Place · Level 3
CBT-I & Sleep Hygiene · the Evidence-Based Self-Help for Insomnia
慢性失眠的一线治疗不是安眠药, 是 CBT-I (ACP 2016 / AASM 2021) · 过度觉醒模型 · 刺激控制 (重建条件反射) + 睡眠限制 (Triple-R: 攒睡眠压力) + 认知重构 + 睡眠卫生 (地板不是天花板) · 为什么它长期赢过药 · 数字 CBT-I 普及之路
Story path
- 1Why behavioral, not pills, is first-lineWhy behavioral, not pills, is first-line
- 2The hyperarousal model · why insomnia won't stopThe hyperarousal model · why insomnia won't stop
- 3Stimulus control · re-weld 'bed = sleep'Stimulus control · re-weld 'bed = sleep'
- 4Sleep restriction · the counterintuitive 'sleep less first'Sleep restriction · the counterintuitive 'sleep less first'
- 5Cognitive work + hygiene · break the loop / floor not ceilingCognitive work + hygiene · break the loop / floor not ceiling
- 6What to do · evidence, digital CBT-I, when to see a clinician + atlas loopWhat to do · evidence, digital CBT-I, when to see a clinician + atlas loop
Chapter 1
Why behavioral, not pills, is first-line
Why behavioral, not pills, is first-line
If you sleep badly for a long stretch, the first instinct is usually 'take something' — melatonin, an OTC sleep aid, or a Z-drug from a doctor. But the first-line treatment for chronic insomnia in every major guideline is not a drug — it is a behavioral/cognitive program: CBT-I (Cognitive Behavioral Therapy for Insomnia).
This island is the 'how to treat it' of the sleep cluster — its division of labor with the other two islands is clean:
`sleep-architecture` is how the building of sleep is constructed (N3 deep sleep / REM / the two-process model) — the 'why'`insomnia` is how insomnia is classified and diagnosed, and what the drug risks are — the 'what it is'This island is how the genuinely effective non-drug intervention works — the 'how to do it,' taking the five CBT-I components mentioned in `insomnia` and unpacking each one's mechanism
How blunt the guidelines are
The American College of Physicians (ACP) 2016 guideline (Qaseem et al.): explicitly recommends that all adults with chronic insomnia receive CBT-I as the initial treatment, with medication only as an add-on 'when CBT-I alone is insufficient,' decided through shared decision-making weighing benefits and harms. This is an official document that strongly recommends CBT-IThe American Academy of Sleep Medicine (AASM) 2021 clinical practice guideline (Edinger et al.): gives a strong recommendation (high certainty) for multicomponent CBT-I in adults with chronic insomnia
Why behavioral and not a pill — the one-line mechanism
A drug (sedation) treats the surface — 'I can't sleep right now.' CBT-I treats the maintaining mechanism — 'why won't this insomnia stop' (the hyperarousal model, next scene). The former covers the symptom; the latter dismantles the root — which is exactly why drugs rebound on discontinuation while CBT-I's effects persist after treatment ends (Trauer 2015, later).
One honest line: CBT-I is not 'toughing it out without drugs' — it is a treatment that is more causal and better value long-term than drugs. It is not mysterious and is very low-risk; the real barrier is availability — most people with insomnia were never told it exists (detailed in the last scene).
This island is the 'how to treat it' of the sleep cluster — its division of labor with the other two islands is clean:
`sleep-architecture` is how the building of sleep is constructed (N3 deep sleep / REM / the two-process model) — the 'why'`insomnia` is how insomnia is classified and diagnosed, and what the drug risks are — the 'what it is'This island is how the genuinely effective non-drug intervention works — the 'how to do it,' taking the five CBT-I components mentioned in `insomnia` and unpacking each one's mechanism
How blunt the guidelines are
The American College of Physicians (ACP) 2016 guideline (Qaseem et al.): explicitly recommends that all adults with chronic insomnia receive CBT-I as the initial treatment, with medication only as an add-on 'when CBT-I alone is insufficient,' decided through shared decision-making weighing benefits and harms. This is an official document that strongly recommends CBT-IThe American Academy of Sleep Medicine (AASM) 2021 clinical practice guideline (Edinger et al.): gives a strong recommendation (high certainty) for multicomponent CBT-I in adults with chronic insomnia
Why behavioral and not a pill — the one-line mechanism
A drug (sedation) treats the surface — 'I can't sleep right now.' CBT-I treats the maintaining mechanism — 'why won't this insomnia stop' (the hyperarousal model, next scene). The former covers the symptom; the latter dismantles the root — which is exactly why drugs rebound on discontinuation while CBT-I's effects persist after treatment ends (Trauer 2015, later).
