Why do I feel anxious at night?

Nothing about your problems changes at 11pm. What changes is everything that was standing between you and them.

Through the day, worry competes with tasks, conversations, screens, and the sense that you could still act on whatever is bothering you. At night those competitors leave one by one. The list is closed, the people have gone, the room is dark, and there is nothing to be done until morning. Worry gets the room to itself.

There is a physical side too. The body is meant to wind down in the evening, with cortisol falling and the parasympathetic system taking over. Worry is a form of cognitive arousal, and arousal pushes the other way. Allison Harvey's cognitive model of insomnia describes the sequence: worry triggers arousal, arousal makes sleep harder, the difficulty becomes a new thing to worry about, and attention starts scanning for signs of not sleeping. Each loop tightens the next.

Anxiety at night is not only louder; it is less accurate. Sleep loss raises next-day anxiety by around 27 percent in Matthew Walker's lab studies, and it does so by disconnecting the prefrontal cortex from the amygdala, the part of the brain that flags threat. Tired brains rate ordinary problems as bigger and their own ability to cope as smaller. The worry that keeps you awake was produced by a brain already running short, so it is worth treating as a biased witness.

1. A day with unfinished business or a hard conversation.
2. Bedtime removes the distractions. Worry starts.
3. Worry raises arousal. Sleep is delayed or broken.
4. Short sleep makes the next day's worry bigger and harder to regulate.
5. Repeat, now with the added worry of "I won't sleep tonight either".

You can interrupt this at almost any step. The two below have the best evidence.

Thomas Borkovec's stimulus-control method for worry, from 1983 and replicated many times since, is simple: set a fixed 15 to 20 minute period earlier in the evening, and when a worry arrives outside it, note it and postpone it to the window. During the window, write the worries down and, for each one, either the next small step or "nothing to do until Tuesday". A 2023 meta-analysis found small but consistent reductions in worry frequency and duration, and a 2025 randomised trial found the same benefit extended to sleep.

The window works because it gives the worry an appointment. The brain stops raising it at midnight once it trusts it will be heard at eight.

The single most effective piece of insomnia treatment is called stimulus control: if you are awake in bed for what feels like 20 minutes, get up, do something dull in dim light, and go back when sleepy. It feels counterproductive and it is not. Lying awake and anxious in bed teaches the brain that bed is where anxiety happens. Getting up breaks that association. How to sleep when your brain won't switch off goes through the full routine.

• Caffeine after mid-afternoon. It keeps arousal up long after you have forgotten the cup. The caffeine cut-off.
• Alcohol to get off to sleep. It works for the first half of the night and rebounds in the second, often with a burst of anxiety. Alcohol, sleep and hangxiety.
• Checking the time. Every glance is a small arithmetic problem ("if I fall asleep now I get five hours") and a small spike.
• Problem-solving in the dark. Real decisions need a rested brain. Anything decided at 2am should be re-decided at 9.