Pain can hurt more if you are not expecting it. (Antonio Guillem/Shutterstock)
In a nutshell
• Pain feels more intense when it surprises you. Researchers found that unexpected or delayed pain is amplified by the brain because of the mismatch between what we predict and what we experience.
• Your sense of control matters. Pain perception was only affected by surprise when participants actively controlled the virtual knife themselves, highlighting how our sense of agency influences pain intensity.
• This finding could transform pain treatment. Understanding that pain is processed as “Bayesian surprise” could lead to new therapies for chronic pain that focus on managing expectations and reducing uncertainty.
TSUKUBA, Japan — Stubbing your toe can be surprisingly painful, leaving you wincing more than anticipated. It turns out that there is actually a scientific reason for that. According to research from the University of Tsukuba in Japan, our brains amplify pain when it catches us by surprise.
In a study published in Cognition, researchers have discovered that when pain comes unexpectedly or differs from what we anticipate, we actually feel it more intensely. This challenges conventional wisdom about pain perception and opens new avenues for understanding chronic pain conditions.
The research team investigated a phenomenon they call “Bayesian surprise” in pain perception. This concept refers to how our brains process pain differently when reality doesn’t match our expectations. Pain experiences vary dramatically from person to person and even within the same person at different times. Sometimes an injury hurts terribly, while a similar one at another time feels much less intense.
How our expectations affect pain
This inconsistency happens because pain perception depends heavily on our expectations and how certain we are about what’s coming. While scientists have known for years that expectations influence pain (think placebo effects), this study digs deeper into exactly how our brains integrate predictions with actual sensory experiences as they happen.
The research team created a clever experiment using virtual reality. Participants wore VR headsets and used controllers to manipulate a virtual knife with their left hand, stabbing it into their virtual right forearm at specific moments indicated by beeps. Meanwhile, a thermal device applied real heat to the corresponding spot on their actual arm.
First, they varied the timing of the heat stimulation. Sometimes, it happened right when the virtual knife made contact, and other times, it was delayed by about a second. They occasionally made the virtual knife suddenly vanish just before contact. This setup lets them test how prediction errors affect pain perception in different scenarios.
Scientists have proposed two main theories about how the brain processes pain. The “Estimate Hypothesis” suggests the brain calculates pain intensity based on predictions, while the “Surprise Hypothesis” proposes that pain is experienced as the difference between prediction and reality, essentially, how surprised your brain is. The researchers designed their experiment specifically to see which explanation holds up better.
Participants reported much higher pain intensity when the heat stimulus was delayed and when the knife unexpectedly disappeared. People felt pain most strongly when the prediction error was large, which strongly supports the Surprise Hypothesis. This contradicted what the Estimate Hypothesis would predict, that the disappearance of the threat should have reduced pain perception. Instead, the mismatch between expectation and reality actually amplified the pain experience.
Could this help chronic pain sufferers?
By comparing the two competing theories mathematically, they found overwhelming evidence supporting the Bayesian surprise model. This suggests our brains perceive pain not just based on the physical stimulus itself, but on how much that stimulus violates our expectations. When something surprising or unpredictable happens related to pain, that prediction error itself becomes part of what we feel.
This discovery has huge potential for chronic pain conditions. People suffering from chronic pain often experience vague pain-related fears and anxieties. The researchers suggest that this uncertain gap between expectation and reality might actually intensify their pain. Reducing this gap, this “surprise” factor, could potentially help reduce pain intensity.
The researchers conducted a follow-up experiment where participants just passively watched the knife movement instead of controlling it themselves. Interestingly, the timing and disappearance effects nearly vanished. This highlights how important our sense of agency, feeling in control of our actions, is to how we anticipate and experience pain.
From an evolutionary standpoint, this heightened sensitivity to surprising pain makes perfect sense. Unexpected pain often signals danger or injury requiring immediate attention. By amplifying surprising pain, our brain ensures we pay extra attention to threats we failed to predict. This mechanism probably helped our ancestors survive by making unexpected painful experiences more memorable and significant.