One honest line: CBT-I is not 'toughing it out without drugs' — it is a treatment that is more causal and better value long-term than drugs. It is not mysterious and is very low-risk; the real barrier is availability — most people with insomnia were never told it exists (detailed in the last scene).
Chapter 2
The hyperarousal model · why insomnia won't stop
The hyperarousal model · why insomnia won't stop
To understand why CBT-I works, first understand how chronic insomnia maintains itself. The most widely accepted framework is the hyperarousal model (Riemann et al. 2010).
The core idea: people with insomnia run 'too awake,' everywhere, all the time
Riemann's *Sleep Medicine Reviews* review systematically maps the evidence and concludes that the core of primary insomnia is hyperarousal across levels — not merely 'can't sleep at night,' but a round-the-clock tilt toward the 'on' state, from molecule to system:
Autonomic: reduced heart-rate variability, elevated nighttime heart rate — sympathetic tone never switches offNeuroendocrine: heightened hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.-axis activity (cortisol/ACTH) — directly bordering the HPA axis in `chronic-stress`Electrophysiological: more high-frequency (β/γ) activity in the sleep EEG — the brain never truly goes offlineCognitive/emotional: racing pre-sleep thoughts, worry about and monitoring of sleep itselfImaging: at sleep onset, the wake-promoting regions that should 'cool down' don't fully come down
Why this explains 'won't stop' — the 3P maintenance model
Clinically, Spielman's 3P model is used to read the timeline of insomnia:
Predisposing: an inborn higher arousal level, anxious temperament — your 'baseline noise'Precipitating: a stressor / illness / jet lag ignites the acute insomniaPerpetuating: after the trigger is gone, insomnia keeps itself alive through the 'behaviors used to cope with insomnia' — going to bed early to 'catch up,' lying down by day, lying in bed scrolling and anxious, trying harder and harder to sleep
The key insight: acute insomnia mostly resolves on its own; what turns it chronic are the behaviors done 'to fight the insomnia,' which precisely feed the hyperarousal.
'Lie there longer and I'll surely sleep more' → the longer you lie awake in bed, the stronger the bed↔wakefulness/anxiety conditioning (exactly what stimulus control, next scene, dismantles)'I must get 8 hours tonight' → that effort to sleep and self-monitoring is itself a form of arousal — the harder you try, the less you sleep (exactly what cognitive restructuring dismantles)
CBT-I's whole logic follows from this: since insomnia self-perpetuates through 'hyperarousal + maintaining behaviors,' the treatment is to strip away the maintaining factors one by one and bring arousal down — not to 'knock the person out' with a drug. The next three scenes are the three matching wrenches.
The core idea: people with insomnia run 'too awake,' everywhere, all the time
Riemann's *Sleep Medicine Reviews* review systematically maps the evidence and concludes that the core of primary insomnia is hyperarousal across levels — not merely 'can't sleep at night,' but a round-the-clock tilt toward the 'on' state, from molecule to system:
Autonomic: reduced heart-rate variability, elevated nighttime heart rate — sympathetic tone never switches offNeuroendocrine: heightened hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.-axis activity (cortisol/ACTH) — directly bordering the HPA axis in `chronic-stress`Electrophysiological: more high-frequency (β/γ) activity in the sleep EEG — the brain never truly goes offlineCognitive/emotional: racing pre-sleep thoughts, worry about and monitoring of sleep itselfImaging: at sleep onset, the wake-promoting regions that should 'cool down' don't fully come down
Why this explains 'won't stop' — the 3P maintenance model
Clinically, Spielman's 3P model is used to read the timeline of insomnia:
Predisposing: an inborn higher arousal level, anxious temperament — your 'baseline noise'Precipitating: a stressor / illness / jet lag ignites the acute insomniaPerpetuating: after the trigger is gone, insomnia keeps itself alive through the 'behaviors used to cope with insomnia' — going to bed early to 'catch up,' lying down by day, lying in bed scrolling and anxious, trying harder and harder to sleep
The key insight: acute insomnia mostly resolves on its own; what turns it chronic are the behaviors done 'to fight the insomnia,' which precisely feed the hyperarousal.
