Circadian chronobiology and melatonin in RLS

Synthesis of the evidence that Restless Legs Syndrome (RLS) has an intrinsic circadian (clock-driven), not merely rest-dependent, pattern of symptom expression; the phase relationship between symptom severity, core body temperature, and endogenous melatonin; the effect of exogenous melatonin and bright light on RLS symptoms and PLMS; clock-gene and circadian-neurotransmitter evidence; and what this means for timing non-drug interventions (light, activity, medication) in people already on dopamine agonists or alpha-2-delta ligands. All claims below cite a study note filename (in knowledge-base/studies/) or a PMID. Medical disclaimer: nothing here is individual medical advice; any change to RLS medication, especially tapering a dopamine agonist, must be done only under physician supervision and never stopped abruptly (dopamine agonist withdrawal syndrome, DAWS, can be severe).

What is established (grade A/B)

• RLS's "worse in the evening or at night" criterion and its "worse at rest" criterion are related but statistically separable phenomena, not the same thing measured twice: controlled studies holding body position and sleep/wake state constant across the 24-hour day (recumbent testing, forced desynchrony, sleep deprivation nights) all still find an independent circadian effect on symptoms and periodic limb movements (2023-walters-circadian-rest-night-review.md, a review of this literature; the primary evidence behind it is graded C - see Emerging, below).
• Extending dopaminergic coverage into the second half of the night pharmacologically corrects late-night symptom/PLM recurrence: a randomized, double-blind, placebo-controlled crossover RCT (n=30) found adding sustained-release L-dopa at bedtime to regular-release L-dopa significantly reduced late-night PLM index and improved subjective sleep quality versus rr-L-dopa plus placebo (p<0.0001 for PLM index; 1999-collado-seidel-sustained-release-levodopa-rct.md).
• Evening/nighttime hormonal manipulation measurably changes RLS severity in a controlled trial: low-dose evening hydrocortisone modulated symptom severity in a placebo-controlled RCT (2008-hornyak-evening-hydrocortisone-rct.md), and near-infrared light therapy (delivered at fixed times) significantly reduced RLS severity versus sham in a single-blind RCT in hemodialysis patients (2018-mohammadi-near-infrared-light-therapy-rct.md).
• Exogenous melatonin worsens the motor (PLM) manifestation of RLS: in a controlled, within-subject crossover study, 3 mg melatonin given in the evening significantly increased the Suggested Immobilization Test PLM index compared to baseline and to bright light, while bright light (which suppresses endogenous melatonin) produced a small but significant reduction in sensory leg discomfort with no effect on PLM (2010-whittom-melatonin-bright-light-rls.md). This is the single most important controlled-trial finding behind the master plan's melatonin safety warning.
• A systematic review and methodological appraisal of RLS RCTs using the Suggested Immobilization Test (SIT) confirms that capturing RLS's circadian fluctuation properly requires multiple same-day SIT measurements, not a single-timepoint severity score, and that SIT-based PLM measures outperform single-night polysomnography for detecting treatment effects (2025-gorantla-sit-methodology-review.md).

