Depression, Anxiety and HRV: What the Research Shows
Editorial: ONDA Life (Yakiv Bilenko) · Updated November 10, 2026

On average, groups of people with some mental disorders have lower heart rate variability than healthy groups. For depression and generalized anxiety this evidence is weaker than for conditions such as PTSD, and the links are small and differ by metric. Lower HRV only slightly anticipates later depressive symptoms. HRV cannot diagnose, screen for or predict depression or anxiety in one person.
Key points
- Group studies link lower HRV with several mental disorders, but for depression and generalized anxiety the evidence is weaker than for PTSD or schizophrenia.
- Links between HRV and anxiety severity are weak to moderate and appear only for some combinations of metric and questionnaire.
- In people without depression, lower HRV is linked only slightly with more depressive symptoms later.
- In PTSD, the HRV difference from controls is consistent but its size depends on how long the recording is.
- The direction is unknown: HRV differences may follow the condition, its treatment, sleep or activity rather than cause it.
- HRV is not a diagnostic or screening test for any mental disorder, and a low reading is not a sign of one.
What is the link between HRV, depression and anxiety?
Heart rate variability (HRV) is the natural variation in the time between heartbeats. Research has compared it between people with mental disorders and people without them for decades. The short version: on average, groups with some disorders show lower HRV, but the differences are modest, they depend on the disorder and the metric, and they do not work as a test for an individual [S1].
This page is about what those studies found. It is not about treating depression or anxiety, and it does not recommend any practice for them.
How does it work?
HRV measures such as RMSSD and high-frequency power are read as reflections of vagally mediated changes in heart rate; Vagal tone cannot be measured directly; HRV measures such as RMSSD reflect vagally mediated changes in heart rate [S6]. The vagus nerve links the brainstem with the heart, and the same brain networks are involved in emotion and in regulating the heart. That is why researchers expected HRV and mood to be related.
The direction of the link is not known. Lower HRV may be part of a vulnerability, or it may follow the condition through poor sleep, less physical activity, other illnesses or medication. A meta-analysis of longitudinal studies describes HRV as reflecting mechanisms of vulnerability, with small effects [S3]. Treatment itself can change HRV: before-and-after studies of antipsychotic medication, brain stimulation, physiotherapy and psychotherapy found altered HRV, on weak evidence [S1]. So a person's HRV during treatment does not show the condition alone.
How is it measured?
Most studies record HRV with an ECG in a laboratory, usually for a few minutes at rest, sometimes over a whole day. Depression and anxiety are measured separately, with interviews and questionnaires. The results depend on both choices. In PTSD, for example, the difference from controls was larger in short recordings than in all-day ones [S4]. Recording length, setting, breathing and analysis all shape the value, and guidelines ask for them to be reported and considered [S6].
Consumer watches and rings measure HRV differently, mostly at night and from the pulse at the skin. The research on mental disorders was not done with these values, and its findings do not transfer to them directly. How wearables measure HRV is covered in can you trust HRV from a smartwatch or ring?
What affects it?
HRV is shaped by many things that also change with low mood or anxiety: sleep, physical activity, alcohol, illness, breathing during the recording, and medication [S1] [S6]. This is one reason why studies disagree, and why a difference between groups cannot be pinned on the mental disorder itself. The everyday factors are covered in why is my HRV low?
What does the evidence show?
Mental disorders in general — context-dependent. An umbrella review of meta-analyses (Wang et al., 2025) compared HRV between people with mental disorders and healthy controls across many conditions [S1]. On the authors' grading, the strongest evidence of lower HRV reached only a "suggestive" level, for dementia, PTSD, somatic symptom and functional somatic disorders, and schizophrenia [S1]. No disorder reached a higher level. No two disorders showed the same pattern, and the credibility of the findings varied by metric [S1].
Depression and generalized anxiety — context-dependent, weaker. For major depressive disorder and generalized anxiety disorder, the same review rated the evidence weaker, below the suggestive level [S1]. A difference may exist, but it is less certain than many popular articles suggest.
