Glycine for sleep: the thermoregulatory mechanism and what the polysomnography data actually shows.

What glycine is and why it matters for sleep

Glycine (2-aminoacetic acid) is the simplest amino acid. The body synthesizes it from serine and from threonine, but dietary intake from collagen-rich foods contributes meaningfully. It is classified as non-essential in healthy adults, though researchers describe it as "conditionally essential" under high metabolic demand because endogenous synthesis may fall short of requirements [1].

In the nervous system, glycine acts as an inhibitory neurotransmitter at glycinergic synapses in the brainstem and spinal cord — distinct from its role as a co-agonist at N-methyl-D-aspartate (NMDA) receptors, where it facilitates excitatory signalling. The inhibitory and excitatory roles are anatomically compartmentalised, which is part of why glycine has a relatively benign safety profile even at gram-level doses: it does not globally suppress central nervous system (CNS) activity the way GABA (gamma-aminobutyric acid) agonists do.

The sleep application rests on a more peripheral pathway: glycine's ability to lower core body temperature (CBT) in the first hours after ingestion. That CBT drop — not sedation, not GABA receptor agonism — is the primary hypothesised mechanism behind the improvements in sleep latency and slow-wave sleep (SWS) duration seen in controlled trials.

The thermoregulatory mechanism

Healthy sleep onset is tightly linked to a fall in core body temperature. Under normal circadian conditions, CBT begins declining in the early evening, reaches its nadir in the middle of the night, and begins rising again a few hours before waking. The CBT decline is an active process driven by peripheral vasodilation that shunts heat from the body's core to the extremities. Czeisler and Klerman documented this in detail: sleep onset probability rises sharply as the rate of CBT decline increases [2]. Warm hands and feet — a proxy for peripheral heat dissipation — are one of the better physiological predictors of sleep onset.

Glycine appears to accelerate this process through vasodilatory action in the skin microvasculature. Inagawa and colleagues demonstrated that 3 g of glycine taken 30 minutes before bed produced a measurable rise in peripheral skin temperature within 90 minutes — consistent with increased cutaneous blood flow — alongside a corresponding drop in rectal temperature [3]. The magnitude was modest: approximately 0.3–0.5°C at the core. That range is physiologically meaningful given that natural CBT decline from evening peak to nadir is roughly 0.8–1.0°C in total.

The proposed molecular pathway involves glycine acting on NMDA receptors and on NMDA-independent pathways in the hypothalamic preoptic area — a region central to thermoregulatory control. Whether the effect is primarily central or peripheral remains an open mechanistic question. What the human trials can directly assess is the downstream outcome: does the temperature shift translate to measurable changes in sleep architecture?

Polysomnography evidence in humans

Polysomnography (PSG) is the gold-standard sleep assessment. It records electroencephalography (EEG), electromyography (EMG), electrooculography (EOG), and respiratory parameters simultaneously, enabling precise staging: wake, N1, N2, N3 (slow-wave sleep), and rapid eye movement (REM) sleep.

Yamadera and colleagues published the first PSG-controlled glycine sleep trial in 2007 in a crossover design enrolling volunteers reporting unsatisfactory sleep [4]. Participants received either 3 g of glycine or placebo dissolved in water, 30 minutes before bed, across four consecutive nights per arm. PSG findings:

• Sleep latency (time to sleep onset) was reduced by approximately 4 minutes versus placebo — statistically significant in this population with sub-clinical sleep difficulty.
• N3 (slow-wave sleep) duration increased. Participants entered the first SWS episode faster and spent more total time in N3 across the night.
• REM sleep duration was not significantly changed, but latency to the first REM episode was modestly reduced — suggesting the first sleep cycle was compressed toward deeper stages.
• Total sleep time was not significantly different, meaning glycine shifted architecture rather than extending duration.

Effect sizes are real but modest. A 4-minute reduction in sleep latency is less meaningful for someone lying awake 45 minutes; more meaningful for someone at 15 minutes wanting to reach 10. The SWS increase is the more interesting finding from a recovery standpoint: N3 is the stage most associated with physical restoration, growth hormone (GH) secretion, and memory consolidation.

Bannai and Kawai followed with a parallel-design study examining glycine 3 g in subjects with acutely restricted sleep — 25% curtailment by advancing wake time — measuring next-day performance [5]. PSG confirmed the SWS-augmenting pattern: glycine subjects spent proportionally more time in N3 even within the compressed window. If you cannot extend sleep duration, shifting a larger fraction toward N3 may partially offset the performance cost of restriction. This is a conditional benefit, not a replacement for adequate duration.

