Ketamine Lifts Depression, Anxiety and Insomnia in Resistant Cases, but Sleep Improves on Its Own Track

Ketamine has stunned psychiatry before, but a new analysis from researchers at the National Institute of Mental Health and their collaborators suggests that the drug’s dramatic effects on mood, anxiety and sleep may be less entangled than clinicians have long assumed. In a study of seventy-one patients with treatment-resistant depression, a single intravenous infusion of ketamine at the widely used dose of 0.5 milligrams per kilogram of body weight produced significant reductions in all three symptom domains within twenty-four hours. Yet when the team examined how the changes in each domain related to one another, a striking pattern emerged: improvements in anxiety moved in lockstep with improvements in depression, while changes in insomnia appeared to travel on an independent track, statistically disconnected from either.

The findings, published in Translational Psychiatry, come from a clinical trial registered as NCT00088699 and conducted under the Intramural Research Program of the National Institutes of Health. Treatment-resistant depression, typically defined as depression that fails to respond adequately to two or more adequate antidepressant trials, affects a substantial minority of the millions of people living with major depressive disorder worldwide. For these patients, insomnia and anxiety are not incidental companions but core burdens that shape daily functioning, suicide risk and quality of life. Understanding whether ketamine relieves these burdens together or separately is therefore far more than an academic exercise; it determines how clinicians should monitor, counsel and follow up with patients in the critical first day after infusion.

The study’s design was straightforward but rigorous. Each participant was assessed at baseline, immediately before receiving the infusion, and again twenty-four hours afterward. Depression severity was quantified with the Montgomery–Åsberg Depression Rating Scale, a clinician-administered instrument widely regarded as sensitive to rapid change. Anxiety was measured with the Hamilton Anxiety Rating Scale, and insomnia was captured using the insomnia items embedded in the Hamilton Depression Rating Scale. To analyze the data, the researchers employed linear mixed-effects models, a statistical framework that can account for repeated measurements within the same individual while estimating the average pre-to-post change across the group. The results were unambiguous: depression scores fell by an estimated 11.72 points on the MADRS, anxiety scores dropped by 7.87 points on the HAM-A, and insomnia scores declined by 0.89 points, with all three changes reaching statistical significance at P less than 0.001.

Those headline numbers confirm what a growing body of literature has suggested since the early 2000s, when researchers first reported that a subanesthetic dose of ketamine could lift depressive symptoms within hours in patients who had exhausted conventional options. Ketamine acts primarily as an antagonist of the N-methyl-D-aspartate receptor, a glutamate-gated ion channel, and its rapid antidepressant effects are thought to be mediated downstream through activation of AMPA receptors, stimulation of brain-derived neurotrophic factor signaling, and the restoration of synaptic connections eroded by chronic stress. The metabolite (2R,6R)-hydroxynorketamine has also been implicated in the antidepressant response, a discovery that has fueled an entire pipeline of ketamine-derived therapeutics now in clinical development.

But the real novelty of the new study lies in its dissection of how the three symptom domains interact. At baseline, before any drug was given, insomnia severity showed a moderate correlation with anxiety severity, with a Spearman rank correlation coefficient of 0.36, and anxiety correlated with depression with a Pearson correlation coefficient of 0.39. After ketamine, the association between anxiety and depression actually strengthened, rising to a Pearson correlation of 0.57, while the insomnia-anxiety link held steady at a Spearman coefficient of 0.36. Most tellingly, when the researchers computed change scores for each domain and correlated them, changes in anxiety were strongly associated with changes in depression, with a Spearman coefficient of 0.61, whereas changes in insomnia were not significantly associated with changes in either of the other two domains.

This dissociation carries real clinical weight. If insomnia improvement were simply a downstream consequence of mood lifting, one would expect patients whose depression improved most to also sleep most dramatically better. The data say otherwise. Instead, the study found that the single best predictor of how much a patient’s insomnia improved was how severe that insomnia was at baseline: greater baseline insomnia severity was associated with larger absolute reductions in insomnia scores. In other words, ketamine appears to act on sleep complaints in proportion to how disturbed sleep was to begin with, rather than as a ripple effect of the drug’s antidepressant action. This pattern hints at partially separable neurobiological mechanisms, perhaps involving ketamine’s effects on arousal circuitry, glutamatergic signaling in sleep-regulating regions, or the drug’s known sedative and psychotomimetic properties during the acute infusion period.

