Increased breathlessness in post-COVID fatigue despite normal breathing behaviour in a rebreathing challenge - A Bayesian brain perspective on post-COVID fatigue

Background:Multiple pathophysiological changes have been identified in patients with post-COVID syndrome. However, a comprehensive understanding of the underlying mechanism explaining the symptoms is still lacking. Here, we investigate processing of respiratory information for breathing control and symptom perception by measuring the perceptual, behavioural and physiological responses to a controlled rebreathing challenge.Methods:In this pre-registered rebreathing paradigm, we investigated 40 patients suffering from severe post-COVID fatigue (N=22 with breathlessness) and 40 healthy participants matched for age, gender and BMI. Participants were only included if lung function testing, neurological and neurocognitive examination were within normal limits on the day of the experiment. During the experiment, respiratory measures (physiology and behaviour) and breathlessness ratings were recorded. Groups were compared using Bayesian repeated measures ANOVA.Results:Patients’ breathlessness is strongly increased (BF10,baseline=8.029, BF10,rebreathing=11636, BF10,recovery=43662) compared to controls, also in patients without post-COVID breathlessness. When excluding patients who hyperventilated (N=8, 20%) during the experiment from the analysis, differences in breathlessness remain (BF10,baseline=1.283, BF10,rebreathing=126.812, BF10,recovery=751.282). In contrast, for physiology and breathing behaviour, all evidence points towards no difference between the two groups (0.307>BF10<0.704).Conclusion:While breathing control is mostly intact, processes for symptom perception are impaired in patients with post-COVID fatigue. We propose different computational mechanisms that could underlie this erroneous processing by adopting a Bayesian brain perspective.