicks produced headache in 20.7% and tinnitus in 15.2%
of the active group (n=46), against 10.7% and 0% in sham (n=28). In a healthy cohort it rated 6.2 points
higher for discomfort on a 0–10 scale than steady light.
2 · Invisible spectral flicker (ISF)
Modulates the light's spectral composition while holding perceived luminance constant — the eye sees a
steady field, the visual cortex still gets a 40Hz drive. All three visual variants produce a 40Hz
steady-state evoked potential, but ISF scores significantly better on comfort and perceived flicker.
Reducing brightness had no significant effect on the response, and peripheral viewing angles cost only
slightly — which may free a patient from gazing directly at the light. Now in randomised, triple-masked,
placebo-controlled trial.
Here it is approximated by alternating two colours matched for relative luminance
(sRGB-linearised, the WCAG formula) but differing in spectral content. The luminance match is computed in
the browser, not eyeballed.
3 · Auditory
40Hz clicks, or a carrier amplitude-modulated at 40Hz. Works with the eyes closed and sidesteps
photosensitivity entirely. Produced the highest EEG response and increased regional cerebral blood flow in
healthy participants; in dementia patients it enhanced default-mode-network connectivity, strengthening
frontal–parietal rhythmic synchrony, and those improvements correlated with memory performance.
Because visual gamma drive is more likely to provoke seizure or pre-seizure activity in photosensitive
people, auditory stimulation has been proposed as the safer route for that population — which is why it
is the one method unlocked by default here.
4 · Combined audiovisual
What the pivotal clinical work runs. The protocol is consistent across trials: one hour per day, at
home, daily. A Phase 2A pilot over 3 months in mild probable Alzheimer's showed lesser ventricular
dilation and hippocampal atrophy, increased default-mode-network connectivity, better face-name delayed
recall, and improved daily activity rhythmicity versus control. Adherence over the 6-month study averaged
85% among completers.
Status: Cognito Therapeutics completed enrolment of 670 participants in the HOPE study
(NCT05637801). Re-read all of the above against that readout. And the 2016 reason for interest in 40Hz
— lucid dreaming — has largely fallen apart; the therapeutic gamma literature and the lucid-dream
literature are separate stories and should not prop each other up.
Practices — the half that needs no hardware
Everything above needs a screen, headphones or a device you built. These need nothing at
all. They are graded on the same scale, for the same reason: this corner is thick with confident
instruction and thin with evidence, and the useful thing is saying which is which.
None of this is a treatment for insomnia. Sleep that stays broken for
weeks is a clinical matter, not a technique problem.
Getting to sleep
What to do with a mind that will not stop. The folk answer is wrong and the literature says what to do instead.
Imagery distraction (what to do instead of counting sheep)moderate10 min
Build one specific, interesting, absorbing scene and stay inside it — not a count, not a list. The scene has to be engaging enough that re-engaging with your worries would take effort.
Before bed, pick ONE scene you find genuinely interesting. Specific, not generic: a particular shop you know, a route you have walked, a workshop you would like to build.
Lying down, enter the scene and furnish it. What is underfoot, what is the light, what can you hear, what is on the shelf to your left.
Keep elaborating. New detail, not repetition. The task is to occupy the space a worry would otherwise fill.
When you notice you have drifted back to a worry — you will — return to the scene rather than fighting the worry.
Evidence. Harvey & Payne (2002, Behaviour Research and Therapy) gave 41 people with insomnia one of three instructions: distract with imagery, distract generally, or nothing at all. Imagery distraction produced shorter sleep-onset latency and less frequent, less distressing pre-sleep thought than no instruction. GENERAL distraction — the counting-sheep shape — did not.
Caution. Not a treatment for chronic insomnia. If sleep stays broken for weeks, that is a clinical matter, not a technique problem.
The most famous sleep technique in the world, and the one study that actually tested its shape found it did nothing. Listed because people ask, and because the correction is more useful than the omission.
The traditional instruction: count imagined sheep passing, one by one, until sleep comes.
What the evidence says: a repetitive, unengaging count is "general distraction", and general distraction did not shorten sleep onset.