'Lie there longer and I'll surely sleep more' → the longer you lie awake in bed, the stronger the bed↔wakefulness/anxiety conditioning (exactly what stimulus control, next scene, dismantles)'I must get 8 hours tonight' → that effort to sleep and self-monitoring is itself a form of arousal — the harder you try, the less you sleep (exactly what cognitive restructuring dismantles)
CBT-I's whole logic follows from this: since insomnia self-perpetuates through 'hyperarousal + maintaining behaviors,' the treatment is to strip away the maintaining factors one by one and bring arousal down — not to 'knock the person out' with a drug. The next three scenes are the three matching wrenches.
Chapter 3
Stimulus control · re-weld 'bed = sleep'
Stimulus control · re-weld 'bed = sleep'
Stimulus control is the most central component of CBT-I and the one with the hardest single-component evidence. What it repairs is exactly the conditioning broken in the last scene: bed ↔ wakefulness/anxiety.
Mechanism: this is a conditioning problem
In healthy sleep, the cues of bed, bedroom, and lying down are registered over years as the 'time to sleep' signal — get into bed and sleepiness arrives. But someone with chronic insomnia repeatedly experiences wakefulness + tossing + anxiety in bed, and over time the same cues are re-registered as 'time to be alert.' Hence the signature experience: 'I doze off instantly on the couch but snap awake in bed' — the bed itself has become a trigger for arousal.
Stimulus control uses a set of counterintuitive but precise instructions to weld that reflex back:
Only get into bed when genuinely sleepy (not 'because it's the time')The bed is only for sleep and sex — no working, scrolling, watching shows, or being anxious in bedIf still awake after ~20 minutes, get up, go to another room and do something quiet and low-stimulation (read something dull under dim light), return only when sleepyIf you wake at night and can't fall back asleep, get up too — don't lie in bed wrestling with insomniaNo matter how poorly you slept, get up at the same fixed time every day; no daytime catch-up, no long naps
Why 'get up if you can't sleep' actually helps sleep (rather than making you more tired)
Every 'lying in bed awake and anxious' episode reinforces the wrong association; leaving the bed means you stop paying into that wrong associationA fixed wake time + no daytime catch-up also banks sleep pressure (Process S in `sleep-architecture`) for the night, when it should be spentAfter a week or two, the bed is paired only with 'fall asleep fast' again — the reflex is rewritten
Evidence strength: the AASM 2021 guideline lists stimulus control alone as one of the separately recommendable behavioral therapies (Edinger et al.). It is often considered the single biggest contributor within multicomponent CBT-I, but the full program beats any single component (as `insomnia` stresses).
One honest line: in the first few days, 'get up if not asleep in 20 minutes' feels like 'now I'll sleep even less.' That is expected — you are trading a few uncomfortable days to win back a reflex you trained wrong yourself. Push through it and the bed becomes a place that makes you sleepy again. (For how sleep pressure accumulates at the mechanistic level, dive to the two-process model in `sleep-architecture`; for the classification of 'can't sleep,' dive to `insomnia`.)
Mechanism: this is a conditioning problem
In healthy sleep, the cues of bed, bedroom, and lying down are registered over years as the 'time to sleep' signal — get into bed and sleepiness arrives. But someone with chronic insomnia repeatedly experiences wakefulness + tossing + anxiety in bed, and over time the same cues are re-registered as 'time to be alert.' Hence the signature experience: 'I doze off instantly on the couch but snap awake in bed' — the bed itself has become a trigger for arousal.