How our brain predicts pain
This research fits into the broader “Bayesian brain” theory, which proposes that our brains constantly generate predictions about the world and update them based on what actually happens. The research team suggests pain perception follows this same pattern, with our expectations continuously updating in real-time as we interact with our environment.
This might explain why paper cuts feel so awful when they catch you by surprise, or why an anticipated injection at the doctor’s office sometimes hurts less than you expected. It might also explain why distractions can reduce pain. Focusing elsewhere might reduce attention to pain-related prediction errors.
For doctors and pain specialists, these findings suggest that managing patients’ expectations and reducing uncertainty could be effective pain management strategies. Simply giving clear information about when and how pain might occur during medical procedures might help reduce pain intensity.
Could some chronic pain sufferers have a dysfunction in their brain’s prediction system? Are they experiencing constant “Bayesian surprise” due to problems with how they process sensory information? These questions could lead to completely new treatment approaches. If pain is indeed amplified by prediction errors, then therapies that reduce these errors or change how the brain responds to them might provide relief.
The next time you experience an unexpected painful stimulus, remember that your brain’s prediction system is actively shaping how intense that pain feels. Your pain isn’t just a direct reflection of the physical stimulus; it’s a complex construction shaped by your expectations, your sense of control, and how surprised your brain is by what actually happens. In the world of pain perception, surprise really does pack a punch.
Paper Summary
Methodology
The researchers conducted two experiments using virtual reality combined with heat stimulation. In the first experiment, 23 participants controlled a virtual knife with their left hand and stabbed it into their virtual right forearm. Meanwhile, a device applied heat to their actual arm at different timings—either simultaneously with the virtual knife contact or delayed by about a second. The knife either remained visible or suddenly vanished just before contact. After each trial, participants rated their pain intensity. In the second experiment, 26 different participants observed the knife movement without controlling it themselves, allowing researchers to examine how the sense of control influences pain perception.
Results
The key findings were clear: pain intensity was significantly higher when heat stimulation was delayed compared to when it was simultaneous with knife contact, and this effect was even stronger when the knife unexpectedly vanished. These effects only appeared in the group who controlled the knife themselves, not those who merely observed it. Mathematical analysis strongly favored the “Bayesian surprise” hypothesis over the “Bayesian estimate” hypothesis, suggesting that pain perception is based more on how much a stimulus violates our expectations than on a simple integration of what we predict and what we feel.
Limitations
The study has several limitations: it was conducted on healthy volunteers in a controlled setting rather than people with chronic pain conditions; it used only heat stimulation rather than other types of pain; and the virtual reality environment, while immersive, differs from real-world experiences. The research also doesn’t account for long-term adaptation to repeated pain stimuli, and further neuroimaging studies would be needed to identify the specific brain mechanisms involved.
Discussion and Takeaways
The findings suggest that pain perception follows different principles than some other sensory experiences—while we perceive things like object size as a blend of predictions and sensory input, we seem to experience pain more as a function of prediction error or “surprise.” This may serve an important evolutionary purpose by highlighting potentially dangerous unexpected stimuli. The researchers suggest that their findings could lead to new approaches to pain management that focus on reducing prediction errors or changing how the brain responds to them, which could be particularly valuable for treating chronic pain conditions.
Funding and Disclosures
The research was supported by JSPS KAKENHI (grant numbers 19H05729 and 23KJ0261), which are grants from the Japan Society for the Promotion of Science. The authors declared no competing interests. The study was approved by the Institutional Review Board at the University of Tsukuba, and all participants provided informed consent.
Publication Information
This study, “Bayesian surprise intensifies pain in a novel visual-noxious association,” was published in Cognition (Volume 257, 2025, Article 106064) and was available online as of January 16, 2025. The paper was authored by Ryota Ishikawa, Genta Ono, and Jun Izawa from the University of Tsukuba in Japan. The article is available under the CC BY license.