Use imagery distraction instead — same idea, but a specific absorbing scene rather than a monotonous count.
Evidence. In Harvey & Payne (2002) the general-distraction arm was predicted to be WORSE than no instruction at all; that prediction was not supported either. So the honest summary is that it neither helped nor measurably hurt — it simply did not do the thing it is famous for. Imagery, in the same study, did.
The induction family. Two techniques have real comparative data behind them; the rest are tradition.
MILD — Mnemonic Induction of Lucid Dreamsmoderate10 min
Prospective memory. You rehearse the intention to notice you are dreaming, attached to a specific remembered dream, so the intention is waiting for you when the dream starts.
Wake from a dream (an alarm ~5 hours in, or a natural waking) and recall it in as much detail as you can.
Pick the moment in that dream that was most obviously impossible — the dreamsign.
Repeat, meaning it rather than reciting it: "Next time I am dreaming, I will remember that I am dreaming."
While repeating, SEE yourself back in that dream, reaching the dreamsign, and recognising it. Vividly, not abstractly.
Let yourself fall asleep while still holding the intention.
Evidence. The International Lucid Dream Induction Study (Adventure-Heart 2020, Frontiers in Psychology) ran 355 participants across five technique combinations. MILD and SSILD came out SIMILARLY EFFECTIVE, and a hybrid of the two showed no advantage over either alone. Success predicted by good general dream recall and by falling asleep within ten minutes of finishing the technique. No adverse effect on sleep quality was found.
The ten-minute figure is the practical one. If the technique leaves you wide awake, it is working against itself.
Cycle attention through sight, hearing and touch in slow repeated passes, then simply go to sleep. No visualisation and no affirmation — which makes it the technique of choice if MILD keeps you awake.
Wake after ~4-5 hours of sleep. Lie comfortably, eyes closed.
Three or four QUICK cycles: attend to what you see behind closed eyelids, then to what you hear, then to what you feel in your body. A few seconds each.
Then three or four SLOW cycles: same three senses, roughly thirty seconds each. Observe; do not strain to produce anything.
Stop, and go to sleep normally without holding any intention. That is the whole technique.
Evidence. In the same 355-person International Lucid Dream Induction Study, SSILD performed comparably to MILD. That is the useful finding: two mechanistically different techniques, similar results, so choose the one that lets you fall back asleep quickly.
Not an induction technique on its own — a multiplier. Waking late in the night and returning to sleep puts you into REM-dense sleep with an alert enough mind to hold an intention.
Set an alarm for roughly 4.5 to 6 hours after you fall asleep — late enough that REM periods are long.
Get up. Stay awake 10 to 30 minutes: read about dreaming, write the dream you just had, do the technique.
Go back to bed and run MILD or SSILD as you fall asleep.
Evidence. Sleep-laboratory work supports the wake-then-return structure (Erlacher and colleagues, 2020), and a 2022 study examined how the TIMING of the interruption changes the result — earlier interruptions are not automatically better, so the late-night window matters.
Caution. This deliberately fragments your night. Do not run it before a day that needs you sharp, and not at all if you are already sleep-deprived.
DILD is not a technique — it is the CATEGORY of becoming lucid from inside a dream. Reality testing is the daytime habit meant to produce it: checking, repeatedly, whether you are awake.
Several times a day, genuinely ask whether you are dreaming — genuinely, not as a formality.
Test it physically. Read text, look away, read it again (it changes in dreams). Try to push a finger through your palm. Pinch your nose shut and try to breathe in.
Pair the check with things that recur in your dreams, so the habit fires where it is needed.
Evidence. Reality testing is the most widely taught technique and among the least well supported on its own. In the International Lucid Dream Induction Study it was a component of the combinations tested rather than the active ingredient, and reviews of induction techniques repeatedly find single daytime methods underperform the night-time memory techniques. Graded weak alone, useful as a component.
WILD — Wake Initiated Lucid Dreamtraditional20 min
Cross from waking directly into a dream without losing awareness. The most dramatic technique in the family and the one with the least controlled evidence behind it.
