How the magnitude and precision of pain predictions shape pain experiences

Abstract

In Bayesian and predictive coding frameworks, the magnitude and precision of pain expectancies are key determinants of perceptual experience However, relatively few studies have directly tested whether this holds for pain, and results so far have been inconclusive. Here, we investigated expectancy effects on pain experiences and affective responses. In two studies, healthy participants (n=30 in each study) received painful electrical stimuli preceded by explicit pain predictions. In study 1, the magnitude of pain predictions and administered pain intensities were varied. In study 2, the magnitude and precision of pain predictions were varied, while administered pain intensity was kept constant. Experienced pain intensity was the primary outcome in both studies. Study 1 found that pain experiences assimilated towards both under- and overpredictions of pain for moderate and very high pain stimuli, but this was not observed for non-painful stimuli. In addition, effects were not necessarily larger with predictions of greater magnitude. Study 2 found assimilation of pain experiences upon under- and overpredictions of pain, regardless of precision level, but no effects on EMG eyeblink startle responses. Moreover, results of both studies showed that under- and overpredictions caused disappointment and relief, respectively. In conclusion, the influence of pain predictions on pain may be more complex than assumed in current theoretical frameworks, with no systematically stronger assimilation of pain experiences to larger and more precise predictions. Since overpredictions are associated with relief, but underpredictions with disappointment, these findings underline the importance of providing realistic predictions when preparing for upcoming painful procedures.

How should uncertainty about upcoming painful procedures be communicated? An experimental study into highly uncertain pain predictions

“…Using a within-subject study design, participants were repeatedly presented with highly uncertain pain predictions by a hypothetical doctor and electrical stimuli that were individually calibrated to be not, moderately, or very highly painful. The design was partly based on our recent experiments (26).…”

“…A sample size of 30 participants was considered sufficient to obtain enough statistical power for this design and analytical approach (26). Recruitment was done through printed flyers, verbally, or through social media.…”

“…Participants had to be between 18 and 35 years old, healthy, and fluent in English. The exclusion criteria were identical to our recent studies (26). All participants were either compensated monetarily or with study credits according to the Institute's standard rate (2 course credits or €8.50 per hour).…”

“…Potential moderators) in Qualtrics (Qualtrics, Provo, UT). Subsequently, a previously used individualized calibration procedure (26) was used to select electrical stimulus intensities matching to no pain, low pain, moderate pain, high pain, and very high pain. This was followed by a task assessing the variability of pain intensity ratings (see 2.6.…”

“…Learning what others experience (i.e., observational learning) has been found to affect pain perception by shaping expectancies (15). Indicating a range of pain intensities that can be expected can be used to express uncertainty (25)(26)(27), where a very broad range may be indicated in case of high uncertainty. We tested potential moderation of the response to uncertainty by individual characteristics (i.e., variability of pain ratings, state and trait anxiety, optimism, intolerance of uncertainty, and pain vigilance) (15,28).…”