Worry gets louder at 3 a.m., and a few tested tricks can turn it down

Key takeaways

• In a University of California, Berkeley study, one sleepless night raised anxiety by up to 30%, and a full night of sleep brought it down, most in people who got more deep sleep¹.
• In a daily diary study of 50 heavy worriers, a day of worry predicted a worse night, but a bad night didn’t predict more worry the next day³.
• Putting off worry until a set 30-minute window cut how long and how often people worried in a 2024 review of eight samples, though the effects were small⁵.
• People with insomnia who watched a clock while trying to fall asleep worried more and overestimated how long they lay awake⁷.
• Waking in a panic is common in panic disorder, and studies suggest cognitive behavioral therapy helps, but crushing chest pain or pain spreading to your arm or jaw means call 911⁸¹⁰.

Eighteen young adults gave up a night’s sleep for science at the University of California, Berkeley¹. Each was tested twice: once after a normal night in bed, and once after staying up all night¹.

After each night, the team scanned their brains while they watched emotional video clips, and they filled out a standard anxiety questionnaire¹.

The sleepless night pushed anxiety scores up by as much as 30%¹. After it, half of the volunteers landed above the cutoff researchers use for a clinical anxiety disorder².

The scans hinted at why. The medial prefrontal cortex, an area just behind the forehead that acts as a brake on anxiety, had gone quiet. Deeper emotional regions were overactive¹.

Then came the hopeful part. A full night brought anxiety down, and the people who spent the most time in deep, slow-wave sleep calmed the most¹. A second group of 30 volunteers showed the same pattern¹.

That study, published in late 2019 in Nature Human Behaviour, was led by Eti Ben Simon, with sleep scientist Matthew Walker as senior author¹. It was small, and it tested a full night without sleep, not the ordinary rough night most of us know. But it shows how tightly sleep and worry are tied, and why the loop can feel impossible to escape at 3 a.m.

This article covers what tested techniques can do, how night panic differs from a bad dream, and a plan for the next time you wake up worried.

Worry and lost sleep feed each other, though not always equally

The researchers then looked beyond the lab. In an online study, 280 people of all ages reported on their sleep and anxiety for four days in a row. How long and how well someone slept on a given night predicted how anxious they felt the next day¹.

Other research points the arrow the other way. In 2016, Sarah Kate McGowan, Evelyn Behar and Maike Luhmann had 50 heavy worriers keep daily records for three weeks³.

On days when people worried more, they slept worse that night. But a bad night didn’t predict more worry the next day³. For these chronic worriers, daytime worry seemed to drive the sleep trouble, not the reverse.

The findings aren’t as contradictory as they look. One tested total sleep loss in young adults; the other tracked ordinary nights in chronic worriers. Either way, the practical lesson is the same: deal with worry before bed, and protect your sleep.

Why does worry feel louder at night? Part of the answer may be simple: in the dark and quiet, little competes for your attention. And as the Berkeley scans suggest, a short-slept brain has weaker brakes on emotion¹.

Give your worry an appointment, well before bed

Worry postponement sounds too simple to work. You jot down each worry as it pops up during the day, then put it off until a set “worry window,” usually 30 minutes in the evening⁵.

In 2024, researchers Dippel, Brosschot and Verkuil pooled the trials in the International Journal of Cognitive Therapy. They found seven studies in six papers, with eight groups and 999 people, mostly women and students⁵.

Postponing worry shortened how long people worried, and to a lesser degree how often, compared with control groups⁵. Both effects were small, by the usual statistical yardstick⁵.

The limits are real. Worry was self-reported, some studies raised risk-of-bias concerns, and none followed people afterward to see if the benefit lasted⁵.

One early test came from Brosschot and van der Doef in 2006. Among 171 teenagers, six days of postponing worry to a 30-minute evening window led to shorter worry spells and fewer health complaints⁵.

What about sleep? In a master’s thesis study at the University of Illinois Chicago, 46 people were randomly assigned to two weeks of a daily 30-minute worry period, held at a set time and place⁴. The comparison group was taught to accept worry rather than push it away⁴. The worry-period group improved more on worry, anxiety and insomnia symptoms⁴.

A 2013 paper by McGowan and Behar also found that worry postponement cut worry and anxiety more than a comparison condition⁵.

