Bayes goes tango: Twenty years of experience alter joint walking dynamics, but not the use of multisensory information
This study demonstrates that while twenty years of tango experience significantly alters joint walking coordination dynamics, it does not change how dancers integrate multisensory information, suggesting that Bayesian multisensory integration is a fundamental sensorimotor mechanism rather than a skill acquired through practice.
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 This is an AI-generated explanation of a preprint that has not been peer-reviewed.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
Every day, we move in sync with others without thinking about it. We step aside to let someone pass, walk in step with a friend, or navigate a crowded room without bumping into anyone. These moments of shared movement rely on a complex, split-second conversation between our eyes, our ears, and our sense of touch. Our brains must constantly guess where another person is, how fast they are moving, and what they will do next, all while dealing with the fact that our senses are not perfect. They are noisy and sometimes slow. To make sense of this uncertainty, the human brain has a built-in strategy: it combines information from different senses, weighing the most reliable clues more heavily to form a single, clear picture of the world. This process, known as multisensory integration, allows us to act smoothly even when the world is confusing. But a question remains: does this strategy change as we get better at a specific task? If we spend decades mastering a complex form of movement, do we learn to rely on different senses, or does our brain stick to the same fundamental rules?
To answer this, researchers at the University of Bern turned to the Argentine tango, a dance style that relies entirely on improvisation rather than memorized steps. In this dance, one person leads and the other follows, with the follower needing to react instantly to the leader's subtle movements. The researchers recruited 48 women who had been dancing tango for varying amounts of time, ranging from beginners with less than a year of experience to experts with over twenty years of practice. They asked these women to perform a simplified version of the dance: walking back and forth in time with music, facing a leader who guided them with subtle cues. The goal was to see how the followers' movements changed as their expertise grew, and how they coped when the researchers took away specific sensory clues.
The experiment was designed to test the limits of the followers' senses. In the standard version of the task, the follower could see the leader, hear the music, and feel the leader's hand on their chest. In other versions, the researchers removed one of these clues at a time. Sometimes the follower wore opaque goggles to block vision. Other times, they were asked to keep their arms at their sides so they could not feel the leader's touch. In a third variation, the leader wore headphones so the follower could not hear the music. By measuring the precise movements of the dancers' chests with high-speed cameras, the researchers could calculate exactly how well the follower matched the leader's speed and timing.
The results revealed a fascinating difference between how experts and beginners moved. The women with the most experience did not simply copy the leader more perfectly; instead, they moved with a more dynamic, spring-like quality. They seemed to wait a fraction of a second longer to gather information before reacting, and then they accelerated more sharply to catch up. This "wait and surge" pattern was a hallmark of the experts, distinguishing them from the beginners who moved more cautiously. However, when the researchers looked at how stable the coordination was—how much the timing and speed varied from step to step—they found that experience did not make the dancers more consistent. The stability of the movement depended almost entirely on which senses were available, not on how many years the dancer had spent practicing.
The most critical finding concerned which sense mattered most. When the researchers blocked the sense of touch by having the dancers walk without holding the leader, the coordination fell apart more than when vision or hearing was blocked. This confirmed that the physical connection between partners is the most vital clue for this type of movement. Surprisingly, the researchers found that the way the dancers used these clues did not change with experience. Even after twenty years of dancing, the experts relied on touch, sight, and sound in the exact same proportions as the beginners. They did not learn to ignore the music or the visual cues in favor of touch; they simply became better at using all of them together.
This lack of change suggests that the brain's method for combining sensory information is a fundamental, unchangeable rule of how humans move, rather than a skill that is learned through practice. The researchers tested this idea using a mathematical model based on how the brain weighs uncertain information. They found that the dancers' behavior was well-captured by this model, matching the predictions of Bayesian principles closely. While there were systematic deviations in the data—likely due to motor noise added after the brain combined the senses—the overall pattern confirmed that the brain acts like a smart calculator, blending the available clues to reduce error. When a clue is removed, the brain relies more heavily on the remaining ones, but the underlying rule for how to mix them stays the same. The study concludes that while we can learn to move more dynamically and expressively through years of practice, the basic sensorimotor mechanism that allows us to coordinate with others remains constant. We do not rewire our brains to become better dancers; we simply get better at applying the same, ancient rules of perception to a complex new challenge.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.