Stimulus control uses a set of counterintuitive but precise instructions to weld that reflex back:
Only get into bed when genuinely sleepy (not 'because it's the time')The bed is only for sleep and sex — no working, scrolling, watching shows, or being anxious in bedIf still awake after ~20 minutes, get up, go to another room and do something quiet and low-stimulation (read something dull under dim light), return only when sleepyIf you wake at night and can't fall back asleep, get up too — don't lie in bed wrestling with insomniaNo matter how poorly you slept, get up at the same fixed time every day; no daytime catch-up, no long naps
Why 'get up if you can't sleep' actually helps sleep (rather than making you more tired)
Every 'lying in bed awake and anxious' episode reinforces the wrong association; leaving the bed means you stop paying into that wrong associationA fixed wake time + no daytime catch-up also banks sleep pressure (Process S in `sleep-architecture`) for the night, when it should be spentAfter a week or two, the bed is paired only with 'fall asleep fast' again — the reflex is rewritten
Evidence strength: the AASM 2021 guideline lists stimulus control alone as one of the separately recommendable behavioral therapies (Edinger et al.). It is often considered the single biggest contributor within multicomponent CBT-I, but the full program beats any single component (as `insomnia` stresses).
One honest line: in the first few days, 'get up if not asleep in 20 minutes' feels like 'now I'll sleep even less.' That is expected — you are trading a few uncomfortable days to win back a reflex you trained wrong yourself. Push through it and the bed becomes a place that makes you sleepy again. (For how sleep pressure accumulates at the mechanistic level, dive to the two-process model in `sleep-architecture`; for the classification of 'can't sleep,' dive to `insomnia`.)
Chapter 4
Sleep restriction · the counterintuitive 'sleep less first'
Sleep restriction · the counterintuitive 'sleep less first'
Sleep restriction therapy (SRT) is the most counterintuitive — and most dramatically effective — part of CBT-I: it asks someone who is already not sleeping enough to first shorten their time in bed even more. It sounds cruel, but the mechanism is very clean.
How it's done (the essentials)
First use a sleep diary to estimate your actual time asleep (e.g. 8 hours in bed, only 5 actually asleep)Compress time in bed to near your actual sleep time (e.g. set a 5–5.5-hour sleep window), with a fixed wake timeNo lounging in bed when not sleeping (pairs with stimulus control)When sleep efficiency (actual sleep ÷ time in bed) is ≥ 85–90% for a full week, add 15 minutes to the window each week, titrating up toward the duration you truly need
Mechanism · Maurer's Triple-R model (Maurer, Espie & Kyle 2018)
This *Sleep Medicine Reviews* mechanistic review breaks down why SRT works into three pathways acting at once — landing squarely on the hyperarousal model:
R1 · Restrict (→ bank sleep pressure): cutting time in bed over successive nights directly strengthens homeostatic sleep drive (Process S, adenosine pressure) and dampens pre-sleep hyperarousal — with high enough sleep pressure, you fall asleep faster and sleep more solidlyR2 · Regularise (→ tighten the clock): a fixed sleep window and wake time tighten circadian (Process C) control over sleep–wake, making sleep more stable and continuous (the S×C product from `sleep-architecture`)R3 · Recondition: when almost all time in bed is spent asleep → the bed↔sleep association is re-strengthened (same direction as stimulus control)
Key: this is 'pain first, reward later' by design
The first week or two feels sleepier (this is precisely the 'banking sleep pressure' process — and evidence it is working)But fragmentation and middle-of-the-night wakefulness drop fast, and sleep gets consolidated — the same sleep packed into a shorter window, so efficiency soarsThen the window is widened week by week, ending with shorter time in bed, higher efficiency, fewer awakenings
Evidence strength: Maurer's review reports a medium-to-large effect of SRT on sleep-continuity measures (onset latency, wake after sleep onset, sleep efficiency) and a large effect on insomnia severity (ISI); AASM 2021 gives a conditional recommendation for sleep restriction alone (Edinger et al.).
Safety red line (important): SRT transiently increases daytime sleepiness, so it needs therapist supervision and is not advised as full DIY — in particular, drivers/machine operators, people with bipolar disorder (sleep deprivation can trigger mania), uncontrolled epilepsy, and untreated sleep apnea (which needs CPAP, dive `sleep-apnea`) should not attempt it on their own. This is one reason CBT-I belongs in a professional frame.