Emerging C

• The circadian pattern is endogenous, not simply masked by rest/activity or sleep timing: a 28-hour modified constant-routine study (posture, light, and feeding held constant) found significant circadian variation in leg discomfort and PLM in both RLS patients and controls (p<0.01), and among vigilance, core temperature, and salivary melatonin, only melatonin's rise preceded (by roughly 2 hours) the worsening of RLS symptoms (2004-michaud-circadian-rhythm-melatonin.md). A forced-desynchrony protocol (20-hour imposed sleep-wake cycle, dissociating clock time from sleep pressure) found periodic limb movements in sleep have a statistically independent circadian rhythm (p<0.0001) even in subjects without diagnosed RLS (2010-duffy-forced-desynchrony-plms-circadian.md) - though in that non-RLS sample PLMS peaked just before, rather than after, the endogenous melatonin peak, the reverse order from Michaud's RLS finding.
• The two founding 1999 studies (Hening et al. and Trenkwalder et al.) independently established that symptom/PLM peaks track the falling limb of each patient's own core-body-temperature curve and persist through a full night of enforced sleep deprivation, arguing against a purely sleep-dependent explanation (1999-hening-circadian-motor-restlessness-temperature.md, 1999-trenkwalder-circadian-sleep-deprivation-plm.md); peak symptom/PLM expression clustered around midnight-1am with a trough around 9-11am in both studies.
• Central and peripheral nervous system excitability show a circadian pattern in RLS that is present at night but not in the day: TMS-measured active motor threshold was significantly lower at night in RLS than controls (28.5+/-6.2% vs 40.4+/-8.4%, p=0.006), with daytime measures equivalent between groups (2012-gunduz-circadian-cortical-excitability.md); brainstem blink-reflex (R2) recovery, which normally decreases at night in healthy people, instead stayed elevated at night in RLS (2016-uzun-blink-reflex-circadian-brainstem.md); and spinal flexor-withdrawal, crossed-extensor, and plantar reflexes were each significantly larger/faster in the evening than the morning specifically within the RLS group (2017-dafkin-flexor-withdrawal-crossed-extensor-circadian.md, 2017-dafkin-plantar-reflex-circadian-rls.md).
• Circadian dopamine receptor sensitivity and dopamine metabolism differ by time of day specifically in RLS: a randomized day/night L-dopa neuroendocrine challenge found significantly greater nighttime prolactin suppression and growth-hormone rise in RLS patients than controls, correlating with PLM index (2004-garcia-borreguero-neuroendocrine-ldopa-circadian.md); however, real-time striatal dopamine transporter density itself showed no diurnal difference on PET (see Contested, below), suggesting the oscillating element is downstream of transporter density (receptor sensitivity, extracellular dopamine dynamics, or the melatonin-dopamine interaction).
• Molecular/genetic evidence for clock involvement is accumulating: CLOCK gene haplotypes are associated with RLS risk in two independent case-control studies (2014-jung-clock-npas2-schizophrenia-rls.md, 2021-seo-clock-gene-variants-rls-koreans.md), and an exploratory transcriptomic study found 12 circadian clock genes significantly downregulated in RLS alongside upregulated ferroptosis/iron-oxidative-stress pathways (2025-mogavero-ferroptosis-circadian-genes.md), with a second, independent PBMC transcriptomic comparison against Parkinson's disease also flagging circadian-pathway genes as relatively distinctive in RLS (2026-mogavero-pbmc-transcriptomics-pd-rls.md).
• Peripheral microvascular/hypoxia physiology also varies diurnally in RLS: laser Doppler flowmetry found significantly higher intramuscular leg blood flow in the morning (17.9 perfusion units) than evening (12.1 perfusion units, p=0.004) in RLS patients, a difference not seen in controls (13.1 vs 12.0, ns), consistent with an evening/night peripheral-hypoxia component layered on top of central circadian drivers (2014-oskarsson-diurnal-leg-blood-flow.md; reviewed in 2023-tang-circadian-rhythm-rls-review.md).
• Cognition and large-scale brain connectivity also show a diurnal pattern in RLS, though not uniformly in the "worse at night" direction: RLS patients made significantly more errors on an evening flanker interference task than in the morning (p=0.023), with reduced attentional-selection EEG activity in the evening (2017-zhang-eeg-circadian-cognition-flanker.md); yet the same research group found RLS patients paradoxically performed better on a related Simon conflict task in the evening, due to reduced automatic visuo-motor priming (2018-zhang-simon-task-visuomotor-circadian.md). Resting-state fMRI found thalamic default-mode-network connectivity increased in the morning and decreased in the evening in RLS, with morning thalamic connectivity negatively correlated with symptom severity (2017-ku-default-mode-network-diurnal.md).
• Real-world circadian entrainment problems track with RLS: rotating night-shift nurses with RLS reported worse sleep quality tied to chronotype and shift pattern (2019-uekata-shift-work-chronotype-rls-nurses.md); a case report of a totally blind adolescent documented a non-24-hour sleep-wake disorder with comorbid RLS symptoms (2019-tubbs-blind-non24-rls-case-report.md); and RLS prevalence has been reported to vary with geographic latitude, a pattern proposed to relate to photoperiod/melatonin exposure (2011-koo-latitude-rls-prevalence-melatonin.md).
• One exploratory, actigraphy-based pilot (n=21 RLS vs 21 insomnia controls) estimated dim light melatonin onset (eDLMO) non-invasively and found a non-significant trend toward delayed phase in RLS (10:01 PM vs 9:33 PM, p=0.069), with a later estimated phase modestly associated with greater symptom severity within the RLS group (beta=0.073, p=0.038) (2026-bang-actigraphy-dlmo-rls.md).
• Real-world pharmacovigilance corroborates the melatonin-worsens-RLS signal from an entirely independent data source: a 20-year FAERS analysis of the melatonin-receptor agonist ramelteon flagged RLS among its rare but clinically notable adverse-event signals (2025-wu-ramelteon-faers-rls-signal.md), and a published clinical case report documented RLS symptoms resolving within one week of stopping self-initiated melatonin (2024-buntak-melatonin-discontinuation-case.md).

