Insomnia Treatment: What Actually Works (and What Doesn't)

The DSM-5 definition of insomnia disorder requires difficulty initiating sleep, difficulty maintaining sleep, or early morning awakening with inability to return to sleep — occurring at least three nights per week for at least three months, causing clinically significant daytime impairment (fatigue, cognitive difficulty, mood disturbance, impaired performance), despite adequate opportunity for sleep.

Acute vs. chronic. Acute insomnia — lasting days to a few weeks, typically triggered by a stressor — is extremely common and usually self-resolves. Chronic insomnia (three months or more) has a different character: it often persists well beyond the original stressor because of the behavioral and cognitive patterns that develop around it. The bed becomes associated with wakefulness and anxiety; the effort to sleep itself creates arousal that prevents it.

The hyperarousal model is the most useful framework for understanding chronic insomnia. Insomnia is not primarily a problem of insufficient tiredness — most people with insomnia are exhausted. It is a problem of hyperarousal: the brain is stuck in a threat-detection mode that keeps it alert and activated at the exact time it needs to wind down. Physiological arousal (elevated cortisol, higher body temperature), cognitive arousal (racing thoughts, worry, planning), and emotional arousal (anxiety about sleep itself) all feed this loop. Sleep medication suppresses symptoms without addressing the underlying arousal — which is why behavior-based treatment outperforms medication in the long run.

Sleep and mental health are deeply entangled, and the relationship runs in both directions. Insomnia is simultaneously a symptom of most major psychiatric conditions and a driver that worsens them. Getting the direction right matters for treatment.

Depression has a characteristic sleep signature: early morning awakening, waking at 3–4am and being unable to return to sleep. This is distinct from the sleep-onset insomnia more typical of anxiety. Depression also disrupts sleep architecture — reducing restorative slow-wave sleep and dysregulating REM. When depression is the primary driver, treating the depression typically improves sleep; but insomnia can also persist independently and may need its own treatment.

Anxiety produces sleep-onset insomnia — difficulty falling asleep because the brain is activated with racing thoughts, worry, and planning. The bed becomes associated with the mental experience of lying awake and ruminating, which creates a conditioned arousal response that persists even after the acute anxiety is addressed.

Bipolar disorder: Reduced need for sleep — feeling rested after only 2–4 hours — is a warning sign of hypomania or mania, not insomnia. This is a critical distinction. A person in a hypomanic episode who reports sleeping 3 hours a night and feeling great needs mood stabilization, not sleep medication.

ADHD involves delayed sleep phase — the internal clock runs late, making it genuinely difficult to fall asleep before midnight even when the person wants to. Racing thoughts at night, often described as the quietest part of the day being when the brain finally “turns on,” is a common ADHD sleep complaint distinct from anxiety-driven insomnia.

The bidirectionality matters: poor sleep is not just a symptom. Sleep deprivation impairs emotional regulation, amplifies amygdala reactivity, worsens working memory and executive function, and lowers the threshold for mood episodes. Treating insomnia in someone with depression or anxiety often produces downstream improvement in the psychiatric condition — sometimes more than treating the psychiatric condition alone.

Cognitive Behavioral Therapy for Insomnia (CBT-I) is the first-line treatment for chronic insomnia according to every major clinical guideline — more effective than sleep medication in head-to-head trials, and with effects that last after treatment ends (unlike medication, whose effects stop when you stop taking it). Multiple meta-analyses confirm this. Most people see significant improvement in 6–8 weeks.

CBT-I has several components, and all of them matter:

Sleep restriction therapy

The most powerful and most counterintuitive component. Sleep restriction consolidates fragmented sleep by temporarily limiting time in bed to your actual sleep time — building sleep pressure (adenosine accumulation) and reconditioning the bed-sleep association. It feels terrible for the first week (that's the point) and produces dramatic improvements in sleep quality and efficiency within 2–3 weeks. Most people resist this component and most people are glad they did it.

Stimulus control

The goal is to make the bed a strong cue for sleep rather than for wakefulness, worry, and screen time. Rules: use the bed only for sleep (and sex); if you're awake for more than 20 minutes, get up and return when sleepy. This feels wrong — lying in bed trying feels productive. It isn't. Extended wakefulness in bed strengthens the bed-wakefulness association that maintains insomnia.

Cognitive restructuring

Targeting the catastrophic beliefs about sleep that amplify arousal: “I can't function on less than 8 hours,” “I'll have a terrible day if I don't sleep,” “I haven't slept well in years and I never will.” These beliefs create performance anxiety around sleep, which produces exactly the arousal that prevents sleep.

Relaxation training

Progressive muscle relaxation, diaphragmatic breathing, and imagery techniques target the physiological arousal component of insomnia. These are skills that require practice — they don't work the first time.

CBT-I is available through therapists trained in it, but also digitally: Sleepio is an evidence-based digital CBT-I program, and Somryst is an FDA-cleared prescription digital therapeutic for chronic insomnia. Both deliver CBT-I effectively in a self-guided digital format.

When medication is appropriate — as a bridge, in combination with CBT-I, or when CBT-I isn't accessible — it helps to understand what each option actually does and where it fits.

