Insomnia and Anxiety: The Bidirectional Trap

3:38 am. The thought arrived about twenty minutes ago: what if I never fix this. Not tonight’s sleep specifically, but all of it. What if I am someone who does not sleep, and the not-sleeping eventually dismantles the parts of my life that still function. The thought is detailed and persuasive and it has brought several supporting thoughts with it, each one more alarming than the last.

This is not just insomnia. This is insomnia and anxiety doing what they do best: feeding each other.

I have had nights where the insomnia came first and the anxiety followed, lying there in the quiet while the brain constructed elaborate disaster scenarios from the raw material of sleeplessness. And I have had nights where the anxiety arrived at bedtime like an uninvited guest, bringing with it a heart rate and a vigilance level that made sleep impossible. The order varies. The result is the same: a cycle that tightens with each repetition, and that neither condition on its own can fully explain.

The bidirectional mechanism

The relationship between insomnia and anxiety is not a one-way street. Alvaro et al.’s systematic review confirmed what most insomniacs intuit: insomnia increases the risk of developing anxiety, and anxiety increases the risk of developing insomnia [1]. The two conditions are bidirectionally linked, and each one makes the other harder to treat.

The mechanism works at the neural level. Goldstein et al.’s review of sleep and emotional brain function describes the pathway clearly [6]. The prefrontal cortex, the part of the brain responsible for regulating emotional responses, requires adequate sleep to function properly. When sleep is curtailed, prefrontal control over the amygdala (the threat-detection center) weakens. The result is increased emotional reactivity: the same stimulus that produces mild concern after a good night’s sleep produces disproportionate alarm after a bad one.

Ben Simon and Walker’s fMRI studies have shown this directly [2]. Sleep-deprived subjects showed amplified amygdala activation and reduced functional connectivity between the prefrontal cortex and the amygdala. The brain was literally less capable of regulating its own fear response. The same bad news that you could have absorbed at noon, after seven hours of sleep, becomes a crisis-level event at 3am after four hours of fragmented rest.

This is measurable. It is not a matter of being “too sensitive” or “overthinking.” The prefrontal cortex is offline to a degree that it would not be after adequate sleep, and the amygdala is compensating by running a threat assessment with the brakes off.

How anxiety maintains insomnia

Anxiety raises both cognitive and physiological arousal. The cognitive component is the racing thoughts, the what-ifs, the catastrophic predictions about tomorrow. The physiological component is elevated heart rate, muscle tension, and cortisol, which together produce a body state incompatible with sleep onset.

Harvey’s cognitive model of insomnia applies directly here [4]. Anxious individuals monitor their internal state more closely (“Am I relaxed enough to sleep?”), which paradoxically increases arousal. They engage in safety behaviors (going to bed early, lying very still, trying hard to clear their mind) that reinforce the bed-wakefulness association. And they catastrophize about the consequences of not sleeping, which generates more anxiety, which generates more arousal.

The monitoring is worth looking at closely because it is subtle. You lie down and begin scanning: is my heart rate slowing? Am I relaxed? Am I closer to sleep than I was five minutes ago? Each check is a small dose of cognitive activation. The scanning itself prevents the thing it is scanning for. It is like trying to fall asleep by closely watching yourself fall asleep, which is exactly the kind of paradox that insomnia specializes in.

The safety behaviors compound the problem. Going to bed early to “make sure” you get enough sleep extends time in bed, diluting sleep pressure. Lying very still so as not to “disrupt” the sleep process keeps you physically tense. Trying to suppress anxious thoughts, as anyone who has tried it knows, amplifies them. The effort to control the anxiety becomes its own source of arousal.

The cycle is self-tightening. Each bad night provides evidence for the anxious prediction that tonight will also be bad. Each anxious night confirms the insomniac’s belief that they cannot control their sleep. The evidence accumulates in one direction, which makes the predictions feel less like predictions and more like facts.

How insomnia maintains anxiety

The less obvious direction is equally important. Insomnia is not just a symptom of anxiety; it is a contributor to it. Chronic sleep deprivation degrades the neural circuits responsible for emotional regulation, making anxious responses more frequent and more intense [6].

