Publication date: Jul 29, 2025

Severe symptoms in the absence of measurable body pathology are a frequent hallmark of post-COVID syndrome. From a Bayesian Brain perspective, such symptoms can be explained by incorrect internal models that the brain uses to interpret sensory signals. In this pre-registered study, we investigate whether induced breathlessness perception during a controlled COrebreathing challenge is reflected by altered respiratory measures (physiology and breathing patterns), and propose different computational mechanisms that could explain our findings in a Bayesian Brain framework. We analysed data from 40 patients with post-COVID syndrome and 40 healthy participants. Results from lung function, neurological and neurocognitive examination of all participants were within normal limits on the day of the experiment. Using a Bayesian repeated-measures ANOVA, we found that patients’ breathlessness was strongly increased (BF=8. 029, BF=11636, BF=43662) compared to controls. When excluding patients who hyperventilated (N = 8, 20%) during the experiment from the analysis, differences in breathlessness remained (BF=1. 283, BF=126. 812, BF=751. 282). For physiology and breathing patterns, all evidence pointed towards no difference between the two groups (0. 307 > BF 

Open Access PDF

Semantics

Original Article