Melatonin

Melatonin signals the brain that it's time to sleep — it is a circadian rhythm regulator, not a sedative. It works well for circadian rhythm issues (jet lag, shift work, delayed sleep phase syndrome) and for sleep-onset problems when circadian timing is the issue. It is weak for sleep maintenance insomnia (waking in the middle of the night) and for the hyperarousal-driven insomnia common in anxiety. Dose matters: 0.5–1 mg is physiological; most OTC products are 5–10 mg (supraphysiological) and can cause next-day grogginess.

OTC antihistamines (diphenhydramine / Benadryl)

Diphenhydramine is the active ingredient in most OTC sleep aids (ZzzQuil, Unisom, Tylenol PM). Tolerance develops within days — by the fourth or fifth night, the sedating effect is substantially reduced. Next-day sedation (“sleep hangover”) impairs cognitive performance and driving. There is emerging evidence of cumulative anticholinergic effects with long-term use. Not recommended as a regular sleep aid.

Trazodone (low dose, off-label)

Low-dose trazodone (25–100 mg) is the most commonly prescribed off-label sleep medication in the US. At low doses, it acts primarily as an H1 antihistamine and alpha-1 antagonist — producing sedation without significant serotonin reuptake inhibition at the sleep doses used. It has no dependence liability, is not a controlled substance, and does not cause tolerance in the same way as Z-drugs. Side effects include next-day grogginess (especially at higher doses) and, rarely, priapism (prolonged erection — a medical emergency; patients should be counseled about this).

Doxepin low-dose (Silenor)

Low-dose doxepin (3–6 mg) is FDA-approved specifically for insomnia characterized by difficulty maintaining sleep. At these doses, it acts as a selective H1 antagonist — blocking histamine to promote sleep maintenance — without the anticholinergic effects of higher antidepressant doses. It has no dependence liability and minimal next-day sedation at approved doses.

Orexin antagonists: suvorexant (Belsomra) and lemborexant (Dayvigo)

This newer class works by blocking orexin (hypocretin) — the neurotransmitter that promotes wakefulness — rather than inducing sedation. They are FDA-approved for both sleep onset and sleep maintenance insomnia. They carry lower dependence risk than Z-drugs, have a mechanism that more closely mimics natural sleep onset, and have a favorable long-term profile compared to older sleep medications. They are more expensive (not always well covered by insurance) and can cause vivid dreams or next-day drowsiness.

Z-drugs: zolpidem (Ambien), eszopiclone (Lunesta)

Z-drugs act on GABA-A receptors (similar to benzodiazepines) and are effective for sleep onset. They are approved for short-term use. Long-term use is associated with tolerance, dependence, rebound insomnia on discontinuation, next-day cognitive and psychomotor impairment (driving is genuinely impaired — the FDA lowered recommended doses for this reason), and complex sleep behaviors (sleepwalking, sleep eating, driving while asleep). They remain widely prescribed but are generally not the best long-term choice.

Benzodiazepines for insomnia

Benzodiazepines are not recommended as a long-term treatment for insomnia. They suppress deep sleep and REM, cause tolerance and dependence, produce rebound insomnia on discontinuation, and carry cognitive and fall risks (especially in older adults). Short-term use as a bridge in specific circumstances is sometimes appropriate with a clear taper plan.

Some of the most common insomnia remedies actively make the problem worse. Understanding why is useful both for stopping them and for explaining CBT-I's apparently counterintuitive approach.

Alcohol

Alcohol produces sedation that helps with sleep onset — which is why so many people use it as a sleep aid. But it reduces REM sleep in the first half of the night and produces rebound arousal in the second half (waking up at 3am and being unable to return to sleep is often alcohol, not just insomnia). It also worsens sleep apnea, which amplifies the next-day fatigue. Regular alcohol use to facilitate sleep reliably worsens overall sleep quality.

Sleeping in late to “catch up”

Sleeping in after a bad night feels like compensation but actually disrupts the circadian rhythm and reduces sleep pressure (adenosine buildup) for the following night — making it harder to fall asleep the next evening and perpetuating the cycle. Keeping a consistent wake time, even after a bad night, is one of the most powerful things a person can do for chronic insomnia. It is also one of the hardest.

Lying in bed awake for hours

Extended wakefulness in bed strengthens the learned association between the bed and being awake and anxious. The bed becomes a cue for hyperarousal rather than sleep. This is the mechanism CBT-I's stimulus control targets directly: get out of bed if you are not sleeping. It feels counterproductive. It isn't.

Phone in bed

Blue light suppresses melatonin secretion and delays sleep onset — that part is real but overstated. The larger issue is cognitive and emotional: engaging with email, news, or social media keeps the brain in an alert, active state that is the opposite of what sleep requires. A phone in bed on “night mode” is still a phone in bed.

When talking to a clinician about insomnia, being specific about the pattern helps guide treatment. Sleep-onset insomnia (can't fall asleep) points toward different causes and treatments than sleep maintenance insomnia (waking in the middle of the night) or early morning awakening. Mention how long the problem has been occurring, what you've already tried, and what psychiatric conditions or medications might be relevant.

Useful questions to ask your prescriber:

• “Do you think the insomnia is driving my anxiety/depression, or the other way around — and does that change the treatment approach?”
• “Can you refer me for CBT-I, or recommend a digital program like Sleepio?”
• “If you're recommending medication, is it for short-term use or longer, and what does stopping look like?”
• “Given what I'm taking for [depression/anxiety/ADHD], is there a medication option that makes sense alongside it?”
• “Is sleep apnea worth ruling out — and if so, what would that involve?”