Expectations underlie the effects of unpredictable pain: a behavioral and EEG study

Abstract

Previous studies on potential effects of unpredictability on pain perception and its neural correlates yielded divergent results. This study examined whether this may be explained by expectations. We presented 41 healthy volunteers with laser heat stimuli of different intensities. The stimuli were preceded either by predictable low, medium or high cues or by unpredictable low-medium, medium-high or low-high cues. We recorded self-reports of pain intensity and unpleasantness and laser-evoked potentials (LEPs). Furthermore, we investigated whether dynamic expectations, i.e. evolving expectations based on actual pain experiences from preceding trials, were better predictors of pain ratings, as compared to fixed expectations based on the cued intensities. Our results replicate previous findings that unpredictable pain is higher than predictable pain for low-intensity stimuli, but lower for high-intensity stimuli. Furthermore, we observed higher ratings for the medium-high unpredictable condition as compared to the medium-low unpredictable condition, in line with an effect of expectations. We also observed that higher certainty of expectations strengthened the influence of expectations on pain intensity ratings. We found significant interactions (N1, N2) for the LEP components between intensity and predictability. However, the few significant differences in LEP peak amplitudes between cue conditions did not survive correction for multiple testing. In line with predictive coding perspectives, dynamic expectations better predicted pain perception than fixed expectations. Considering these findings, we highlight the importance of controlling for expectations and expectation certainty in pain unpredictability manipulations. We also suggest to conceptualize pain expectations more often as dynamic constructs adapting to previous experiences.

The effect of unpredictability on the perception of pain: A systematic review and meta-analysis

“…those from both within and between design studies). Since previous studies have shown that the pain intensity ratings to (un)predictable stimulations are influenced by targeted stimulus pain intensities and by differences in expected pain intensities between the predictable and unpredictable conditions [13,83,128], we conducted meta-regressions with targeted stimulus pain intensity and difference in expected pain intensity (P-U), once independently and once in the same model, to try to disambiguate the effects of these two variables from the effect of unpredictability.…”

“…The effect of difference in expected pain intensity (P-U) showed that pain intensity ratings were higher in the condition with higher expected pain intensity, regardless of whether this was the predictable or unpredictable condition, (difference in expected pain intensity (P-U):𝛽 ̂ = 0.005 ± 0.003, t(94) = 2.131, p = 0.036; Figure 2B). Targeted stimulus pain intensities and expected pain intensities are potential confounders of the effect of unpredictability on pain perception [13,83,128]. Therefore, to try to disambiguate the three effects, we tested targeted stimulus pain intensity and expected pain intensity in the same model, where we selected only Hedges'…”

“…We also observed that safety signaling [100,101], that is the predictability of pain onset involving the predictability of phases without pain, did not appear to reduce pain perception, as shown by the absence of significant difference between the conditions with predictable and unpredictable pain onsets. Since previous work showed that the effect of unpredictability could be (at least partially) explained by expectations and targeted stimulus pain intensity [13,83,128], we tried to control for these variables in our analyses. However, we did not find any significant direct effect of unpredictability on pain intensity ratings even when controlling for these variables.…”

“…However, the role of unpredictability is not clear-cut, especially in healthy volunteers. While some experimental studies confirm that unpredictable painful stimuli result in higher perceived intensity than predictable ones [6,8,127], other studies report unpredictable painful stimuli being perceived as less or equally intense than predictable ones [13,83,84,91,110,129].…”

“…First, predictable and unpredictable conditions can induce differences in expected pain intensity because humans do not remain naive throughout an experiment but acquire expectations based on their past pain experiences which may differ between conditions [128]. These differences in expectations may partly explain unpredictable pain perception [13,83,128]. Second, differences in controllability might emerge in unpredictability manipulations as predictability facilitates controllability [73].…”