Differential Classical Conditioning of the Nocebo Effect: Increasing Heat-Pain Perception without Verbal Suggestions

Abstract

Background: Nocebo effects, including nocebo hyperalgesia, are a common phenomenon in clinical routine with manifold negative consequences. Both explicit expectations and learning by conditioning are known to induce nocebo effects, but the specific role of conditioning remains unclear, because conditioning is rarely implemented independent of verbal suggestions. Further, although pain is a multidimensional phenomenon, nocebo effects are usually assessed in subjective ratings only, neglecting, e.g., behavioral aspects. The aim of this study was to test whether nocebo hyperalgesia can be learned by conditioning without explicit expectations, to assess nocebo effects in different response channels, and to exploratively assess, whether contingency awareness is a necessary condition for conditioned nocebo hyperalgesia. Methods: Twenty-one healthy volunteers were classically conditioned using painful and non-painful heat stimuli that followed two different cues. The conditioned nocebo effect was assessed by subjective ratings of perceived stimulation intensity on a visual analog scale and a behavioral discrimination task, assessing sensitization and habituation in response to the same stimulation following the two cues. Results: Results show a conditioned nocebo effect indicated by the subjective intensity ratings. Conditioned effects were also seen in the behavioral responses, but paradoxically, behavioral responses indicated decreased perception after conditioning, but only for subjects successfully conditioned as indicated by the subjective ratings. Explorative analyses suggested that awareness of the contingencies and the different cues was not necessary for successful conditioning. Conclusion: Nocebo effects can be learned without inducing additional explicit expectations. The dissociation between the two response channels, possibly representing the conditioned and a compensatory response, highlights the importance of considering different outcomes in nocebo responses to fully understand underlying mechanisms. The present results challenge the role of explicit expectations in conditioned nocebo effects and are relevant with implications in clinical contexts, e.g., when transient adverse effects become conditioned.

Keywords: awareness; behavioral psychology; classical conditioning; heat-pain; implicit learning; nocebo effect.

FIGURE 1

(A) Experimental procedure. In the acquisition sequence, 15 trials with the high cue (solid line) always preceded the high heat stimulus (white filling) and 15 trials with the low cue (dashed line) always preceded the low heat stimulus (gray filling). In the subsequent test sequence, both the high and low cue (five trials each) preceded the low heat stimulus. The sequence of cues was pseudorandomized across the acquisition and test sequence. (B) Trial structure of the conditioning task. Starting at baseline, the thermal stimulation increased or decreased to the intensity of the high or low cue (5 s), followed by an increase to the intensity of the high or low heat stimulus. Afterwards participants rated the perceived stimulus intensity on a visual analog scale (VAS). After this rating, the behavioral discrimination task started and participants kept the perceived stimulus intensity constant for 25 s (self-adjusted temperature); an example of sensitization after high heat (down-regulation of the temperature) and an example of habituation after low heat (up-regulation of the temperature) is shown. After this self-adjustment, the stimulation intensity returned to baseline temperature before the next trial started.

FIGURE 2

Visual analog scale (VAS) ratings of the perceived intensity. Mean and standard errors of mean of the VAS ratings for the high and low cue in the acquisition and the test sequence (N = 21). ^∗p < 0.05; ^∗∗∗p < 0.001.

FIGURE 3

(A) Behavioral responses to the stimulation. Mean and standard errors of mean of the difference between self-adjusted temperature and initial temperature of the behavioral discrimination task for the high and low cue in the learning and test sequence of the whole sample (N = 21) and (B) of the responder (N = 13) and non-responder subgroups (N = 8) in the test sequence. ^∗p < 0.05.

FIGURE 4

Correlation of the nocebo response in the subjective ratings and the behavioral responses. Nocebo responses in subjective ratings and behavioral responses are displayed as standardized differences between the high and low cue.