Table 1.
Summary of studies demonstrating that different expectations regarding imminent nociceptive stimuli may alter pain perception in different experimental contexts.
While the study examples showed that the manipulation of perception was dependent on the stimulus context, no study has yet shown that the perception of nociceptive stimuli at threshold can be changed to the opposite of the respective cue.
Fig 1.
Graphical scheme of the experimental settings applied in either group 1 (top bars) or group 2 (bottom bars) subjects.
The paradigm applied to either group is shown in each two lines of differently colored small bars of which each denotes a CO2 stimulus and the associated cues. Specifically, in the upper lines, the CO2 stimuli are shown in their chronological succession differently colored for 25% v/v CO2 stimuli (green), CO2 concentration at the individual pain threshold level (grey) and 75% v/v CO2 (red). Below each stimulus, in the second line the cues given before stimulus application are denoted as “No pain” (green), “Stimulus”, i.e., a neutral cue (grey) and “Pain” (red). In the first group, the far-from-thresholds stimuli were cued randomly, i.e., green and red bars in lines 1 and 2 are mixed, whereas in the second group, the 25 and 75% stimuli were always preceded by correct cues, which can be seen by the consistent agreement between green or red bars in lines 1 and 2. The inequality of the experimental settings between the groups is commented on in the discussion of this paper.
Fig 2.
Bar graph of the percentage of stimuli at threshold level that were perceived as painful, separately for each different cue and for the two groups.
The columns and error bars indicate means and standard deviations. Left (red bars): In group 2, where among the stimuli at pain threshold level, correctly cued, far-from-threshold stimuli were randomly interspersed, a significantly higher percentage of stimuli at pain threshold level were rated as painful after a “pain” cue as compared to stimuli rated as painful after a “neutral” or a “no pain” cue. Right (blue bars, main results): In group 1, where, in contrast, the stimuli at threshold as well as the interspersed far-from-threshold stimuli were randomly and equally often cued as painful, neutral or not painful, s the effect was completely inversed. The group*cue interaction effect was statistically highly significant (Table 2). The dashed lines superimposed onto the bar graphs indicate the fits of the linear model used to mathematically describe the relationship between the percentages of stimuli at pain threshold level rated as painful and the three different cues as (Table 4).
Table 2.
Statistical details of the percentages of stimuli at pain threshold level rated as painful: Top: Means and standard deviations of the pain perception [%] of the CO2 stimuli at pain threshold level (“Group 1”: far-from-threshold stimuli randomly announced as “no pain”, “stimulus”, “pain”; “Group 2”: far-from-threshold stimuli correctly announced, i.e. “no pain” and “pain”).
Bottom: Detailed results of the analyses of variance for repeated measures and the estimates of the respective effect sizes partial eta squared, . Results are provided for main effects and interactions, with the F-value followed by the number of degrees of freedom in parentheses and the p-value.
Table 3.
Statistical details of the VAS ratings of the subjects’ certainty about their judgments of the stimuli at pain threshold level as painful or not: Top: Means and standard deviations of the ratings of the subjects’ certainty about their judgment of threshold CO2 stimuli as painful or not (“Group 1”: far-from-threshold stimuli randomly announced as “no pain”, “stimulus”, “pain”; “Group 2”: far-from-threshold stimuli correctly announced, i.e. “no pain” and “pain”).
Bottom: Detailed results of the analyses of variance for repeated measures and the estimates of the respective effect sizes partial eta squared . Results are provided for main effects and interactions, with the F-value followed by the number of degrees of freedom in parentheses and the p-value.
Table 4.
Parameters of the final regression model.
Specifically, the percentages of stimuli at pain threshold level rated as painful were associated with the three different cues using a linear model of , where Cue = [–1,0,1] for “no pain”, “stimulus or neutral cue, and “pain”, respectively; Group = [1,2]; and Sex = [0,1] for women and men, respectively. The fits are plotted in Fig 2. The final model was the result of goodness-of-fit statistics. Only parameters that provided a statistically significant improvement when free to be estimated remained part of the final model, which therefore, in its short form reads as
.