This means that treating anxiety without addressing insomnia leaves a maintaining factor in place. The anxiety may improve with therapy or medication, but the impaired emotional regulation from ongoing sleep deprivation continues to prime the anxiety system. It is like treating a wound while the thing causing the wound is still present.

Harvey et al.’s review frames insomnia not just as a symptom of psychiatric disorders but as an independent risk factor for developing them [4]. People with untreated insomnia are at higher risk of developing new anxiety disorders, even if they did not have one before the insomnia started. The sleep problem is not merely reflecting the anxiety; it is generating additional anxiety through a physiological pathway that exists regardless of the psychological content.

This reframes the clinical question. The old model treated insomnia as secondary: fix the anxiety, and the sleep would follow. The evidence supports a different approach: insomnia is an independent condition that interacts with anxiety, and it needs its own treatment.

CBT-I works even with anxiety

One of the clearer findings in the comorbidity literature is that CBT-I does not require the absence of anxiety to be effective. Belleville et al. demonstrated that CBT-I reduces insomnia severity in patients with co-occurring anxiety disorders, and that the insomnia improvement was accompanied by reductions in anxiety symptoms [3]. Treating the sleep problem partially treated the anxiety problem, even without anxiety-specific intervention.

Espie et al.’s large effectiveness trial included patients with various psychiatric comorbidities and found that CBT-I produced significant improvements across the board [5]. Anxiety did not prevent the treatment from working. In many cases, improving sleep removed enough of the maintaining fuel that anxiety became more manageable.

The mechanism is straightforward: better sleep restores some prefrontal function, which improves emotional regulation, which reduces the intensity of anxious responses, which makes it easier to sleep the next night. The cycle starts turning in the other direction. Each good night provides a small counterexample to the catastrophic prediction, and each calmer day demonstrates that the anxiety was being amplified by sleep loss rather than reflecting reality.

This does not mean CBT-I replaces anxiety treatment. It means the two treatments are complementary, and starting with either one produces benefits for both. For people who have been told their insomnia is “just the anxiety” and that they should treat the anxiety first, the evidence suggests a different path: treat the insomnia directly, and some of the anxiety will come down with it.

What makes this hard

The hardest part of the insomnia-anxiety comorbidity is that the treatment for one temporarily worsens the other. Sleep restriction, the most potent component of CBT-I, involves spending less time in bed and experiencing more daytime sleepiness for the first one to two weeks. For someone with anxiety, the additional sleep deprivation can temporarily amplify anxious symptoms. The treatment is making things feel worse on the way to making them better.

This is manageable with awareness and, ideally, clinical support. Knowing that the first week of sleep restriction will feel harder for someone with anxiety is useful information. It does not mean the treatment is wrong. It means the treatment has a known cost, and the cost is temporary. Some clinicians adjust the pace of sleep restriction for anxious patients, reducing the window more gradually to keep the anxiety from spiking.

The other hard part is the 3am logic. At 3am, the anxious brain is running its threat assessment with reduced prefrontal oversight and elevated amygdala activation. Every thought feels urgent and true. The recognition that this is a physiological state, not an accurate evaluation of reality, is genuinely difficult to access in the moment. At 3am, your reasoning faculty is not at its best, and the thoughts it produces are disproportionately negative, and you cannot trust the feeling that says “but this time it really is as bad as it seems.”

I find it helps to name it during the day, when the prefrontal cortex is fully online. Write it down, if that helps: the thought at 3am is not your best thinking. It is your most aroused thinking. Those are different things, and remembering that difference, even vaguely, at 3am is sometimes enough to loosen the grip. Not to make the thought disappear. Just to create a small gap between the thought and the certainty that it is true. That gap is where the cycle can begin to break.

If your sleep is affecting daily life, talking to a GP or sleep specialist is a good first step. The American Academy of Sleep Medicine has a provider finder at sleepeducation.org. If insomnia is accompanied by depression, anxiety or other concerns, those are worth raising with your doctor as part of the same conversation.