Daytime performance and subjective fatigue

Bannai and Kawai's sleep-restriction protocol assessed the morning after, using psychomotor vigilance testing (PVT) and subjective fatigue ratings. PVT performance — reaction time and lapse frequency — did not differ significantly between glycine and placebo in the restricted-sleep condition. Subjective fatigue and sleepiness ratings, however, were lower in the glycine arm throughout the morning.

A separate study examined daytime fatigue in individuals self-reporting poor sleep quality, using validated fatigue scales over five days at 3 g/day [6]. Participants reported reduced fatigue, reduced daytime sleepiness, and improved "liveliness" scores. These are subjective endpoints without blinded objective performance confirmation — they carry lower evidentiary weight than PSG-validated outcomes.

The most plausible interpretation: glycine's daytime benefit is downstream of its night-time sleep architecture effect. Better N3 sleep — even modestly better — translates to somewhat improved morning alertness, particularly in people starting from chronic mild sleep disruption. The effect is unlikely to be dramatic in individuals who already sleep well.

The 2023 GeroScience systematic review

Razak and colleagues published a systematic review in GeroScience in 2023 synthesising evidence on glycine supplementation across sleep, metabolic function, inflammation, and cognitive performance, with explicit attention to ageing populations [7]. Sleep outcomes received a dedicated assessment.

Four controlled studies met inclusion criteria for sleep endpoints. Pooled analysis was not possible due to heterogeneity, so the authors used narrative synthesis. Their conclusions:

• Sleep latency: Consistent direction of effect across trials — glycine 3 g shortened time to sleep onset. Effect size was moderate and most pronounced in participants with baseline sleep difficulty.
• Slow-wave sleep: Consistent N3 augmentation. This was the most reproducible finding across studies.
• Subjective sleep quality: Improvement on validated scales (Pittsburgh Sleep Quality Index and proprietary variants) in most trials. Subjective improvement exceeded objective PSG changes in magnitude, which the authors noted may reflect expectation effects.
• Safety: No serious adverse events attributable to glycine at 3 g across any included study. Mild gastrointestinal discomfort was infrequent.

The review's overall evidence grade for glycine sleep effects: "promising but preliminary" — recognising the consistent directional signal while flagging sample size, funder affiliation, and replication gaps. The authors called for independent, adequately powered RCTs in Western populations with objective PSG endpoints. As of this writing those trials have not been published.

For the ageing angle: the review noted that CBT dysregulation is more pronounced in older adults, and thermoregulatory interventions may therefore have larger effect sizes in populations over 60. Biologically plausible but not yet directly tested with glycine in a well-powered older-adult cohort.

Dosing, form, and timing

Every controlled trial reporting positive PSG outcomes used 3 g of glycine taken 30–60 minutes before intended sleep onset. No published dose-ranging trial exists for sleep endpoints in humans. Haghayegh and colleagues' meta-analysis on passive body heating and sleep onset provides context: interventions that lower CBT by 0.3–0.5°C produce sleep-onset benefits; the Inagawa data suggest 3 g glycine approaches but does not clearly exceed this threshold [8]. Higher doses (5–10 g) have been used in metabolic studies without significant adverse effects, but no evidence supports escalating beyond 3 g for incrementally better sleep outcomes.

Form: Glycine is available as a free amino acid powder (the form used in all trials), as capsules, and as a component of compound products such as magnesium glycinate — though in the latter case the glycine content per serving is typically 200–500 mg, not the 3 g studied. If you are targeting the sleep-specific mechanism, a dedicated glycine product at 3 g is the directly-supported approach.

Timing: 30 minutes before bed is the established window. The thermoregulatory response is detectable within 45–90 minutes of ingestion. Taking it immediately before lying down likely still works, but the 30-minute lead allows the process to begin as you transition to sleep.

Combinations: Walker's framework identifies consistent wake time, darkness, and temperature as the primary sleep levers [9]. Glycine is an adjunct to those foundations, not a substitute. Within supplement stacks, glycine combines cleanly with magnesium glycinate and low-dose melatonin. There is no known interaction with cognitive behavioural therapy for insomnia (CBT-I) protocols.

The American Academy of Sleep Medicine (AASM) 2023 clinical practice guidelines on chronic insomnia do not include glycine — CBT-I remains first-line, with pharmacological agents as adjuncts for specific populations [10]. Glycine's absence from AASM guidance reflects the small evidence base rather than any contradictory evidence.