How it's done (the essentials)
First use a sleep diary to estimate your actual time asleep (e.g. 8 hours in bed, only 5 actually asleep)Compress time in bed to near your actual sleep time (e.g. set a 5–5.5-hour sleep window), with a fixed wake timeNo lounging in bed when not sleeping (pairs with stimulus control)When sleep efficiency (actual sleep ÷ time in bed) is ≥ 85–90% for a full week, add 15 minutes to the window each week, titrating up toward the duration you truly need
Mechanism · Maurer's Triple-R model (Maurer, Espie & Kyle 2018)
This *Sleep Medicine Reviews* mechanistic review breaks down why SRT works into three pathways acting at once — landing squarely on the hyperarousal model:
R1 · Restrict (→ bank sleep pressure): cutting time in bed over successive nights directly strengthens homeostatic sleep drive (Process S, adenosine pressure) and dampens pre-sleep hyperarousal — with high enough sleep pressure, you fall asleep faster and sleep more solidlyR2 · Regularise (→ tighten the clock): a fixed sleep window and wake time tighten circadian (Process C) control over sleep–wake, making sleep more stable and continuous (the S×C product from `sleep-architecture`)R3 · Recondition: when almost all time in bed is spent asleep → the bed↔sleep association is re-strengthened (same direction as stimulus control)
Key: this is 'pain first, reward later' by design
The first week or two feels sleepier (this is precisely the 'banking sleep pressure' process — and evidence it is working)But fragmentation and middle-of-the-night wakefulness drop fast, and sleep gets consolidated — the same sleep packed into a shorter window, so efficiency soarsThen the window is widened week by week, ending with shorter time in bed, higher efficiency, fewer awakenings
Evidence strength: Maurer's review reports a medium-to-large effect of SRT on sleep-continuity measures (onset latency, wake after sleep onset, sleep efficiency) and a large effect on insomnia severity (ISI); AASM 2021 gives a conditional recommendation for sleep restriction alone (Edinger et al.).
Safety red line (important): SRT transiently increases daytime sleepiness, so it needs therapist supervision and is not advised as full DIY — in particular, drivers/machine operators, people with bipolar disorder (sleep deprivation can trigger mania), uncontrolled epilepsy, and untreated sleep apnea (which needs CPAP, dive `sleep-apnea`) should not attempt it on their own. This is one reason CBT-I belongs in a professional frame.
Chapter 5
Cognitive work + hygiene · break the loop / floor not ceiling
Cognitive work + hygiene · break the loop / floor not ceiling
Stimulus control and sleep restriction address 'behavior,' but hyperarousal also has a cognitive leg: the catastrophic worry about sleep is itself a form of arousal. These are CBT-I's third and fourth components.
Cognitive restructuring · break the 'the more I want to sleep, the less I sleep' loop
People with insomnia often fall into a set of self-fulfilling catastrophic thoughts that directly raise pre-sleep arousal:
'If I can't sleep tonight, I'm finished tomorrow' → manufactures anticipatory anxiety, tense the moment you get into bed'Normal people sleep 8 hours, I must hit that' → turns sleep into a task and an exam, and sleep is precisely the thing that retreats the harder you push'I haven't slept well all week, my body is going to break' → overestimates the harm of a single bad night
Cognitive restructuring is not 'positive-thinking pep talk' — it is replacing these beliefs with evidence:
'I might be tired, but I'll get through' — one bad night is not catastrophic (you've gotten through before)'My individual sleep need may be 7 hours, not 8' — see 'is 6 hours insomnia' in `insomnia`'Acute insomnia mostly resolves on its own' — the real problem is the reaction to insomnia, not the insomnia itselfPaired with relaxation training (progressive muscle relaxation / diaphragmatic breathing / mindfulness): directly applies a 'brake' to the autonomic nervous system, lowering physiological arousal
The key mechanism: let go of the effort to 'must fall asleep,' arousal naturally drops, and sleep comes back — the cognitive-therapy counterpart of the hyperarousal model.