Anxiety severity — context-dependent. A meta-analysis of studies that measured both HRV and anxiety questionnaires found only a few significant associations, weak to moderate, and only for certain combinations of questionnaire and metric; most combinations showed none [S2]. One of them involved the LF/HF ratio, an index whose physiological meaning is disputed. The authors conclude that HRV is not a replacement for psychometric scales [S2].
Later depressive symptoms — context-dependent. A meta-analysis of longitudinal studies followed people who were not depressed at the start. Lower HRV was linked with more depressive symptoms later, but the correlations were small; the authors call the prospective effects small and context-dependent [S3]. That is far from a prediction for one person.
PTSD — context-dependent. A meta-analysis found vagally mediated HRV consistently lower in people with PTSD than in controls, with a larger difference in short recordings than in all-day ones [S4]. A critical review of the same field describes the differences as small, variable and context-dependent, and does not recommend HRV as a standalone clinical or screening tool [S5].
Unknown. Whether lower HRV comes before these conditions or follows them; whether consumer night-time HRV shows the same group differences; and whether HRV adds anything useful to a clinical assessment.
What it does not tell you
- Not a diagnosis. No HRV value shows that someone has depression, anxiety or PTSD, and no HRV value rules them out [S2] [S5].
- Not a screening test or a forecast. The prospective links are small, and HRV has not been shown to identify who will develop symptoms [S3].
- Not your mood. A single low HRV reading does not by itself mean you are stressed or unwell [S6]. A low night can follow a short sleep, a drink, a cold or a hard workout.
- Not a measure of treatment. Medication and therapy can change HRV, so a change during treatment does not by itself say whether the treatment is working [S1].
- When to get help. If low mood, worry or anxiety last for weeks or get in the way of daily life, talk to a doctor or a mental health professional. If you have thoughts of harming yourself, contact emergency services or a crisis line now. Do not wait for a number on a device to confirm how you feel.
In ONDA
ONDA builds a personal baseline from values stored in Apple Health [S7]. The window is 14 days, and ONDA compares your latest day of data with your own corridor — the average of your earlier days with enough data plus or minus one standard deviation — and flags it only when it is at least 1.5 standard deviations outside and has changed by a minimum amount. A value outside your corridor describes a change in your own data, not its cause. ONDA does not assess mood, does not detect depression or anxiety, does not diagnose any condition and does not replace a doctor or a mental health professional.
Educational information, not a diagnosis or medical treatment.
Evidence at a glance
| Claim | Evidence | Limitation |
|---|---|---|
| Across meta-analyses, lower HRV in mental disorders compared with healthy controls reached at most a suggestive level of evidence, for dementia, PTSD, somatic symptom and functional somatic disorders and schizophrenia. [S1] | Context-dependent | Umbrella review of meta-analyses in clinical samples; inherits their biases; group differences, not individual findings. |
| For major depressive disorder and generalized anxiety disorder, the evidence for HRV differences was weaker, below the suggestive level. [S1] | Context-dependent | Weaker evidence means the difference is uncertain, not that it has been ruled out. |
| No two disorders showed the same pattern of HRV changes, and credibility varied by HRV metric and disorder. [S1] | Context-dependent | Pattern differences come from pooled group data and are not usable to tell disorders apart in one person. |
| Before and after treatments such as antipsychotics, rTMS, physiotherapy and psychotherapy, HRV changed, but the evidence for these changes was weak. [S1] | Emerging | Weak evidence; describes how treatment can change HRV, not a recommendation for any treatment. |
| Only a few combinations of anxiety questionnaire and HRV metric showed significant associations with anxiety severity, and these were weak to moderate. [S2] | Context-dependent | Cross-sectional correlations; about half of the included studies entered the pooled analysis; the LF/HF ratio is not a measure of autonomic balance. |