Contested or negative findings

• Melatonin's effect on periodic limb movements is direction-dependent on diagnosis: it worsens PLM measured by the SIT in primary RLS (2010-whittom-melatonin-bright-light-rls.md), but an earlier open-label trial found 3 mg melatonin nightly significantly reduced PLM index and improved actigraphic movement in patients with periodic limb movement disorder specifically WITHOUT RLS (2001-kunz-melatonin-plmd-open-trial.md) - implying the RLS-specific hyperdopaminergic/hypoadenosinergic state, not periodic limb movements as a general phenomenon, is what makes melatonin risky. A separate randomized trial of 3 mg melatonin in a mixed neuropsychiatric sleep-disorder sample (including some RLS/PLM patients, not analyzed separately) found melatonin improved REM sleep architecture with no reported worsening (2004-kunz-melatonin-rem-sleep-rct.md), underscoring that the safety warning should be scoped to primary RLS, not stated as a blanket claim against melatonin in every sleep disorder.
• Striatal dopamine transporter (DAT) binding potential, measured in vivo by PET, is reduced overall in RLS but shows no significant diurnal (day-vs-night) variation, in either RLS patients or controls (2011-earley-dopamine-transporter-pet-diurnal.md) - a genuine negative finding that argues against placing the simulation's circadian "dial" at the level of transporter density, even though other dopaminergic measures (receptor sensitivity, neuroendocrine challenge response) do show clear time-of-day differences.
• The direction of cognitive circadian change in RLS is not uniformly "worse at night": attentional selection (flanker task) is impaired in the evening, but automatic visuo-motor priming (Simon task) is reduced in the evening in a way that improves performance (2017-zhang-eeg-circadian-cognition-flanker.md, 2018-zhang-simon-task-visuomotor-circadian.md) - a genuine complication for any simulation that assumes a single, monotonic "evening deficit" applies uniformly across all brain functions.
• Whether RLS's circadian pattern is a single unified rhythm or the superposition of several quasi-independent oscillators (dopamine, iron/hypoxia, opioids, adenosine) acting on overlapping but not identical schedules remains unresolved; different studies place the phase of maximal excitability, maximal symptom severity, and melatonin acrophase in slightly different, not perfectly aligned, clock windows (compare 2004-michaud-circadian-rhythm-melatonin.md, 2012-gunduz-circadian-cortical-excitability.md, and 2014-oskarsson-diurnal-leg-blood-flow.md), a discrepancy the primary Walters and Zee review explicitly attributes partly to differing temporal resolution between studies (2023-walters-circadian-rest-night-review.md).

