Sleepless and Stuck in Fear: Insomnia Disrupts the Brain’s Ability to Let Go of Threat

For millions of people with insomnia, the nightly struggle to fall or stay asleep is only the most visible symptom of a deeper disturbance. A new study published in PLOS Mental Health suggests that insomnia disorder may also leave an imprint on the body’s most primitive threat machinery, blunting the ability to extinguish learned fears and to recall that extinction the next day. The research, led by Tianyu Hwang, Saman Khazaei, Edward F. Pace-Schott, Ryan Bottary, and Rose T. Faghih, took an unusual approach: rather than relying solely on behavioral reports or brain imaging, the team reconstructed the hidden dynamics of sympathetic arousal from a single, unobtrusive signal — the electrical conductance of the skin.

Skin conductance has long been a workhorse of psychophysiology. When the sympathetic branch of the autonomic nervous system fires, sweat glands in the palms and soles fill, and the skin becomes a slightly better conductor of electricity. These tiny fluctuations, measured in microsiemens, carry information about moment-to-moment changes in internal arousal. But the raw signal is a messy mixture: fast, discrete bursts of activity ride on top of a slowly drifting baseline known as the tonic component. Disentangling the two has historically been as much art as science, which is precisely where the new study makes its technical contribution.

The researchers recruited 48 participants, evenly split between 24 individuals meeting criteria for insomnia disorder and 24 good sleepers. Each participant completed a classic fear conditioning and extinction protocol, the experimental backbone of decades of research into anxiety and trauma. In such paradigms, one cue is repeatedly paired with an aversive stimulus until it elicits a conditioned fear response, while another cue remains safe. During extinction, the feared cue is presented again and again without any aversive consequence, allowing a new memory — that the cue no longer predicts harm — to form alongside the original fear memory. How well that new memory is learned, and how well it is recalled later, is considered a laboratory proxy for the brain’s capacity to regulate fear in daily life.

Throughout the procedure, the team recorded skin conductance and then applied a mathematical framework to estimate the underlying arousal state over time. The approach builds on prior work by Faghih and colleagues in which skin conductance measurements are deconvolved to extract the discrete activations of the autonomic nervous system — the individual sympathetic bursts — along with the tonic component of the signal. These extracted features then serve as observations feeding into a state-space model, a statistical architecture in which an unobservable internal state, here the person’s arousal level, evolves over time and generates the measurable signal. The model continuously updates its estimate of arousal as each new data point arrives, producing a dynamic trace of the nervous system’s internal state that no single raw measurement could provide.

The study’s central methodological question was how best to model the relationship between the observations and the hidden arousal state. The researchers compared two options. The first, a marked point process model, or MPP, used only the discrete autonomic activations extracted from the skin conductance signal, treating each sympathetic burst as an event whose timing and magnitude carry information about arousal. The second, called MPP+Cont, added the tonic component as a continuous observation channel, so that both the discrete bursts and the slow baseline drift informed the arousal estimate. The distinction may sound like a technical footnote, but it turned out to be decisive for the study’s findings — and, the authors argue, for how future studies of autonomic regulation should be designed.

From the estimated arousal trajectories, the team computed a high arousal index, or HAI, a summary measure capturing how much time a participant spent in elevated arousal states during each phase of the fear conditioning and extinction task. Comparing this index between the insomnia and good sleeper groups across conditioning, extinction learning, and extinction recall allowed the researchers to pinpoint exactly where the groups diverged. The answer was telling: individuals with insomnia disorder showed evidence of impaired fear extinction learning and impaired extinction recall, meaning they were slower to learn that the feared cue had become safe and less able to retain that safety learning afterward.

Perhaps the most striking result, however, was what the comparison of the two models revealed. When the analysis relied on the MPP model alone, using only the discrete autonomic bursts, the two groups were statistically indistinguishable. The signature of insomnia emerged only when the tonic component was incorporated through the MPP+Cont model. In other words, the slow, baseline level of sympathetic activity — the part of the signal most often discarded or averaged away in conventional analyses — carried the critical information separating insomniacs from good sleepers. This finding elevates the tonic component from a nuisance variable to a central player in arousal estimation, and it suggests that previous studies relying on event-based skin conductance measures alone may have missed meaningful group differences.

The results fit into a broader and increasingly influential framework in sleep and emotion research. Sleep, and rapid eye movement sleep in particular, has been proposed to serve as a window in which the brain consolidates extinction memories and strips learned fears of their emotional charge. Chronic sleep disruption, as seen in insomnia disorder, may interfere with this overnight processing, leaving fear memories intact and easily reactivated. The present study adds a physiological dimension to that account: the autonomic nervous system of people with insomnia appears to remain in a state of heightened and poorly regulated sympathetic arousal when confronted with threat-related cues, and that dysregulation persists into the recall of safety learning. Insomnia, on this view, is not merely a nighttime inconvenience but a condition of altered threat physiology around the clock.

The clinical implications reach well beyond sleep medicine. Impaired fear extinction is a hallmark of post-traumatic stress disorder, and insomnia is one of its most common comorbidities. If disrupted arousal regulation is a shared mechanism, then treatments that improve sleep — or that directly target autonomic dysregulation through approaches such as biofeedback, relaxation training, or pharmacological interventions — could strengthen exposure-based therapies, which depend on exactly the extinction processes this study measured. The authors suggest that arousal state estimation could serve as a sensitive tool for identifying which patients are most likely to struggle with fear-centered comorbidities, and for tracking physiological change over the course of treatment in a way that self-report questionnaires cannot.

There are, of course, limits to what a single laboratory study can establish. The sample was modest, with 24 participants per group, and the cross-sectional design cannot determine whether autonomic dysregulation is a cause of insomnia, a consequence of chronic sleep loss, or a shared vulnerability underlying both. The fear conditioning paradigm, while well validated, captures only a slice of emotional life. Yet the methodological advance may prove as important as the clinical finding. By demonstrating that a wearable-friendly signal — skin conductance — can be transformed into a continuous estimate of arousal state through carefully constructed observation models, the study points toward a future in which the invisible dynamics of the autonomic nervous system become as measurable and clinically informative as heart rate or brain waves. For the roughly one in ten adults who meet criteria for insomnia disorder, that future could bring a clearer physiological understanding of why sleepless nights so often come with anxious days, and better tools for breaking the cycle.

Subject of Research: Sympathetic arousal estimation from skin conductance reveals disrupted fear extinction in insomnia disorder

Article Title: Disrupted fear regulation in insomnia disorder revealed by arousal estimation from skin conductance

Article References: Hwang, T., Khazaei, S., Pace-Schott, E. F., Bottary, R., & Faghih, R. T. (2026). Disrupted fear regulation in insomnia disorder revealed by arousal estimation from skin conductance. PLOS Mental Health, 3(10), e0000720. https://doi.org/10.1371/journal.pmen.0000720

Image Credits: AI Generated

DOI: 10.1371/journal.pmen.0000720

Keywords: insomnia disorder, fear extinction, skin conductance, sympathetic arousal, autonomic nervous system, tonic component, marked point process, state-space model, high arousal index, fear conditioning, sleep, emotion regulation