Table 1.
Factors that could explain previous contradicting results.
Fig 1.
An illustration of the three experimental tasks: stimulus-response task (top panel), expectation task (middle panel), and cued-perception task (bottom panel; performed during an fMRI scan). When two rectangles are presented on top of each other (e.g., for the ratings in all tasks), the lower one illustrates pain trials and the upper one illustrates visual trials.
Fig 2.
Behavioral results - expectation task.
(A) Participants’ expectation ratings as a function of the mean and variance of the cue’s values, in pain and vision trials. (B) Participants’ expectation ratings as a function of the skewness of the cue’s values, in pain and vision trials. In both panels, bars represent the mean across participants, error bars represent the standard error of the mean across participants, and points represent single participants.
Fig 3.
Computational model of cue-based expectation generation.
(A) Simulations of the expectation model: Mapping of cue values, V (10 per cue), to weights for expectation computation, based on the two free parameters of the model: k and b. When k = 1 (black line), inliers and outliers are equally weighted. When k < 1 (cold colors), inliers are over-weighted, and when k > 1 (hot colors), outliers are over-weighted. When b < 0 (left panels), values below the mean are over-weighted, when b > 0 (right panels) values above the mean are over-weighted, and when b = 0 (middle panel), values are equally weighted. The dashed line represents the cue mean. (B) Correlations between the observed and predicted (based on the computational model) expectation ratings were very high across participants. Each line represents a single participant, and each dot represents a single trial. Data from different participants are presented in different colors. (C) The weight function for each modality, based on the median k and b values across the group. Weight functions of individual participants are overlaid with thin, gray lines. The dashed line represents equal rating of all cue values (V). Note that panels A and C are based on a symmetric cue consisting of equally distributed values between 20 and 60.
Fig 4.
Behavioral results, cued-perception task.
(A) Participants’ perception ratings as a function of the mean (x axis) and variance (color) of the cue’s values, in pain and visual trials (column). (B) Participants’ perception ratings as a function of the skewness (x axis and color) of the cue’s values, in pain and visual trials (columns). In panels A and B, bars represent means across participants, error bars represent the standard error of the mean across participants, and points represent single participants. (C) The effect of cue mean (color) on pain and visual contrast ratings over time, for low vs. high stimulus intensity (columns), in pain and visual perception (rows). Error bars represent standard error of the mean across participants. The lines represent the linear fit, and the shading represents 95% CIs of the linear fit. Vertical dashed lines separate between different blocks (note that a new skin site was used for each block, and thus the increase in the averaged pain rating for the first trial of each block stems from site-nonspecific sensitization and site-specific habituation [71]).
Fig 5.
(A) The NPS and SIIPS neuromarkers. Adapted from Botvinik-Nezer et al. (2024), Nature Communications (DOI: https://doi.org/10.1038/s41467-024-50103-8) [62], under CC BY 4.0. (B) NPS (left) and SIIPS (right) score as a function of the stimulus intensity (x axis), cue mean (color) and modality (column). Bars represent means across participants, error bars represent the standard error of the mean across participants, and points represent single participants. (C) Multilevel mediation analysis with pain neuromarkers. Solid lines represent significant effects and dashed lines represent non-significant effects. Asterisks represent the level of significance (* p < .05, *** p < .001).
Fig 6.
(A) ROIs are shown with a contour for each modality (rows) and processing stage (columns). Regions with higher activity for more intense stimuli are presented with a green contour. Regions with a significant cue mean effect are presented in red (higher activity for higher cue mean) or blue (higher activity for lower cue mean). (B) The ROI score in the two regions with significant cue mean effect during pain perception (including other parts of the NAc) as a function of the stimulus intensity level (x axis) and cue mean (color). Bars represent means across participants, error bars represent standard error of the mean across participants, and points represent single participants. Asterisks represent the level of significance (* p < .05, *** p < .001). Abbreviations: aIns = anterior insula; aMCC = anterior midcingulate cortex; dlPFC = dorsolateral prefrontal cortex; dpIns = dorsal posterior insula; L = left; Med = medial; NAc = nucleus accumbens; OFC = orbitofrontal cortex; PAG = periaqueductal gray; R = right; thal = thalamus.