Mechanistic picture in 5 bullet points

• A circadian pacemaker output (indexed by core body temperature and, especially, endogenous melatonin onset) times the well-known evening dip in central dopaminergic tone: melatonin's known inhibitory effect on central dopamine release, combined with the observation that melatonin's rise reliably precedes RLS symptom worsening by about 2 hours, makes melatonin secretion the most plausible proximal circadian trigger identified to date (2004-michaud-circadian-rhythm-melatonin.md, reviewed in 2023-walters-circadian-rest-night-review.md).
• Dopaminergic treatment itself has chronobiotic (phase-shifting) properties, not just symptomatic ones: 3 weeks of open-label L-dopa advanced DLMO from a pretreatment baseline of ~21:00 to ~18:50 (p<0.05), with a larger advance in patients who went on to develop augmentation (2004-garcia-borreguero-dlmo-levodopa-chronobiotic.md) - a candidate molecular link between chronic dopaminergic exposure and the progressively earlier symptom onset that defines augmentation.
• This circadian dopaminergic dip does not operate through a change in raw dopamine transporter density (no diurnal PET difference; 2011-earley-dopamine-transporter-pet-diurnal.md) but through time-of-day differences in downstream receptor/neuroendocrine sensitivity (greater nighttime prolactin/GH response to an L-dopa challenge; 2004-garcia-borreguero-neuroendocrine-ldopa-circadian.md) and, at the molecular level, coordinated downregulation of clock genes alongside iron/ferroptosis pathway activation (2025-mogavero-ferroptosis-circadian-genes.md).
• The evening/night dopaminergic dip is expressed downstream as measurable central and peripheral hyperexcitability confined to the symptomatic window: cortical motor threshold drops, brainstem blink-reflex inhibition is lost, and spinal reflexes speed up specifically at night in RLS but not in controls (2012-gunduz-circadian-cortical-excitability.md, 2016-uzun-blink-reflex-circadian-brainstem.md, 2017-dafkin-plantar-reflex-circadian-rls.md), while peripheral leg muscle blood flow independently drops in the evening, adding a hypoxic component on the same schedule (2014-oskarsson-diurnal-leg-blood-flow.md).
• The rhythm is genuinely endogenous (clock-driven) rather than purely a consequence of rest, posture, or accumulated sleep pressure: constant-routine and forced-desynchrony protocols that hold these confounds constant still find a significant, independent circadian effect on symptoms and PLMS (2004-michaud-circadian-rhythm-melatonin.md, 2010-duffy-forced-desynchrony-plms-circadian.md), consistent with genetic (CLOCK-variant) association evidence for a true molecular-clock contribution (2014-jung-clock-npas2-schizophrenia-rls.md, 2021-seo-clock-gene-variants-rls-koreans.md).

Implications for a non-drug long-term plan for people currently on dopamine agonists or alpha-2-delta ligands