The studies used evening windows. For sleep, we suggest early evening, well before bed, so the worry session doesn’t become your bedtime routine. That’s our advice, not something the trials tested.

Write down the worry, then write the next step

Constructive worry adds one move: for each worry, you write a possible solution or next step. The idea is to finish the thinking at a desk, not in bed.

Colleen Carney tested it for her 2003 doctoral research at Louisiana State University⁶. She assigned 33 college students with insomnia to five nights of writing down worries and solutions, or writing down worries and filling out worry questionnaires⁶.

The solution writers reported less mental arousal before sleep, meaning fewer racing thoughts, compared with the other group and with their own starting point⁶.

This is the honest catch. Their sleep diaries and wrist sleep monitors showed no significant change, and their anxiety didn’t change either⁶. So constructive worry may quiet your mind at bedtime without, in this small study, making you sleep longer.

A calmer mind at lights-out is still worth having. Our evening routine article covers a related trick, a bedtime to-do list, and our upcoming journaling article goes deeper.

Turn the clock around

Few habits feel as natural as checking the time at night. In 2007, researchers Tang, Schmidt and Harvey tested what clock-watching does⁷.

In the sharper of their two experiments, 38 people with diagnosed insomnia tried to fall asleep facing either a real digital clock or a look-alike display showing random numbers⁷. The fake display meant both groups had something to watch.

Worry scores in the real-clock group nearly doubled, from 6.7 to 13. In the random-number group, they fell, from 12.3 to 7.7⁷. The clock-watchers also thought falling asleep took longer than before, about 37 minutes instead of 27⁷.

Their wrist monitors told a different story: about 11 to 13 minutes in both groups, with no real difference⁷. Clock-watchers overestimated how long they lay awake by about 24 minutes, versus about 11 for the comparison group⁷.

The other experiment asked half of 60 good and poor sleepers to keep an eye on a clock. Those clock-watchers worried more and took longer to fall asleep, both by their own reports and by their monitors⁷.

The studies were small and measured sleep with wrist monitors, not full sleep-lab recordings⁷. The Wellcome Trust, a British research charity, funded the work⁷. The fix costs nothing: turn the clock to the wall and put your phone across the room.

Waking in a panic is different from a bad dream

Some people don’t drift up into worry. They jolt awake already in full alarm, with symptoms much like a daytime panic attack⁹.

That’s nocturnal panic: waking from sleep already in a state of panic⁸. It starts in non-REM sleep, not the dream-heavy REM stage. Panic researcher Michelle Craske and a colleague describe it as distinct from nightmares, sleep terrors and sleep apnea⁸.

It isn’t rare among people with panic disorder. In clinic surveys, 44% to 71% had had at least one, and 30% to 45% had repeated episodes⁸⁹. Panic disorder itself affects about 2.7% of U.S. adults in a given year⁹.

Craske and Joanna Arch describe a cycle. People who panic at night may come to dread sleep and try to put it off, and the lost sleep may then set off more attacks⁹. One model of panic holds that small body changes a person never consciously notices can trigger fear once they’ve been linked to panic⁹.

The good news: studies suggest cognitive behavioral therapy helps nocturnal panic. It targets the habit of reading body sensations as danger, the urge to overbreathe and learned fear of internal cues⁸⁹. A 2005 review found no controlled studies of medicines aimed at night panic specifically⁸.

Nightmares are a separate problem. When bad dreams keep coming back and cause real distress or get in the way of daily life, the American Academy of Sleep Medicine (AASM) calls it nightmare disorder. It affects about 4% of adults¹¹.

In a 2018 position paper, the AASM’s one full recommendation went to image rehearsal therapy, for nightmare disorder and for nightmares tied to PTSD¹¹. You rewrite the bad dream with new, more positive images, then practice the new version while awake¹¹. The medicine prazosin “may be used,” and the AASM advised against clonazepam and venlafaxine for nightmare disorder¹¹. Our upcoming article on when to seek a sleep evaluation covers nightmares and trauma in more depth.

When worry becomes an anxiety disorder

Ordinary worry comes and goes. An anxiety disorder, as the National Institute of Mental Health (NIMH) describes it, is anxiety that doesn’t go away, shows up across many situations and can get worse over time¹⁴.

About a third of U.S. adolescents and adults have an anxiety disorder at some point¹⁴. In national survey data from 2001 to 2003, generalized anxiety disorder (GAD) affected 2.7% of adults in a given year, 3.4% of women and 1.9% of men¹². About a third of those adults had serious impairment¹².