| The authors of the anxiety meta-analysis conclude that HRV is not a replacement for psychometric scales. [S2] | Context-dependent | Authors' conclusion; the suggested complementary role has not been tested as a clinical tool. |
| In initially non-depressed people, lower HRV showed small negative correlations with later depressive symptoms. [S3] | Context-dependent | Small correlations with heterogeneous instruments and follow-up; publication bias not excluded; not an individual prediction. |
| The prospective effects of HRV on depressive symptoms are small and depend on context. [S3] | Context-dependent | Prediction models were synthesized qualitatively; usefulness for screening has not been shown. |
| Vagally mediated HRV is consistently lower in PTSD than in controls, and the difference is larger in short recordings than in all-day recordings. [S4] | Context-dependent | Group-level meta-analysis; recording duration changes the size of the effect. |
| In PTSD, HRV differences from controls are small, variable and context-dependent, and HRV is not recommended as a standalone clinical or screening tool. [S5] | Context-dependent | Narrative review; authors' assessment of measurement reliability and validity. |
| HRV is not a direct readout of vagal tone or of sympathetic–parasympathetic balance. [S6] | Guideline / expert consensus | Methodological guidance (expert consensus); says nothing by itself about mental health. |
| The input signal, recording length and setting, breathing and the analysis method all affect the reliability and interpretation of an HRV measurement. [S6] | Guideline / expert consensus | Methodological guidance for research recordings; consumer wearables add further limits. |
| A single low HRV reading does not by itself mean a person is stressed or unwell; HRV is not a specific marker of a state. [S6] | Guideline / expert consensus | Methodological guidance (expert consensus), not a study of mental states. |
| ONDA builds a personal baseline from values in Apple Health and compares new values with the user's own corridor. [S7] | Established | Describes app behaviour only; not evidence for any health claim. |
Sources
- [S1] Wang et al. (2025). Heart rate variability in mental disorders: an umbrella review of meta-analyses. Translational Psychiatry. DOI 10.1038/s41398-025-03339-x · PMID 40155386
- [S2] Óvári et al. (2026). Heart rate variability and anxiety severity: A systematic review and meta-analysis. Comprehensive Psychiatry. DOI 10.1016/j.comppsych.2026.152738 · PMID 42543052
- [S3] Galin et al. (2026). Heart rate variability as a predictor of future depressive symptoms: A systematic review and meta-analysis. Journal of Affective Disorders. DOI 10.1016/j.jad.2026.121388 · PMID 41692051
- [S4] Tucker, Taylor & Robinson (2026). A multi-level meta-analysis of vagally-mediated heart rate variability and post-traumatic stress disorder. Neuroscience and Biobehavioral Reviews. DOI 10.1016/j.neubiorev.2026.106585 · PMID 41616929
- [S5] Neary, Carleton & Guzik (2026). Critical re-evaluation of heart rate variability as a predictor of posttraumatic stress disorder: a narrative review. Neuroscience. DOI 10.1016/j.neuroscience.2026.07.003 · PMID 42413601
- [S6] Carter et al. (2026). Guidelines for rigor and reproducibility of heart rate variability within human cardiovascular research. Am J Physiol Heart Circ Physiol. DOI 10.1152/ajpheart.00041.2026 · PMID 42495990
- [S7] ONDA — product documentation: How ONDA works. How ONDA works.
Related
- ScienceWhat a Single HRV Value Can and Can't Tell You
- ScienceAcute Stress and HRV: What Happens and How You Recover
- ScienceCortisol and the Stress Response: What It Does and What Is Myth
- ScienceHRV Baseline: Why Your Own Normal Matters More Than Any Norm
- ScienceHRV and Heart Rate During Sleep: Why the Night Is the Best Window
- ScienceWhy Is My HRV Low? Common Causes and What the Evidence Shows
- GlossaryHeart Rate Variability
- ArticleYour HRV Reading Is Low. Now What?
- ToolApple Watch Baseline
How ONDA Science pages are made: every number comes from one checked list of facts, every claim is mapped to its sources and graded by strength of evidence, and sources need a DOI or PMID (manufacturer documentation is used only for device facts).