Best attempted after WBTB, when sleep pressure will carry you across quickly.
Lie still and let the body fall asleep while attention stays on one quiet anchor — the breath, or the shapes behind the eyelids.
Expect the transition: hypnagogic imagery, sounds, a sense of weight or vibration, sometimes sleep paralysis.
Do not grab at the imagery. Let it thicken until it is a place, then step in.
Evidence. Rich first-person and instructional literature; very little controlled data specific to WILD as a discrete technique. Graded traditional for that reason, not because practitioners are wrong.
Caution. Sleep paralysis is a normal and harmless part of this transition and can be genuinely frightening the first time. It ends on its own. Knowing that in advance is most of the remedy.
After a brief waking, make tiny alternating finger movements — as if playing two piano keys — while the body falls back asleep, then reality-test. Popular, fast, and completely untested.
Wake after 4-6 hours. Move as little as possible and keep your eyes shut.
Rest two fingers on the mattress and make micro-movements, alternating, as if pressing two piano keys very lightly. Barely move at all.
Continue for roughly 20-30 seconds while letting sleep take the rest of you.
Reality-test — the nose-pinch breath test is the usual one here, because it works even if you cannot see clearly.
Evidence. No controlled study. FILD is a community technique with a large body of anecdote and zero published trials. Graded traditional on that basis, which is a statement about the evidence and not about whether practitioners experience something.
Listed because it is one of the most-asked-about techniques. If it works for you, that is a report worth filing.
DEILD — Dream Exit Initiated Lucid Dream (chaining)traditional5 min
Re-enter the dream you just left. When you wake from a dream, stay completely still, keep your eyes closed, and slide straight back in while the dream is still warm.
The instant you wake from a dream, do not move and do not open your eyes. Movement is what ends it.
Do not think about the day. Hold the dream you just left.
Let yourself sink back, expecting to arrive in the same scene.
Reality-test as soon as anything forms.
Evidence. No controlled trials. The underlying observation — that dream re-entry is easiest immediately after a REM awakening — is consistent with sleep-laboratory work on the wake-then-return structure, but DEILD itself has not been tested as a discrete technique.
ADA and CAT — all-day awareness, cycle adjustmenttraditional1 min
Two daytime approaches. ADA: hold continuous sensory awareness through the day so the habit carries into sleep. CAT: shift your wake time earlier for a period so the body clock puts you in light REM-rich sleep when you would normally be deeply asleep.
ADA — several times an hour, take deliberate stock of all senses at once: what you see at the edges, what you hear behind you, what your feet feel.
CAT — for one week, wake 90 minutes earlier than usual and stay up. On alternate weeks, return to your normal time and reality-test hard on the mornings you sleep in.
Both are habit-formation plays rather than night-of techniques.
Evidence. Neither has controlled data. They are included because both are widely taught and because the reasoning behind CAT — exploiting circadian REM distribution — is at least mechanistically coherent, which is more than can be said for some of the field.
Caution. CAT deliberately manipulates your sleep schedule. Do not run it alongside shift work or while sleep-deprived.
The single strongest predictor of whether any induction technique works for you.
Dream recall trainingmoderate5 min
The unglamorous one that the data says matters most. In the 355-person study, superior general dream recall was a predictor of successful induction — so this is the prerequisite, not the accessory.
Keep paper and pen within reach. Screens light you up and cost you the dream.
On waking, do not move and do not open your eyes. Hold still and let the dream come back first.
Write it immediately, even a fragment, even one image. Fragments train recall as well as full dreams.
Read back over the journal weekly and mark what recurs — those recurrences are your dreamsigns, and MILD needs them.
Evidence. Adventure-Heart (2020) identified superior general dream recall as a predictor of successful lucid induction across 355 participants. Recall is trainable, which makes it the highest-leverage place to start.
The one supplement in this library with a proper trial behind it — and the trial found something narrower than what it is sold for. B6 increased how MUCH dream content people recalled. It did not make dreams more vivid, more bizarre, or more colourful.
What was studied: 240 mg pyridoxine hydrochloride before bed, five consecutive nights.
