Pain Expectations: the neural representations of predicted pain in self and others

Predicting another’s pain is a challenging decision. Predictive coding explains individual pain via top-down modulation, but its extension to vicarious pain is unknown. We tested neural representations of expected pain, uncertainty (risk), and their related errors for self versus other to assess predictive coding’s explanatory power in decisions on another’s pain. We hypothesized that the brain encodes expected pain, risk, and prediction errors for others, mirroring self-pain mechanisms within a predictive coding framework. Using model-based fMRI (n=33), individuals in dyads gambled on risky or certain pain relief; priced their choices; experienced or witnessed corollary pain, and rated the latter’s intensity, alternating between deciding for themselves or another. Behaviorally, prosocial biases emerged in the form of greater risk aversion, higher pain relief valuation, and inflated ratings for others. Neurally, expected other-pain showed a graded representation (absent for self). Pain risk activated the dorsal striatum for both self and other, but more so for others. Pain surprise, but not signed prediction errors, elicited anterior insula activity in both targets. These findings show that predictive coding can characterize vicarious pain, suggesting that top-down priors influence decisions on another’s expected and experienced pain. Importantly, the observed prosocial biases and enhanced neural processing of risk for others highlight that uncertainty exerts a stronger influence in other-regarding decision-making compared to self-regarding choice. These findings lead to significant implications for social interactions and clinical contexts involving surrogate decision-makers.