• Warn explicitly against self-initiated exogenous melatonin for sleep. This is the single highest-priority, most actionable safety point from this topic: a controlled crossover trial showed melatonin worsens PLM in RLS (2010-whittom-melatonin-bright-light-rls.md), an independent real-world pharmacovigilance signal flags RLS as an adverse effect of a melatonin-receptor agonist (2025-wu-ramelteon-faers-rls-signal.md), and a published case shows rapid clinical resolution after stopping melatonin (2024-buntak-melatonin-discontinuation-case.md). Ask every patient specifically about OTC melatonin use, since patients themselves may not connect it to their RLS worsening.
• Re-time existing symptomatic medication to the patient's own symptom-onset time, not to a fixed "before bedtime" rule. Because most patients already feel symptoms in the early evening or afternoon, dosing gabapentin, pregabalin, or an immediate-release dopamine agonist only 1-3 hours before bedtime (the standard guideline wording) may be systematically too late; matching each drug's time-to-peak to the individual's actual symptom-onset time, reassessed at each visit, is a plausible, low-risk optimization worth discussing with the prescriber (2024-garcia-malo-circadian-dosing-posology.md) - though this specific re-timing strategy itself has not yet been tested in a trial.
• Consider bright-light exposure timed away from the symptomatic window as an adjunct, not exogenous melatonin, for people who want a circadian-based non-drug tool: bright light modestly reduced sensory discomfort (though not PLM) in a controlled trial (2010-whittom-melatonin-bright-light-rls.md), and near-infrared light delivered on a fixed schedule reduced RLS severity in a separate RCT in hemodialysis patients (2018-mohammadi-near-infrared-light-therapy-rct.md).
• Treat CBT-I sleep-restriction therapy with RLS-specific caution. Experimentally imposed sleep deprivation tends to worsen RLS motor and sensory symptoms rather than simply causing daytime sleepiness (1999-trenkwalder-circadian-sleep-deprivation-plm.md), and RLS patients show a paradoxical, hyperarousal- consistent resistance to the sleepiness that chronic sleep restriction should otherwise produce (2008-gamaldo-daytime-alertness-sleep-restricted-controls.md); standard CBT-I protocols should be modified, not applied unmodified, in this population.
• Screen shift workers and frequent long-haul travelers with RLS proactively. Rotating-shift RLS patients report measurably worse sleep quality tied to shift pattern and chronotype (2019-uekata-shift-work-chronotype-rls-nurses.md), and the circadian model predicts (though this specific claim is drawn from a review's synthesis rather than an independently-confirmed primary trial in this KB) that jet lag and shift work can shift the clock time of symptom onset - both are reasons to ask about occupational and travel schedules when a patient's usual evening dosing plan suddenly stops working (2024-garcia-malo-circadian-dosing-posology.md).
• Do not assume every dopaminergic or cognitive effect gets uniformly worse at night. Because some RLS-associated cognitive processes paradoxically improve in the evening (2018-zhang-simon-task-visuomotor-circadian.md), blanket advice to avoid all cognitively demanding evening activity is not well supported; the specific, well-replicated target for evening/night scheduling is motor and sensory symptom relief (activity, light, and medication timing), not a general "protect the evening" rule.

Open questions

• No RCT has directly tested Garcia-Malo et al.'s proposed re-timing strategy (dosing to individual symptom-onset time rather than to bedtime) against standard guideline-timed dosing (2024-garcia-malo-circadian-dosing-posology.md); this is a concrete, low-cost trial the field is missing.
• Whether bright light can actually phase-shift RLS symptom timing (as opposed to producing a same-day symptomatic effect) has never been tested with circadian-phase markers (DLMO, core temperature) before and after light exposure - the primary review explicitly calls for this study to be done (2023-walters-circadian-rest-night-review.md).
• No dedicated forced-desynchrony or constant-routine study has yet been conducted in patients with confirmed RLS scanned across a fully dissociated circadian/sleep protocol; the only forced-desynchrony PLMS data in this KB comes from non-RLS subjects with elevated PLM index (2010-duffy-forced-desynchrony-plms-circadian.md), and Michaud's RLS constant-routine study, while rigorous, was only 28 hours (2004-michaud-circadian-rhythm-melatonin.md).
• The circadian variation of CSF/serum iron and hepcidin specifically in RLS (as opposed to serum iron's well-known general circadian rhythm, or ferritin/transferrin's general lack of one) is not directly confirmed by a primary study in this knowledge base; the strongest available statement is a secondary review's citation of an early-morning-vs-evening CSF ferritin comparison (2024-garcia-malo-circadian-dosing-posology.md), not an independently verified primary source, so this specific number should be treated as "not independently confirmed" until a primary paper is located and read.
• Real jet-lag/travel and shift-work phase-shift data for RLS symptoms (rather than the sleep-quality survey data that exists, or a secondary review's summary of unconfirmed primary sources) is essentially absent; this is a clearly identified but unfilled gap flagged by the field's own leading circadian reviewers (2023-walters-circadian-rest-night-review.md).
• Whether the apparently distinct circadian "oscillators" this topic identifies (melatonin/dopamine, peripheral blood flow/hypoxia, clock-gene expression, cortical/spinal excitability) are all downstream of one master clock signal or are quasi-independently entrained is unresolved and directly affects how many free parameters the simulation's circadian module needs.

