A new study suggests that successful coordination between people may begin not with copying another person’s movements, but with accurately predicting one’s own actions. When those self-predictions become aligned with the actions unfolding in a partner, smooth coordination can emerge—even when the interaction is continuous, unpredictable and constantly changing. The finding offers a new explanation for how people dance together, pass objects, play music, cooperate in sports and perform delicate tasks without needing to consciously negotiate every movement. It also provides a potential framework for understanding why coordination sometimes feels effortless with one person but awkward with another.
Published in Communications Psychology, the study by Putra and Kano investigates coordination as a dynamic motor process rather than a simple matter of synchronizing visible movements. In a static task, two people might be judged coordinated if their hands reach the same position at the same time. Real interactions are more demanding. Each person must move while observing the other, anticipate what may happen next, compensate for delays in perception and movement, and continuously revise their actions as the situation changes. The researchers’ central argument is that coordination depends on how each individual predicts the sensory consequences of their own movement, and how closely those predictions become aligned with the evolving interaction.
The idea is rooted in the brain’s internal models of action. Every voluntary movement generates predictions about what the body should feel and see next: where the hand will be, how quickly it will travel, what force will be required and how the external world will respond. These predictions are compared with incoming sensory information from vision, touch and proprioception, the sense of the body’s position in space. When there is a mismatch, the nervous system updates its motor commands. This process occurs rapidly and largely outside conscious awareness. In a social interaction, however, the predicted consequences of one person’s movement can also become relevant to another person’s behavior. A partner’s action changes the environment in which the next movement must be made.
The study’s important contribution is its focus on “self-predictions.” People do not merely predict what a partner will do; they also predict how their own body will move in response to the partner and the surrounding task. If one person expects to accelerate, slow down or change direction at a particular moment, that expectation shapes the motor commands issued by the brain. Coordination improves when these predictions are compatible with the partner’s actual behavior and with the shared dynamics of the task. Put simply, two people can become coordinated when each person’s internal forecast of their own next movement fits the interaction well enough for the other person to anticipate it too.
This perspective challenges the intuitive idea that coordination is mainly a process of observing another person and imitating them. Visual matching can help, but it is often too slow and too fragile for real-time interaction. Sensory information arrives with delays, and human movement is inherently variable. If people waited for a partner’s action to be completed before responding, coordinated behavior would quickly break down. Predictive control solves part of this problem by allowing the brain to act ahead of incoming feedback. The individual estimates how the interaction is likely to evolve, prepares a response and then uses sensory error to refine it. When the two people’s predictive processes converge, their movements can stabilize into a shared pattern.
The researchers examined this process in a dynamic motor setting, where participants had to coordinate movement while the interaction continued to change. Such tasks are more revealing than simple rhythmic tapping because they require participants to manage evolving trajectories, timing and feedback. Performance can be assessed not only by whether movements coincide, but also by the degree of temporal delay, spatial error, variability and adaptation between partners. Computational approaches can then estimate how strongly a participant’s movement reflects an internal prediction versus a direct reaction to observed sensory input. The results indicate that high coordination was associated with the alignment of these self-generated predictions, rather than with mechanical synchrony alone.
That distinction is technically significant. Two people may display similar movements while relying on very different control strategies. One may closely track the other’s visible behavior, producing a delayed imitation. Another may anticipate the partner’s movement and adjust before the sensory evidence is complete. The second strategy can produce more stable coordination, especially when the interaction contains perturbations or sudden changes. In control-theory terms, the nervous system is not simply minimizing an immediate positional error; it is estimating the future state of a coupled system that includes the body, the partner and the environment. Coordination therefore emerges from reciprocal prediction and correction across time.
The findings may also explain why coordination can improve rapidly without explicit instruction. During repeated interaction, each person gathers information about the other’s timing, movement tendencies and responses to unexpected events. Those regularities become incorporated into the person’s internal model. The resulting predictions do not need to be identical. Effective coordination may instead depend on their being mutually compatible: one person’s expected movement creates conditions that support the other’s expected movement, while each remains flexible enough to adjust when the situation changes. This could explain why experienced partners often appear to “read” each other’s intentions, even when neither can describe exactly how that understanding occurs.
The implications extend beyond ordinary social behavior. A predictive account of coordination could inform the design of collaborative robots that move safely and naturally alongside people. Rather than programming a machine simply to copy human trajectories, engineers could build systems that estimate how a person expects their own movement to unfold and adapt the robot’s behavior accordingly. The same principles may be relevant to rehabilitation, where patients relearn motor control through interaction with therapists or assistive devices. They may also help researchers study conditions in which predictive motor coupling is disrupted, including neurological disorders that affect timing, sensory integration or the ability to use feedback effectively.
The study does not suggest that people consciously calculate one another’s movements or that prediction alone guarantees perfect cooperation. Coordination remains vulnerable to noise, fatigue, unfamiliar movement patterns and conflicting goals. Instead, the work identifies a hidden mechanism that can make joint action robust: the gradual alignment of internal forecasts about what each person’s own body is about to do within a shared, changing environment. The broader message is that social coordination may be constructed from individual predictions. When those predictions become sufficiently compatible, a joint pattern of action can arise without a central controller, explicit commands or exact imitation—turning separate nervous systems into a temporarily coupled motor system.
Subject of Research: Human motor coordination, predictive sensorimotor control and dynamic interpersonal interaction
Article Title: Successful coordination emerges from aligned self-predictions in a dynamic motor interaction
Article References: Putra, P.U., Kano, F. Successful coordination emerges from aligned self-predictions in a dynamic motor interaction. Communications Psychology 4, 119 (2026). https://doi.org/10.1038/s44271-026-00523-7
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s44271-026-00523-7
Keywords: motor coordination, self-prediction, predictive processing, sensorimotor control, interpersonal interaction, motor synchronization, computational neuroscience, social cognition