Sleep hygiene · necessary but not sufficient — it is the floor, not the ceiling
This is the most misunderstood piece. The public often treats 'sleep hygiene' as the treatment for insomnia, but the honest evidence is that sleep hygiene alone works weakly for chronic insomnia — in many CBT-I trials it is even used as the 'control / placebo' arm (Trauer 2015 does exactly this).
The hygiene checklist itself is not wrong (it removes the obvious obstacles to sleep):
Caffeine cutoff in early afternoon (half-life 5–6 h, dive `caffeine-l-theanine`)Less alcohol before bed (it dismantles REM and fragments the second half of the night, dive `sleep-architecture`)A cool, dark, quiet bedroom; control blue light for the hour before bedRegular exercise (but not too late)
Why 'hygiene only' isn't enough: it clears the precipitating / interfering factors, but the engine of chronic insomnia is hyperarousal + maintaining behaviors (conditioning, catastrophizing, burning time in bed). Hygiene never touches that engine.
The right framing: sleep hygiene is the floor (fill the obvious holes first), while CBT-I's stimulus control + sleep restriction + cognitive restructuring is the ceiling (actually dismantling the engine). Treating hygiene as the whole thing is the most common 'doing something correct but insufficient' in insomnia care.
Cognitive restructuring · break the 'the more I want to sleep, the less I sleep' loop
People with insomnia often fall into a set of self-fulfilling catastrophic thoughts that directly raise pre-sleep arousal:
'If I can't sleep tonight, I'm finished tomorrow' → manufactures anticipatory anxiety, tense the moment you get into bed'Normal people sleep 8 hours, I must hit that' → turns sleep into a task and an exam, and sleep is precisely the thing that retreats the harder you push'I haven't slept well all week, my body is going to break' → overestimates the harm of a single bad night
Cognitive restructuring is not 'positive-thinking pep talk' — it is replacing these beliefs with evidence:
'I might be tired, but I'll get through' — one bad night is not catastrophic (you've gotten through before)'My individual sleep need may be 7 hours, not 8' — see 'is 6 hours insomnia' in `insomnia`'Acute insomnia mostly resolves on its own' — the real problem is the reaction to insomnia, not the insomnia itselfPaired with relaxation training (progressive muscle relaxation / diaphragmatic breathing / mindfulness): directly applies a 'brake' to the autonomic nervous system, lowering physiological arousal
The key mechanism: let go of the effort to 'must fall asleep,' arousal naturally drops, and sleep comes back — the cognitive-therapy counterpart of the hyperarousal model.
Sleep hygiene · necessary but not sufficient — it is the floor, not the ceiling
This is the most misunderstood piece. The public often treats 'sleep hygiene' as the treatment for insomnia, but the honest evidence is that sleep hygiene alone works weakly for chronic insomnia — in many CBT-I trials it is even used as the 'control / placebo' arm (Trauer 2015 does exactly this).
The hygiene checklist itself is not wrong (it removes the obvious obstacles to sleep):
Caffeine cutoff in early afternoon (half-life 5–6 h, dive `caffeine-l-theanine`)Less alcohol before bed (it dismantles REM and fragments the second half of the night, dive `sleep-architecture`)A cool, dark, quiet bedroom; control blue light for the hour before bedRegular exercise (but not too late)
Why 'hygiene only' isn't enough: it clears the precipitating / interfering factors, but the engine of chronic insomnia is hyperarousal + maintaining behaviors (conditioning, catastrophizing, burning time in bed). Hygiene never touches that engine.
The right framing: sleep hygiene is the floor (fill the obvious holes first), while CBT-I's stimulus control + sleep restriction + cognitive restructuring is the ceiling (actually dismantling the engine). Treating hygiene as the whole thing is the most common 'doing something correct but insufficient' in insomnia care.
Chapter 6
What to do · evidence, digital CBT-I, when to see a clinician + atlas loop
What to do · evidence, digital CBT-I, when to see a clinician + atlas loop
With the mechanisms covered, the last step is how hard the evidence is, how to actually access it, and when you must see someone.