Parameters for the simulation

• Symptom/PLM peak clock time: approximately midnight-1:00 AM; trough: approximately 9:00-11:00 AM (1999-hening-circadian-motor-restlessness-temperature.md, PMID 10566908; 1999-trenkwalder-circadian-sleep-deprivation-plm.md, PMID 9918351).
• Broader "higher-risk zone": approximately 11:00 PM-4:00 AM; "protective zone": approximately 9:00 AM-2:00 PM (2024-garcia-malo-circadian-dosing-posology.md, PMID 39698169).
• Untreated RLS DLMO (dim light melatonin onset) baseline: ~21:00 +/- 1:20; after 3 weeks of L-dopa, DLMO phase-advances to ~18:50 +/- 0:55 (p<0.05), a shift of roughly 2h10m (2004-garcia-borreguero-dlmo-levodopa-chronobiotic.md, PMID 15223002).
• Actigraphy-estimated DLMO in RLS (non-invasive pilot estimate): ~22:01 vs ~21:33 in insomnia controls (trend, p=0.069); each unit-later eDLMO associated with a 0.073-point rise in IRLS severity (p=0.038) (2026-bang-actigraphy-dlmo-rls.md, PMID 42417186).
• Melatonin-to-symptom phase lag: endogenous melatonin acrophase precedes leg-discomfort/PLMS acrophase by approximately 2 hours in RLS patients (2004-michaud-circadian-rhythm-melatonin.md, PMID 14991815) - the opposite temporal order (PLMS peaking just before melatonin) was found in non-RLS elevated-PLM subjects under forced desynchrony (2010-duffy-forced-desynchrony-plms-circadian.md, PMID 21093364), an unresolved discrepancy the simulation should flag rather than silently resolve.
• Amplitude of evening motor-cortical excitability increase: active motor threshold 28.5% (RLS, night) vs 40.4% (controls, night), p=0.006, with no group difference by day (2012-gunduz-circadian-cortical-excitability.md, PMID 22305328).
• Amplitude of evening/morning peripheral blood-flow drop: 17.9 perfusion units (AM) vs 12.1 (PM), p=0.004, in RLS vs a flat 13.1/12.0 profile in controls (2014-oskarsson-diurnal-leg-blood-flow.md, PMID 24521165).
• Exogenous melatonin dose/timing shown to worsen PLM experimentally: 3 mg given at approximately 19:00 (2010-whittom-melatonin-bright-light-rls.md, PMID 20226733); bright light suppressing endogenous melatonin (3,000 lux, 19:00-midnight) modestly reduced sensory discomfort in the same trial, per the fuller description in 2023-walters-circadian-rest-night-review.md (PMID 37181571).
• Guideline vs proposed drug time-to-peak/dosing-time mismatch (for re-timing modelling): gabapentin time-to-peak 2-4h (guideline dosing: 2h pre-bed); gabapentin enacarbil time-to-peak 5-7.3h (guideline dosing: fixed 5:00-7:00 PM); pregabalin time-to-peak 1.5h (guideline dosing: 2h pre-bed); immediate-release pramipexole time-to-peak ~2h (guideline dosing: 2-3h pre-bed) (2024-garcia-malo-circadian-dosing-posology.md, PMID 39698169).
• Clock-gene transcriptomic signal: 12 circadian clock genes significantly downregulated in RLS in an exploratory transcriptomic comparison (n=17 RLS vs 39 controls) (2025-mogavero-ferroptosis-circadian-genes.md, PMID 40867628) - not yet independently replicated, so treat as a provisional, not confirmed, magnitude.