Trouble falling or staying asleep is on NIMH’s list of GAD symptoms¹³. Doctors diagnose it when hard-to-control worry shows up most days for at least six months, along with at least three of these: restlessness, fatigue, trouble concentrating, irritability, muscle tension or sleep problems¹³.

NIMH calls cognitive behavioral therapy (CBT) the gold standard psychotherapy for GAD¹³. Medicines such as SSRI and SNRI antidepressants can help but take weeks to work¹³. Benzodiazepines act fast, but tolerance and dependence can develop, so providers prescribe them only for short periods¹³.

Healthy habits count, too. Cutting caffeine, getting enough sleep, exercise and mindfulness can ease symptoms, but NIMH says they can’t replace treatment¹³.

If worry turns into hopelessness or thoughts of harming yourself, call or text the 988 Suicide & Crisis Lifeline at 988, any time¹⁴. In a life-threatening emergency, call 911¹⁴.

Your 3 a.m. plan and a 15-minute worry worksheet

These tools fit what many natural-health readers value: no pills, no cost and your own attention as the main instrument. They’re our design, built from the studies above, and they haven’t been tested as a package.

The 3 a.m. plan

1. Don’t check the time. Turn the clock to the wall before bed. Clock-watching raised worry in the lab⁷.
2. Check your body first. Chest pain or pressure, pain spreading to your arm or jaw, or trouble breathing means call 911¹⁰.
3. Park the worry. Write it in a few words on a bedside notepad, with “worry time, 6 p.m.” next to it. It’s the same postponing move the trials used by day⁵.
4. Settle your attention. Try slow breathing or muscle relaxation; MedlinePlus lists relaxation among ways to ease anxiety about sleep¹⁵. Our upcoming relaxation article gives scripts.
5. Get up if you’re wound up. If you’ve been awake a long while and feel tense, go somewhere dim, do something calm and dull, and return when you’re sleepy. Our upcoming insomnia article explains why.
6. Leave the phone alone. Our screens article explains the cost of late-night scrolling.
7. Keep your usual wake time, even after a bad night. Our article on sleep schedules explains why.

The 15-minute worry worksheet (early evening, at a table, not in bed)

Worry, in a few words	Can I act on it?	Smallest next step	When
Example: unpaid bill	Yes	Log in and pay it	Tomorrow, 9 a.m.
Example: test results	Not yet	Write down questions for the doctor	At the visit
Example: “What if I lose my job?”	Not tonight	Note one fact for and one against	Next worry time

• Minutes 1 to 5: read the day’s notes and cross off anything that no longer matters.
• Minutes 6 to 12: fill in the table, writing a next step for each worry, the move Carney tested⁶. If a worry has no next step, write “not tonight” and let it wait.
• Minutes 13 to 15: close the notebook and leave it outside the bedroom.

The trials used 30-minute windows⁴⁵. Start with 15 and lengthen it if worries pile up.

What the evidence says, rated

Technique or claim	Evidence level	What studies found
Sleep loss raises next-day anxiety	Promising but early	Small lab studies plus a four-day online study1
Worry postponement cuts daily worry	Promising but early	Small effects across eight samples; no follow-up5
Worry postponement eases insomnia	Promising but early	One small trial found fewer insomnia symptoms4
Constructive worry calms the mind at bedtime	Promising but early	Less racing thought in 33 students; no measured sleep change6
Hiding the clock reduces bedtime worry	Promising but early	Two small experiments; mostly self-reported effects7
CBT treats generalized anxiety disorder	Strong evidence	Called the gold standard therapy by NIMH13
CBT for nocturnal panic	Promising but early	Reviews from one research group report benefit89
Image rehearsal therapy for nightmares	Strong evidence	The one therapy the AASM fully recommended in 201811
Clonazepam for nightmare disorder	Evidence says no, or unsafe	Listed by the AASM as not recommended11

When to get help

Talk with a health care provider or therapist if any of these apply:

• anxiety causes problems at work, school or in your relationships¹³;
• hard-to-control worry happens most days for six months or more¹³;
• falling asleep often takes you longer than half an hour, or you wake several times and can’t drift off again¹⁵;
• you often feel sleepy during the day¹⁵;
• you snore loudly, snort, gasp or stop breathing for short spells¹⁵;
• you wake in a panic again and again⁸⁹;
• nightmares keep coming back and cause you distress¹¹.

