Differential effects of heartbeat perception training on interoceptive accuracy in high vs. low symptom reporters

Abnormal interoception may be a hallmark of medically-unexplained symptoms (MUS), although its underlying mechanisms remain unclear. We investigated whether MUS are associated with enhanced interoceptive learning, potentially facilitated by post-learning stress. From 177 volunteers providing self-reports on MUS, 22 individuals were recruited from the upper and 26 from the lower quartile (high/HSR and low symptom reporters/LSR) and randomly assigned to a stress or a control condition (LSR-stress/n = 12; LSR + stress/n = 14; HSR-stress/n = 11; HSR + stress/n = 11). Participants underwent a heartbeat perception training as interoceptive perceptual learning task, followed by a socially-evaluated cold pressor stress or a control intervention. Interoceptive accuracy was measured using the heartbeat counting and discrimination task at baseline (T₁), 30 min. (T₂) and 24 h (T₃) after the stress/control intervention. Sympatho-adreno-medullary (SAM) and hypothalamic-pituitary-adrenal (HPA) axes were measured in terms of heart rate, blood pressure (SBP/DBP) and salivary cortisol reflected reactivity. The stress test induced a robust increase of SBP/DBP in both stress groups, and a cortisol increase in the LSR + stress group only. Post-learning stress induced an increase of interoceptive accuracy (heartbeat discrimination) the HSR + stress group only. Correlation/regression analyses showed a positive association between changes in interoceptive accuracy (heartbeat discrimination) from T₁ to T₂ with MUS, and changes from T₂ to T₃ with SAM and HPA axis responsiveness. Post-learning stress may enhance interoceptive learning in MUS, which may explain why benign body sensations are perceived as symptoms. While MUS-specific heightened attention to body sensations may be responsible for early effects, physiological stress axes may explain late effects of interoceptive learning, including overnight consolidation.