Single neurons in the human substantia nigra encode social learning signals
This study demonstrates that single dopaminergic neurons in the human substantia nigra encode social learning signals, specifically tracking the valence of norm prediction errors with greater activity during interactions with humans compared to computers.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Human beings are wired to navigate a complex web of unwritten rules. We know that in a group, certain behaviors are expected, and when someone breaks those expectations, we feel it. This ability to sense when a social offer is unfair or when a norm has been violated is a cornerstone of how we interact. For decades, scientists have understood that the brain uses a system of prediction and error correction to learn these rules. When we expect a certain outcome and get something different, our brains register a "prediction error," a signal that helps us update our understanding of the world. While we know this happens in the brain, we have largely been guessing about the specific machinery involved, relying on blurry images of brain activity or studies in animals. The question of how a single human neuron reacts to a social slight has remained out of reach, hidden deep within the brain's core.
A team of researchers has now looked directly at this machinery by recording the activity of individual neurons in the brains of people undergoing brain surgery. They focused on a tiny, deep structure called the substantia nigra, a region known for its role in movement and reward. By listening to the electrical signals of single cells while patients played a game of fairness, the scientists discovered that these neurons do more than just track money or movement; they specifically track the feeling of being treated fairly or unfairly by other people. The study suggests that the brain's reward system is deeply entangled with our social lives, firing up when we receive a better offer than expected from a human, but staying quiet when the same offer comes from a computer.
The experiment took place in an operating room at the Icahn School of Medicine at Mount Sinai. The participants were patients with Parkinson's disease who were undergoing surgery to implant a device to help control their symptoms. Before the doctors implanted the final device, they paused to perform a unique research task. The patients, who were awake and alert, played a version of the Ultimatum Game. In this game, a partner proposes how to split a sum of twenty dollars. The player can either accept the split, in which case both parties get the money, or reject it, in which case neither gets anything. The offers ranged from one to nine dollars, and the partners were either other people represented by avatars or a computer program.
While the patients made their choices, the researchers recorded the electrical activity of individual neurons in the substantia nigra. They also recorded activity in a neighboring area called the globus pallidus to see if the effect was specific to the substantia nigra or a general response across the brain. To understand what the patients were thinking, the researchers used a computer model to estimate the "norm prediction error" on every single turn. This is a measure of how much the offer differed from what the player expected to be fair. If a player expected a five-dollar offer and got seven, that is a positive error. If they got two, that is a negative error. The model also tracked whether the player was playing against a human or a computer.
The results showed a clear and specific pattern. When the patients played against human avatars, the neurons in the substantia nigra lit up in response to the fairness of the offer. Specifically, these cells fired more rapidly when the offer was better than expected and less when it was worse. This signal was about the feeling of the offer being fair or unfair, not just the amount of money. Crucially, this reaction was much stronger when the partner was a human. When the same players faced a computer making the exact same offers, the neurons did not show this same sensitivity to fairness. The cells seemed to care deeply about the social context, distinguishing between a human judgment and a machine calculation.
In contrast, the neurons in the globus pallidus, the neighboring brain region, did not show this pattern at all. They did not track the fairness of the offer, nor did they react differently to human versus computer partners. This finding suggests that the ability to encode social learning signals is a specific job of the substantia nigra, not a general function of the entire area. Furthermore, the researchers checked if these neurons were simply reacting to the patient's mood. They found that while the patients felt worse after rejecting an unfair offer, the neurons themselves did not track these mood changes. The cells were strictly focused on the social prediction error, the gap between what was expected and what was received.
The study also revealed that the patients had different expectations depending on who they were playing. Even though they rejected unfair offers in both scenarios, the computer model showed that they expected higher offers from humans than from computers. They held a higher standard of fairness for people. The neurons in the substantia nigra tracked this specific social nuance, firing in a way that reflected the player's internal sense of what was fair for a human partner versus a machine. This work provides a rare, direct look at how the human brain processes social rules at the most fundamental level, showing that the cells responsible for learning from rewards are also the ones helping us navigate the complex world of human interaction.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.