Imagine standing in line for a flu shot when the person ahead of you walks out wincing, muttering that it was agony. Does their experience change yours? According to a new study from Dartmouth College, the answer is yes — and the mechanisms behind that shift are more deeply embedded in human cognition than previously understood.

Published in the Proceedings of the National Academy of Sciences, the research demonstrates that unreinforced social cues — essentially, suggestions from other people about how an experience will feel — can meaningfully alter how individuals perceive physical pain, empathy for others in pain, and even the mental effort required to complete a cognitive task. What’s more, these altered perceptions tend to persist over time, resisting correction even when the underlying reality contradicts them. The culprit is a well-documented but poorly understood cognitive mechanism: confirmation bias in learning.

“Our results suggest that when expectations are shaped by social information, people tend to hold onto those expectations which in turn impacts how we feel in a long-lasting way,” said Aryan Yazdanpanah, a doctoral candidate in Dartmouth’s Department of Psychological and Brain Sciences and the study’s lead author.

The Experiment

The research team recruited 111 healthy participants and ran them through three distinct experiential tasks across multiple sessions. In the first — a somatic pain task — heat was applied to participants’ inner forearms at varying intensities. In the second, a vicarious pain task, participants watched videos of individuals displaying different levels of pain expression. The third was a cognitive effort task, in which subjects mentally rotated three-dimensional objects and judged whether two images were identical.

Before each trial, participants viewed a cluster of dots on a computer screen that they were told represented how 10 previous participants had rated the same stimulus. In reality, those dots were completely randomized — bearing no relationship whatsoever to actual stimulus intensity. They were, in the parlance of the research, “mere suggestions.”

Despite the cues being meaningless, participants’ ratings shifted dramatically in their direction. High-aversive cues led participants to rate stimuli as more painful, more empathy-inducing, and more effortful — even when the objective stimulus intensity was identical to conditions preceded by low-aversive cues. The effect sizes were substantial, ranging from 0.73 to 0.86 across tasks.

A Feedback Loop That Feeds Itself

The study’s deeper contribution lies in its computational modeling of why these effects persist. The researchers found evidence for two interlocking mechanisms. The first is expectation coloring perception: when people anticipate something will be painful or difficult, their experience is genuinely shifted in that direction. The second — and more consequential — is confirmation bias in learning.

“We found that a person will favor the evidence that aligns with their beliefs but will ignore or dampen those which do not align,” said co-senior author Alireza Soltani, an associate professor of psychological and brain sciences at Dartmouth.

In practical terms, this means that when a high-aversive cue is followed by a genuinely painful experience, participants updated their expectations significantly — reinforcing the original cue. But when the same high-aversive cue was followed by a milder-than-expected experience, participants updated far less. The system, in other words, is rigged to confirm what it already believes. Learning rates for cue-congruent prediction errors were significantly higher across all three tasks — a finding confirmed by both behavioral analysis and the computational model that best fit the data.

The best-fitting computational model — labeled Model 3c by the researchers — included separate learning rates for congruent and incongruent prediction errors, a perceptual carryover effect (where one trial’s experience bleeds into the next trial’s expectation), and a weighting parameter reflecting the relative influence of expectations versus actual stimuli on perception. Stimulus intensity still dominated experience in all three tasks, but social expectations carved out a meaningful and durable share.

This mirrors findings from earlier work by Jepma and colleagues, who demonstrated analogous self-reinforcing mechanisms in conditioned pain cues.¹ The current study’s advance is applying similar analysis to unconditioned, unreinforced social suggestions — the kind that circulate freely through conversations, online reviews, and social media — and extending it across three distinct experiential domains.

Placebo Effects, Chronic Pain, and the Hyperconnected World

The implications extend well beyond the laboratory. Co-senior author Tor Wager, the Diana L. Taylor Distinguished Professor of Neuroscience at Dartmouth and a leading researcher in pain neuroscience and placebo effects, framed the findings in broader terms.

“The dynamics we observed can create self-fulfilling prophecies — feedback cycles that affect many kinds of health conditions, including chronic pain and fatigue, as well as beliefs about other people.”

The study draws an explicit line to the placebo and nocebo literature. Decades of research have shown that contextual information — a clinician’s tone, the color of a pill, a warning label — can shape how patients experience symptoms and recover from treatment.²’³ What makes the current findings striking is that the cues here were never reinforced. No one was conditioned to expect pain; they simply received a suggestion, and that was enough to create lasting perceptual drift.

The researchers also speculate about clinical implications for chronic fatigue syndrome and cognitive fatigue, conditions in which the subjective experience of effort may become decoupled from underlying physiology — possibly through exactly the kind of socially mediated feedback loop the study describes. The connection, while promising, the authors caution, remains to be tested directly.

Yazdanpanah offered a concrete example of the stakes: “If a person is truly in severe pain but others believe that the pain is not serious, this social belief may lead you to underestimate or overlook that person’s suffering.”

The finding that individuals show stable susceptibility to social cue effects across all three tasks — pain, vicarious pain, and cognitive effort — suggests a shared cognitive trait, not a domain-specific quirk. A person highly susceptible to social suggestion in one context tends to be susceptible across contexts. This cross-domain consistency, the researchers argue, points toward shared cognitive processes and may open avenues for interventions designed to interrupt these feedback loops.

In an era when reviews, ratings, and shared experiences instantaneously reach millions of people, the capacity of social information to manufacture subjective reality takes on dimensions the study’s authors take seriously. What others tell us about how something feels may matter as much as how it actually feels — and once those expectations take hold, changing them may require more than simply encountering the truth.

---

Endnotes

1. Jepma M, Koban L, van Doorn J, Jones M, Wager TD. Behavioural and neural evidence for self-reinforcing expectancy effects on pain. Nature Human Behaviour. 2018;2:838–855.
2. Tracey I. Getting the pain you expect: mechanisms of placebo, nocebo and reappraisal effects in humans. Nature Medicine. 2010;16:1277–1283.
3. Wager TD, Atlas LY. The neuroscience of placebo effects: connecting context, learning and health. Nature Reviews Neuroscience. 2015;16:403–418.
4. Yazdanpanah A, Jung H, Soltani A, Wager TD. Social information creates self-fulfilling prophecies in judgments of pain, vicarious pain, and cognitive effort. Proceedings of the National Academy of Sciences. 2026;123(7):e2513856123. https://doi.org/10.1073/pnas.2513856123
5. Dartmouth College. “People’s opinions can shape how negative experiences feel.” EurekAlert! March 12, 2026. https://www.eurekalert.org/news-releases/1119856