Evidence · why CBT-I wins long-term over drugs (Trauer et al. 2015)
This *Annals of Internal Medicine* meta-analysis is the gold citation for CBT-I:
20 RCTs, 1162 adults with chronic insomniaPost-treatment: sleep-onset time ↓ ~19 minutes, wake after sleep onset (WASO) ↓ ~26 minutes, sleep efficiency +~10%Key: effects are maintained at 6 months after treatment ends — this is CBT-I's fundamental advantage over drugs
Why it wins long-term (at the mechanism level)
Sleep drugs (Z-drugs / benzodiazepines / antihistamines) deliver sedation and never touch hyperarousal or the maintaining behaviors; stop the drug and the engine is still running → rebound insomnia; long-term they also carry tolerance, dependence, falls/fractures in the elderly, and an association with dementia (drug details: dive `insomnia`)CBT-I dismantles the maintaining mechanism itself — so what it teaches you are portable, reusable skills that keep working after treatment endsACP 2016 therefore orders it: CBT-I first; drugs only as an add-on when CBT-I is insufficient, and short-term/intermittent (Qaseem et al.)
How to actually access it · digital CBT-I is the key to scale
CBT-I's biggest problem is not that it's poor — it's too few hands: trained therapists are far outnumbered by people with insomnia. The way out:
Digital / internet CBT-I (dCBT-I): self-guided app/web programs (e.g. Sleepio), supported by AASM 2021 and ACP as a way to expand access, with RCT effects close to in-personBrief Behavioral Treatment (BBT-I) (2–4 sessions, focused on stimulus control + sleep restriction): suited to primary-care / nurse deliveryA book + sleep diary to self-start, while following the mechanisms above
When to stop self-help and see a clinician (red flags → referral)
Loud snoring + witnessed apneic pauses + extreme daytime sleepiness → possible sleep apnea; CBT-I won't help, needs diagnosis + CPAP (dive `sleep-apnea`)Uncomfortable legs at night with an urge to move → restless legs syndromeInsomnia + persistent low mood / loss of interest, and especially any thoughts of self-harm → urgent psychiatric care / a crisis line; severe untreated depression should be treated first (same as the decision tree in `chronic-stress`)Sudden daytime sleep attacks, violent behaviors during sleep → narcolepsy / REM behavior disorder, needs a specialistBipolar disorder / uncontrolled epilepsy → sleep restriction can be harmful, must be under professional supervision
Atlas loop — the division of labor across the sleep cluster
`sleep-architecture` — the mechanistic spine: N3/REM, the two-process model (S×C); the underlying reason this island's sleep-pressure banking and clock-tightening work`insomnia` — classification and diagnosis + drug risks: 'what it is + why be careful with sleeping pills,' complementary to this island`chronic-stress` — the hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol. axis: the neuroendocrine end of hyperarousal; stress ↔ insomnia are mutually causal`caffeine-l-theanine` — how caffeine masks Process S sleep pressure (the mechanism behind the hygiene 'caffeine cutoff')`melatonin` — melatonin is a clock signal (Process C), not a sleeping pill; don't lump it with CBT-I as 'just take something'`shift-work-circadian` — the situation of chronically pulling S and C apart`all-nighter` — what a one-off all-nighter cuts is exactly deep sleep and REM`sleep-apnea` — the medical cause that needs CPAP, not CBT-I
How to navigate: chronic insomnia (≥ 3 months) → self-start with this island's stimulus control + sleep hygiene, find a dCBT-I app, and do sleep restriction ideally under therapist supervision; to understand the underlying mechanism return to `sleep-architecture`; for classification/drug risks go to `insomnia`; if any red flag above appears, stop self-help and seek care.
Bottom line: the first-line treatment for chronic insomnia is not a drug — it is CBT-I, which is causal (it dismantles hyperarousal), durable long-term, and very low-risk. Sleeping pills have their place — short-term, intermittent, supervised — but should not be the default starting point. Knowing this mechanism, you neither panic-reach for melatonin nor mistake 'ten rules of sleep hygiene' for the whole treatment. This site does not replace a physician: persistent severe insomnia, loud snoring plus extreme daytime sleepiness, abnormal behaviors during sleep, or any emotional crisis warrant medical evaluation.