What it did: significantly increased the amount of dream content recalled.
What it did NOT do: vividness, bizarreness and colour were all unaffected, despite being exactly what the marketing claims.
Why it belongs here anyway: recall is the strongest predictor of successful lucid induction, so more recall is a real lever even if it is not the glamorous one.
Evidence. Aspy and colleagues (2018, Perceptual and Motor Skills) ran a randomised, double-blind, placebo-controlled trial in 100 participants, replicating a 2002 pilot at larger scale. B6 increased recalled dream content only. Notably the B-COMPLEX arm did worse: significantly lower self-rated sleep quality. More B vitamins is not better here.
Caution. This matters more than the effect does. Chronic high-dose pyridoxine causes peripheral sensory neuropathy — numbness and tingling in the hands and feet, sometimes slow to reverse. 240 mg is far above dietary intake and the trial ran FIVE NIGHTS, not indefinitely. Do not take this as a standing nightly supplement, and stop at any pins and needles.
The one application with clinical standing: nightmares. Plus what the meditation evidence really shows.
Lucid dreaming for nightmarespromising
The application with actual clinical standing. If you know you are dreaming inside a nightmare, you can change what happens — and that is used as a treatment.
Usually taught alongside imagery rehearsal therapy: rewrite the nightmare while awake, rehearse the new version, then use lucidity to steer toward it.
The lucid element is not always required for the treatment to work; rescripting alone has the stronger evidence base.
This is done with a clinician when the nightmares are frequent, trauma-linked or disabling.
Evidence. A 2006 pilot in Psychotherapy and Psychosomatics tested lucid dreaming treatment for nightmares; a 2015 study in Acta Neurologica Scandinavica used it as an add-on to Gestalt therapy; and the American Academy of Sleep Medicine best-practice guide for nightmare disorder in adults places these approaches within the recognised options. A 2022 Scientific Reports paper examined mindful acceptance and lucid dreaming against nightmare frequency and distress.
Caution. Trauma-linked nightmares are a clinical matter. Do not self-treat PTSD with a dream technique — this belongs with a clinician.
Long-term meditators do report more lucid dreams than non-meditators. But when an eight-week mindfulness course was tested in a blinded randomised design, it did NOT increase lucid dream frequency. Association is not the same as an intervention that works on your timescale.
The correlation is real: frequent lucid dreaming tracks with meditation practice style, meta-awareness and trait mindfulness.
The causal test at eight weeks came back negative.
The honest reading: whatever long-term practice does, it is not something an eight-week course delivered.
Evidence. Baird and colleagues (2019) used three complementary methods — a cross-sectional comparison of long-term meditators against meditation-naive individuals, a trait-mindfulness analysis, and a BLINDED RANDOMISED CONTROLLED test of an 8-week mindfulness course. Lucid dreaming was more frequent in long-term meditators; the 8-week course did not increase it. A 2024 paper in Brain Sciences replicated the association with practice style and meta-awareness.
Graded weak as an INDUCTION TECHNIQUE, which is what this library grades. That is not a judgement on meditation, only on the claim that taking it up will make you lucid.
Using a smell or a sound in sleep to re-trigger something practised while awake.
Scent-cued reality testingpromising5 min
Practise reality testing while a distinctive smell is present, then reintroduce that smell during early-morning sleep. The cue re-activates the practised behaviour inside the dream. This is where the scent work and the dream work meet.
Choose one distinctive scent you do not otherwise encounter.
For several days, do your reality checks with that scent present, so the two are bound together.
During the early-morning sleep window, have the scent reintroduced — a timer diffuser, or someone else placing it.
Evidence. A 2020 proof-of-concept in Consciousness and Cognition induced lucid dreams by olfactory-cued reactivation of reality testing during early-morning sleep. Proof of concept means exactly that: the mechanism was demonstrated, the effect is not yet established at scale.
Targeted memory reactivation is the general form of this, and it is one of the better-supported ideas in sleep science. The lucid application is the new and unproven part.