Supporting this interpretation, exploratory polysomnography analyses conducted in a subset of participants suggested objective sleep changes in the twenty-four hours after infusion: total sleep time increased, slow-wave sleep, the deep restorative stage most associated with synaptic homeostasis, increased, sleep efficiency improved, and sleep latency, the time it takes to fall asleep, shortened. Polysomnography, the gold standard for sleep measurement, records brain waves, eye movements, muscle tone and breathing overnight, providing an objective counterpoint to questionnaire-based insomnia ratings. The concordance between subjective insomnia improvement and objective sleep architecture changes, while exploratory in this dataset, strengthens the case that ketamine genuinely reshapes sleep physiology rather than merely altering how patients report their nights.

The research team, which included investigators from the NIMH’s Experimental Therapeutics and Pathophysiology Branch, Jena University Hospital, Philipps University of Marburg, the National Institute on Alcohol Abuse and Alcoholism, the University of Wisconsin-Madison and the University of British Columbia, went one step further by attempting to classify which patients would experience insomnia improvement. Using linear discriminant analysis and logistic regression as exploratory tools, they sought statistical signatures that might separate improvers from non-improvers. Such classification approaches, if validated in larger samples, could eventually help clinicians identify before an infusion which patients are likely to gain sleep benefits, allowing targeted adjunctive treatments, such as cognitive behavioral therapy for insomnia, for those less likely to respond on the sleep front.

The study is not without limitations, and the authors are careful to frame the insomnia findings appropriately. The insomnia measure was derived from items within the Hamilton Depression Rating Scale rather than a dedicated insomnia instrument, which may limit sensitivity to the full breadth of sleep complaints. The twenty-four-hour window captures only the earliest phase of ketamine’s action, and the drug’s antidepressant effects typically wane over days to two weeks, raising open questions about whether the dissociation between sleep and mood trajectories persists or collapses over longer follow-up. The sample of seventy-one patients, while respectable for an intramural trial of a hard-to-treat population, limits statistical power for subgroup analyses, and the polysomnography component was explicitly exploratory. The open-access paper, published on 6 October 2026 with the DOI 10.1038/s41398-026-04443-2, also notes that the version shared early is subject to further editorial edits before the final Version of Record.

Even with those caveats, the implications are immediate. Clinicians administering ketamine for treatment-resistant depression should not assume that a patient whose mood and anxiety improve will necessarily report better sleep, nor that a poor sleep response signals treatment failure on the mood front. The two domains deserve separate monitoring, and patients can be counseled that sleep improvements, when they occur, may scale with the severity of their pre-existing insomnia rather than with the depth of their mood response. For researchers, the study adds to a shifting view of ketamine not as a single blunt instrument but as a pharmacological scalpel whose blades cut along partly independent neural pathways, and it sets the stage for mechanistic work aimed at understanding why the brain’s sleep machinery responds to this anesthetic-turned-antidepressant on a schedule of its own.

Subject of Research: Short-term effects of ketamine on depression, anxiety and insomnia symptoms in treatment-resistant depression

Article Title: Insomnia and anxiety exhibit distinct short-term associations following ketamine in treatment-resistant depression

Article References: Hejazi, N. S., Kheirkhah, M., Chholak, R., Riedner, B., Jamalabadi, H., Walter, M., Hofmann, S. G., Duncan, W., Jr, Goldman, D., & Zarate, C. A., Jr (2026). Insomnia and anxiety exhibit distinct short-term associations following ketamine in treatment-resistant depression. Translational Psychiatry. https://doi.org/10.1038/s41398-026-04443-2

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04443-2

Keywords: ketamine, treatment-resistant depression, insomnia, anxiety, depression, polysomnography, NMDA receptor, Translational Psychiatry, sleep architecture, MADRS, HAM-A, rapid-acting antidepressants