For chest pain or trouble breathing, call 911¹⁰. For thoughts of suicide or a mental health crisis, call or text 988¹⁴.

Questions readers ask

The bottom line

Night worry isn’t a character flaw. A tired brain has weaker brakes, the dark offers few distractions, and a day of worry tends to spill into the night. Small, free tools can help: give worry an early-evening appointment, write down the next step, hide the clock and keep your wake time steady.

These techniques have early, modest evidence, so judge them by your own nights. When worry runs your days, CBT has strong support. Waking in a panic is treatable, and sudden chest pain or pressure is a reason to call 911, not to wait it out.

Related reading

• A warm shower and a to-do list do more for sleep than a nightcap
• Blue light gets the blame, but time on screens may cost more sleep
• Moving can ease anxiety, though viral shaking and somatic workouts are less proven
• The symptoms that mean call 911 now
• How to tell psychiatrists, psychologists and therapists apart

Sources

1. Anwar Y. “Deep sleep can rewire the anxious brain” (UC Berkeley news release on Ben Simon E, et al., senior author Walker MP, “Overanxious and underslept,” Nature Human Behaviour, 2019, DOI 10.1038/s41562-019-0754-8). Republished by Futurity, November 5, 2019. futurity.org
2. Genetic Engineering & Biotechnology News. “Overstressed and Underslept: Study Shows How Deep Sleep Can Reset Anxious Brain.” November 2019. genengnews.com
3. McGowan SK, Behar E, Luhmann M. “Examining the Relationship Between Worry and Sleep: A Daily Process Approach.” Behavior Therapy, 2016;47(4):460–474, doi 10.1016/j.beth.2015.12.003 (University of Cologne repository record). doi.org
4. McGowan SK. “The Efficacy of Stimulus Control for Worry.” Master’s thesis, University of Illinois at Chicago, 2012 (advisor Evelyn Behar). indigo.uic.edu
5. Dippel A, Brosschot JF, Verkuil B. “Effects of Worry Postponement on Daily Worry: A Meta-Analysis.” International Journal of Cognitive Therapy, 2024;17:160–178 (Leiden University open-access copy). scholarlypublications.universiteitleiden.nl
6. Carney CE. “The Effects of an Evening Structured Problem-Solving Procedure in Undergraduate College Students With Insomnia.” Doctoral dissertation, Louisiana State University, 2003. repository.lsu.edu
7. Tang NKY, Schmidt DA, Harvey AG. “Sleeping With the Enemy: Clock Monitoring in the Maintenance of Insomnia.” Journal of Behavior Therapy and Experimental Psychiatry, 2007;38:40–55 (University of Pennsylvania copy). med.upenn.edu
8. Craske MG, Tsao JCI. “Assessment and Treatment of Nocturnal Panic Attacks.” Sleep Medicine Reviews, 2005;9(3):173–184, doi 10.1016/j.smrv.2004.11.003 (EM-Consulte abstract). em-consulte.com
9. Arch JJ, Craske MG. “Panic Disorder.” In Craighead WE, Miklowitz DJ, Craighead LW, eds., Psychopathology: History, Diagnosis, and Empirical Foundations. Wiley (2008, per the University of Colorado file name). colorado.edu
10. MedlinePlus. “Chest Pain.” Medical Encyclopedia, National Library of Medicine, reviewed April 28, 2026. medlineplus.gov
11. American Academy of Sleep Medicine. “New Position Paper Recommends Treatment Options for Nightmare Disorder in Adults.” News release, June 14, 2018 (on Morgenthaler TI, et al., Journal of Clinical Sleep Medicine, 2018). aasm.org
12. National Institute of Mental Health. “Generalized Anxiety Disorder” (statistics). NIH, accessed October 2026. nimh.nih.gov
13. National Institute of Mental Health. “Generalized Anxiety Disorder: What You Need to Know.” NIH, revised 2025. nimh.nih.gov
14. National Institute of Mental Health. “Anxiety Disorders.” NIH, last reviewed December 2024. nimh.nih.gov
15. MedlinePlus. “Sleep Disorders.” National Library of Medicine, updated December 4, 2025. medlineplus.gov