Gamma current during REM — the 40Hz result in the dream literaturepromising
The single most striking result in lucid-dream research, and the reason this library sits on a page about 40Hz. Frontal current stimulation in the lower gamma band during REM sleep induced self-reflective awareness in dreams — and other frequencies did not.
This is a laboratory finding, not a home protocol. It is here as evidence, not as instruction.
What was done: frontal transcranial alternating current at 25Hz and 40Hz, applied during ongoing REM sleep.
What happened: self-reflective awareness appeared in dreams, and control frequencies produced nothing.
The buildable version of this hardware is on this page, but applying current to your own head while asleep is not something we are handing you as a recipe.
Evidence. Voss et al. (2014, Nature Neuroscience) established a CAUSAL link where only correlation existed before: fronto-temporal gamma EEG had been associated with dream awareness, and this showed that driving it produces the awareness. The paper states that other stimulation frequencies were not effective, "suggesting that higher order consciousness is indeed related to synchronous oscillations around 25 and 40 Hz."
Hold this next to the 2026 chamber study elsewhere on this page, which found alpha and theta performed EQUIVALENTLY and concluded the immersive context was the active ingredient. Here, frequency was decisive; there, it was not. Both results are real and they are about different things. We publish both rather than the one that flatters the product.
Caution. Do not improvise this. Stimulating your own head while asleep means no one is monitoring you and you cannot end the session. The lab did it with staff, EEG and a stop condition.
TLR — Targeted Lucidity Reactivationpromising45 min
Train reality testing against a specific sound while awake, then replay that sound quietly during REM so the trained behaviour fires inside the dream. The cued cousin of MILD, and the version that has reached a clinical pilot.
Pick a distinctive, non-startling audio cue.
Awake, practise reality testing repeatedly with the cue playing, so cue and check are bound.
Sleep — a morning nap is the studied window, because REM is dense there.
The cue is replayed quietly during REM, below waking threshold.
Evidence. The cueing logic is targeted memory reactivation, one of the better-supported ideas in sleep science. A 2025 pilot in the Journal of Sleep Research combined cognitive behavioural therapy with targeted lucidity reactivation to treat narcolepsy-related nightmares — which is TLR being taken seriously in a clinical setting rather than only a curiosity.
Needs someone or something to deliver the cue at the right time. That timing is the hard part and the reason this is not yet a home technique.
A microcontroller hardware timer gives true 40Hz; a screen usually cannot. 40Hz means a 25 ms
period, so the output toggles every 12.5 ms — a timer alarm at 12500 µs hits that exactly where a
frame loop drifts. A 60Hz display cannot place pulses on equal boundaries,
because 60 is not an integer multiple of 40. A 120Hz display can — exactly three frames per cycle. This page measures your
refresh rate and says which case you are in.
The build — about $35–40
Part
Est.
Notes
ESP32 dev board
$6–10
Wi-Fi built in, so an app can control it
12V LED strip (warm white, ~1m) or 4× 1W LEDs
$8
Diffuse light; not aimed at the eyes
Logic-level MOSFET (IRLZ44N or similar)
$2
Switches the LEDs at 40Hz
Small speaker + PAM8403 amp board
$5
40Hz click train / isochronic tone
12V power supply
$8
Shared with the LED strip
Breadboard, wires, resistors
$5
Building a TENS unit
A TENS unit is a pulse generator driving a current-limited, DC-blocked output into two skin
electrodes. Every part of that sentence is load-bearing, and the failure modes come from getting one of
them wrong — so build it understanding what each stage is protecting you from.
The five stages
Oscillator / timing. Pulse rate 1–150 Hz, pulse width 50–300 µs. Conventional TENS runs
high rate (80–120 Hz) and narrow; acupuncture-like TENS runs low rate (2–10 Hz) and wide. A
microcontroller does this far better than a 555, because you get exact timing and a hard watchdog.
Constant-current output — the stage that matters. Skin impedance swings enormously with
sweat, electrode contact and time. A constant-voltage output therefore delivers wildly varying
current: comfortable at first contact, then rising as the gel wets. A constant-current sink holds
delivered current steady as impedance changes. This is why commercial units feel stable and why bad DIY
units burn people.