Evidence · why CBT-I wins long-term over drugs (Trauer et al. 2015)
This *Annals of Internal Medicine* meta-analysis is the gold citation for CBT-I:
20 RCTs, 1162 adults with chronic insomniaPost-treatment: sleep-onset time ↓ ~19 minutes, wake after sleep onset (WASO) ↓ ~26 minutes, sleep efficiency +~10%Key: effects are maintained at 6 months after treatment ends — this is CBT-I's fundamental advantage over drugs
Why it wins long-term (at the mechanism level)
Sleep drugs (Z-drugs / benzodiazepines / antihistamines) deliver sedation and never touch hyperarousal or the maintaining behaviors; stop the drug and the engine is still running → rebound insomnia; long-term they also carry tolerance, dependence, falls/fractures in the elderly, and an association with dementia (drug details: dive `insomnia`)CBT-I dismantles the maintaining mechanism itself — so what it teaches you are portable, reusable skills that keep working after treatment endsACP 2016 therefore orders it: CBT-I first; drugs only as an add-on when CBT-I is insufficient, and short-term/intermittent (Qaseem et al.)
How to actually access it · digital CBT-I is the key to scale
CBT-I's biggest problem is not that it's poor — it's too few hands: trained therapists are far outnumbered by people with insomnia. The way out:
Digital / internet CBT-I (dCBT-I): self-guided app/web programs (e.g. Sleepio), supported by AASM 2021 and ACP as a way to expand access, with RCT effects close to in-personBrief Behavioral Treatment (BBT-I) (2–4 sessions, focused on stimulus control + sleep restriction): suited to primary-care / nurse deliveryA book + sleep diary to self-start, while following the mechanisms above
When to stop self-help and see a clinician (red flags → referral)
Loud snoring + witnessed apneic pauses + extreme daytime sleepiness → possible sleep apnea; CBT-I won't help, needs diagnosis + CPAP (dive `sleep-apnea`)Uncomfortable legs at night with an urge to move → restless legs syndromeInsomnia + persistent low mood / loss of interest, and especially any thoughts of self-harm → urgent psychiatric care / a crisis line; severe untreated depression should be treated first (same as the decision tree in `chronic-stress`)Sudden daytime sleep attacks, violent behaviors during sleep → narcolepsy / REM behavior disorder, needs a specialistBipolar disorder / uncontrolled epilepsy → sleep restriction can be harmful, must be under professional supervision
Atlas loop — the division of labor across the sleep cluster
`sleep-architecture` — the mechanistic spine: N3/REM, the two-process model (S×C); the underlying reason this island's sleep-pressure banking and clock-tightening work`insomnia` — classification and diagnosis + drug risks: 'what it is + why be careful with sleeping pills,' complementary to this island`chronic-stress` — the hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol. axis: the neuroendocrine end of hyperarousal; stress ↔ insomnia are mutually causal`caffeine-l-theanine` — how caffeine masks Process S sleep pressure (the mechanism behind the hygiene 'caffeine cutoff')`melatonin` — melatonin is a clock signal (Process C), not a sleeping pill; don't lump it with CBT-I as 'just take something'`shift-work-circadian` — the situation of chronically pulling S and C apart`all-nighter` — what a one-off all-nighter cuts is exactly deep sleep and REM`sleep-apnea` — the medical cause that needs CPAP, not CBT-I
How to navigate: chronic insomnia (≥ 3 months) → self-start with this island's stimulus control + sleep hygiene, find a dCBT-I app, and do sleep restriction ideally under therapist supervision; to understand the underlying mechanism return to `sleep-architecture`; for classification/drug risks go to `insomnia`; if any red flag above appears, stop self-help and seek care.
Bottom line: the first-line treatment for chronic insomnia is not a drug — it is CBT-I, which is causal (it dismantles hyperarousal), durable long-term, and very low-risk. Sleeping pills have their place — short-term, intermittent, supervised — but should not be the default starting point. Knowing this mechanism, you neither panic-reach for melatonin nor mistake 'ten rules of sleep hygiene' for the whole treatment. This site does not replace a physician: persistent severe insomnia, loud snoring plus extreme daytime sleepiness, abnormal behaviors during sleep, or any emotional crisis warrant medical evaluation.