Notes in this topic

New notes added in this sweep:

• 2023-walters-circadian-rest-night-review.md - the field's most complete narrative review; establishes worse-at-rest and worse-at-night as separable, reviews CLOCK genes, cortical/spinal/peripheral excitability, cognition, dopamine, and melatonin/light evidence.
• 2024-garcia-malo-circadian-dosing-posology.md - review proposing symptom-onset-timed (not bedtime-timed) dosing; source of the higher-risk/protective-zone clock windows and the PK/dosing-timing table.
• 2026-bang-actigraphy-dlmo-rls.md - actigraphy-estimated DLMO pilot; trend toward delayed phase in RLS, later phase linked to greater severity.
• 2025-mogavero-ferroptosis-circadian-genes.md - transcriptomic study finding 12 circadian clock genes downregulated in RLS alongside upregulated ferroptosis pathways.
• 2026-mogavero-pbmc-transcriptomics-pd-rls.md - independent PBMC transcriptomic comparison (RLS vs Parkinson's) also flagging circadian-pathway genes as relatively distinctive in RLS.
• 2025-stefani-circadian-pattern-handbook.md - current (2025) Handbook of Clinical Neurology chapter, tertiary confirmation of the multi-system circadian picture.
• 2011-earley-dopamine-transporter-pet-diurnal.md - PET study: striatal DAT reduced in RLS but with no diurnal (day/night) variation - a negative/contested finding.
• 2017-zhang-eeg-circadian-cognition-flanker.md - evening-worse flanker task performance and EEG N1 in RLS.
• 2018-zhang-simon-task-visuomotor-circadian.md - companion study: evening-better Simon-task performance via reduced visuo-motor priming - a contested, non-uniform cognitive finding.
• 2025-wu-ramelteon-faers-rls-signal.md - 20-year FAERS pharmacovigilance analysis flagging RLS as a rare adverse-event signal for the melatonin-receptor agonist ramelteon.
• 2004-kunz-melatonin-rem-sleep-rct.md - RCT of melatonin in a mixed sleep-disorder sample (incl. RLS/PLM, not separately analyzed): improved REM sleep architecture, no reported worsening - an important nuance.
• 2025-gorantla-sit-methodology-review.md - systematic review establishing multiple-SIT-per-day methodology as necessary to capture RLS's circadian fluctuation.
• 2017-ku-default-mode-network-diurnal.md - diurnal reversal of thalamic default-mode-network connectivity in RLS, correlated with severity.
• 2001-kunz-melatonin-plmd-open-trial.md - melatonin SUPPRESSES PLM in periodic limb movement disorder without RLS - directionally opposite to the RLS-specific finding, a key contested-findings citation.
• 2016-uzun-blink-reflex-circadian-brainstem.md - brainstem blink-reflex circadian modulation is lost in RLS.
• 2024-buntak-melatonin-discontinuation-case.md - single-patient case report: RLS resolved within a week of stopping self-initiated melatonin.

Pre-existing notes found in the knowledge base (written by other project agents), also relevant to this topic:

• 1999-hening-circadian-motor-restlessness-temperature.md - founding circadian study; symptom/PLM worsening tracks the falling limb of core body temperature, persists through sleep deprivation.
• 2008-gamaldo-daytime-alertness-sleep-restricted-controls.md - RLS patients show paradoxically greater objective alertness than sleep-matched controls at both morning and evening timepoints (t=3.80, p=0.001; U=31.0, p<0.001).
• 1999-trenkwalder-circadian-sleep-deprivation-plm.md - founding circadian study; PLM/symptom peak midnight-1am, trough 9-11am, independent of sleep, and sleep deprivation tends to worsen RLS motor/sensory symptoms further without abolishing the pattern - a caution for CBT-I sleep restriction.
• 1999-collado-seidel-sustained-release-levodopa-rct.md - RCT: adding sr-L-dopa at bedtime reduces late-night PLM recurrence versus rr-L-dopa alone.
• 2004-michaud-circadian-rhythm-melatonin.md - 28-h constant routine; melatonin rise is the only marker that precedes RLS symptom worsening (~2h lag).
• 2004-garcia-borreguero-dlmo-levodopa-chronobiotic.md - L-dopa phase-advances DLMO from ~21:00 to ~18:50; larger advance in patients who later develop augmentation.
• 2004-garcia-borreguero-neuroendocrine-ldopa-circadian.md - randomized day/night L-dopa challenge; greater nighttime prolactin suppression/GH rise in RLS than controls.
• 2005-clemens-tyrosine-hydroxylase-circadian-d3ko.md - animal model: circadian tyrosine-hydroxylase expression is reversed in the spinal cord of D3-receptor-knockout mice.
• 2005-michaud-circadian-immobility-symptoms.md - circadian variation in how immobility (SIT) affects RLS symptoms.
• 2008-hornyak-evening-hydrocortisone-rct.md - RCT: low-dose evening hydrocortisone modulates RLS severity.
• 2010-duffy-forced-desynchrony-plms-circadian.md - forced-desynchrony protocol; PLMS have an independent circadian rhythm (p<0.0001) even without diagnosed RLS.
• 2010-kripke-clock-gene-genotyping-sleep-disorders.md - clock-gene genotyping across sleep-disorder patients, including RLS.
• 2010-whittom-melatonin-bright-light-rls.md - controlled crossover: melatonin worsens PLM, bright light modestly reduces sensory discomfort - the central melatonin-safety study.
• 2011-koo-latitude-rls-prevalence-melatonin.md - RLS prevalence varies with geographic latitude, proposed link to photoperiod/melatonin.
• 2011-lattova-hpa-axis-sleep-disorders.md - HPA-axis/cortisol comparison across sleep disorders including RLS.
• 2011-dowling-d3-receptor-iron-deficient-mice.md - iron deficiency and D3-receptor knockout each shift circadian wheel-running activity in mice.
• 2012-gunduz-circadian-cortical-excitability.md - TMS: motor cortical excitability abnormally increased at night, not day, in RLS.
• 2014-jung-clock-npas2-schizophrenia-rls.md - CLOCK/NPAS2 gene polymorphism association with RLS.
• 2014-oskarsson-diurnal-leg-blood-flow.md - laser Doppler flowmetry: leg muscle blood flow drops significantly from morning to evening in RLS but not controls.
• 2017-dafkin-flexor-withdrawal-crossed-extensor-circadian.md - spinal flexor-withdrawal/crossed-extensor reflexes larger and faster in the evening in RLS.
• 2017-dafkin-plantar-reflex-circadian-rls.md - plantar reflex excitability increased in the evening in RLS.
• 2018-mohammadi-near-infrared-light-therapy-rct.md - RCT: near-infrared light therapy reduces RLS severity in hemodialysis patients.
• 2019-tubbs-blind-non24-rls-case-report.md - case report: non-24-hour sleep-wake disorder with comorbid RLS symptoms in a totally blind adolescent.
• 2019-uekata-shift-work-chronotype-rls-nurses.md - rotating-shift nurses with RLS report worse sleep quality tied to shift pattern and chronotype.
• 2021-seo-clock-gene-variants-rls-koreans.md - CLOCK gene variant association with RLS in a Korean cohort.
• 2023-tang-circadian-rhythm-rls-review.md - review of circadian mechanisms (iron, dopamine, melatonin, peripheral hypoxia/microvascular function) in RLS.