DC blocking. Put a series capacitor in the output path. Any net DC across skin drives
electrochemical reactions at the electrode — that is not a heating burn, it is a chemical one, and
it happens at currents far below what feels alarming. Charge-balanced biphasic pulses plus a blocking
capacitor is the standard answer.
Isolation. Run it from a battery. Never mains-derived, never while charging. This removes the
entire class of fault where a mains transient reaches the electrodes.
Fault behaviour. Decide what happens on a broken wire, a lifted electrode, a brownout, or a
firmware hang — and make the safe state output off. A watchdog that stops output beats one that
resets into a live output.
Verify before it touches skin
This is the step that separates a build from a hazard, and it needs no special equipment beyond a
multimeter and, ideally, a scope:
Load resistor first. 500 Ω–2 kΩ across the output stands in for skin. Never test on
yourself first.
Measure current, not voltage. Across the load, confirm the peak current matches what the dial
claims, and that it stays constant when you swap 500 Ω for 2 kΩ. If current changes with
load, you have built a voltage source and the constant-current stage is not working.
Check for DC offset. Meter on DC volts across the load, output running. You want ≈0. Anything
persistent means your blocking capacitor or charge balance is wrong — fix it before proceeding.
Scope the pulse. Confirm width and that the waveform is biphasic and returns to baseline.
Sweep the dial to maximum on the load and confirm it never exceeds your design ceiling.
Electrodes: use proper self-adhesive hydrogel pads with adequate surface area. Small
electrodes concentrate current density, which is what causes hot spots. Never place electrodes across the
chest, over the carotid sinus, or on broken skin. Do not use with an implanted pacemaker or defibrillator.
Building a tDCS unit
tDCS is a much simpler circuit than TENS and a more consequential one — a low, steady direct current
between two scalp electrodes. There is no oscillator; the whole device is a regulated constant-current
source with a ramp and a cutoff.
Current range. Research protocols overwhelmingly use 1–2 mA. That is the studied
band; build your ceiling there rather than making a device that can exceed it.
Constant current, again. A regulated current sink is the entire device. Scalp and hair make
impedance unstable, so a voltage source is not acceptable here at all.
Ramp up and down. 30 seconds in and out. Abrupt onset produces a bright phosphene flash and
a much less comfortable sensation; ramping is standard practice for both comfort and blinding.
Impedance monitoring and cutoff. Measure the voltage the current source needs to maintain its
setpoint — that is your impedance readout. If it rises past a threshold (electrode drying, poor contact),
shut the output off rather than pushing harder. This is the single feature that separates a safe
build from an unsafe one, and it is entirely achievable: it is one ADC reading and one comparison.
Hard ceiling in hardware. A resistor that physically limits maximum current, independent of
firmware, so a software bug cannot exceed it.
Electrodes. Saline-soaked sponges over conductive rubber, ~25–35 cm². Keep them wet — a
drying sponge raises impedance and concentrates current, which is the usual cause of skin irritation and
small burns. Re-wet between sessions.
Sensation: a mild tingle or itch under the electrodes is expected and usually fades. A
metallic taste, a phosphene at switch-on, and mild transient headache are commonly reported. Stop if
sensation exceeds mild tingling, if it becomes a sharp or burning point, or if the skin under an
electrode is anything more than transiently pink. Never over broken skin.
Evidence, so you know what you are building for: the best-supported territory for tDCS is
depression and chronic-pain research protocols. The lucid-dream literature (Voss 2014 tACS 25/40Hz;
Stumbrys 2013 tDCS over DLPFC) timed stimulation to REM sleep in a lab, results were weak to mixed, and a
replication failed — treat that use as experimental.
This is an enclosure, not a player. Darken the room. Put headphones on. Turn your phone over.
When you cross in, you stay until it ends — about twelve minutes.
The 2026 chamber study found alpha and theta arms worked
equally well. The frequency is not what is doing the work. The room is.
Settling
0:00
Return
Sit for a moment before you stand. Notice what is different and what is not — both are data.