Sleep paralysis — the unsettling experience of waking or falling asleep unable to move, sometimes accompanied by a sensation of pressure on the chest or a feeling of presence in the room — has long been anecdotally linked to anxiety. A new systematic review and meta-analysis published in the Journal of Clinical Sleep Medicine set out to test that link with more rigor than any prior analysis, and found a relationship that only became clear once the data was examined closely.
The Largest Pooled Analysis of Its Kind
Led by Gustavo Garrido and colleagues at the Bahiana School of Medicine and Public Health in Salvador, Brazil, the team followed PRISMA and Cochrane systematic review standards and pre-registered their protocol with PROSPERO before searching PubMed, Embase, and the Cochrane Library for studies examining sleep paralysis alongside anxiety symptoms or formally diagnosed anxiety disorders.
Eleven studies met the inclusion criteria, together covering 5,568 participants across multiple countries — the largest sample size yet assembled to examine this specific question.
A Result That Only Emerged on Closer Inspection
The initial pooled analysis across all 11 studies found no statistically significant association between sleep paralysis and anxiety (risk ratio 1.28; 95% confidence interval 0.59 to 2.77; p = 0.53) — a result that, taken alone, would suggest the long-assumed sleep paralysis–anxiety link doesn't hold up under rigorous pooling.
But the researchers ran a sensitivity analysis to check whether any single study was distorting the overall estimate — a standard meta-analytic check for outliers. Once one outlier study was excluded, the picture changed substantially: the risk ratio rose to 1.80 (95% confidence interval 1.08 to 2.99), a statistically significant result indicating that people with sleep paralysis face nearly double the risk of having an anxiety disorder compared with those who don't experience it.
Why the Outlier Mattered
Meta-analyses are only as reliable as the studies feeding into them, and a single study with an unusual population, measurement approach, or effect size can pull a pooled estimate toward a null result even when most of the underlying evidence points the other way. That appears to be what happened here: with 10 of the 11 studies pointing toward a real association, one outlier was large or divergent enough to erase statistical significance in the full pooled estimate. The sensitivity analysis — a routine part of rigorous meta-analytic methodology — surfaced a signal that a simpler, unadjusted pooling of the data would have missed entirely.
The authors describe their findings as the most rigorous quantitative synthesis to date on whether sleep paralysis and anxiety are genuinely connected, concluding that the association is real and measurable once outlier-driven distortion is accounted for.
What This Means for Patients
For people who experience recurrent sleep paralysis, this analysis adds evidence-based weight to a connection many have already suspected from their own experience: that these episodes cluster with anxiety rather than occurring as isolated, unrelated events. That has practical implications both ways. Someone with frequent sleep paralysis who also has unrecognized anxiety symptoms may benefit from raising both issues with a clinician together, since treating an underlying anxiety disorder could plausibly reduce how often paralysis episodes occur. Conversely, clinicians treating patients for anxiety disorders may want to specifically ask about sleep paralysis, since it's a symptom patients often don't volunteer unprompted, either from embarrassment or because they don't realize it's a clinically recognized phenomenon worth mentioning.
The findings don't establish which condition causes the other, or whether a shared underlying vulnerability — such as heightened arousal during REM sleep transitions — drives both. Larger, prospective studies that track people over time will be needed to clarify the direction of the relationship.
