Reference frames reverse biases in human heading perception

When we move through the world, our eyes do not just see static pictures; they see a rushing stream of visual information known as optic flow. Imagine standing on a train platform as a train pulls away: the tracks and the station walls seem to streak outward from a single point in front of you. This visual pattern is the primary clue our brains use to figure out which direction we are traveling, a sense scientists call heading. For decades, researchers have tried to map exactly how the brain calculates this direction, but they have hit a confusing wall. Some experiments suggested that when people guess their direction of travel, they tend to aim slightly toward the center, as if pulled inward. Other experiments, using different setups, found the opposite: people seemed to aim slightly away from the center, pushed outward. This contradiction left scientists wondering if the brain was simply unreliable or if something about the way the experiments were designed was tricking the participants.

A team of researchers led by Renate Reisenegger at the University of Marburg has now proposed a solution to this decades-old puzzle. They suggest that the brain is not inconsistent; rather, the answer depends entirely on how the person is asked to report their direction. The study, published in a recent issue of a scientific journal, reveals that the mental perspective we adopt—whether we think about our movement from our own point of view or from a detached, map-like view—fundamentally changes the direction of the error. By testing the same people with the exact same visual scenes but asking them to answer in two different ways, the researchers showed that the brain can be pulled inward or pushed outward simply by changing the rules of the game.

To test this idea, the team built a unique laboratory setup using a massive, curved screen that wraps around the viewer, covering a field of view of about 224 degrees. This "edgeless" display allowed participants to see a wide, immersive simulation of walking across a flat ground made of moving dots, without any visible borders to distract them. In the first part of the study, thirty-seven volunteers watched these short clips of self-motion and were asked to report where they thought they were heading. The researchers split the task into two distinct modes. In the first mode, the "egocentric" condition, participants reported their direction from their own perspective, pointing to a spot on a horizontal line that represented their immediate view. In the second mode, the "allocentric" condition, participants were asked to imagine they were looking down from above, like a bird, and to point to their direction on a circular arc or a full circle.

The results were striking and immediate. When participants reported from their own perspective, their guesses consistently drifted toward the straight-ahead direction. If they were actually moving slightly to the left, they reported moving even closer to the center. This is known as a centripetal bias. However, when the very same people switched to the "bird's-eye" view and reported in the allocentric frame, the bias flipped completely. Now, their guesses drifted away from the center. If they were moving slightly to the left, they reported moving even further to the left. This centrifugal bias appeared even though the visual stimulus was identical and the range of possible answers was the same in both cases. The researchers also tested whether the size of the answer range was the cause, as some previous theories had suggested, but found that changing the size of the circle did not change the direction of the bias. The only thing that mattered was the frame of reference.

To understand what was happening inside the brain during this switch, the team conducted a second experiment with twenty-one different volunteers. This time, they used standard computer monitors and recorded the participants' brain activity using electroencephalography, or EEG, which measures electrical signals on the scalp. The participants watched the same moving dots, but instead of pointing to a direction, they had to decide quickly if a brief flash appeared to the left or right of their remembered heading. Crucially, a cue appeared before each trial telling them whether to use their own perspective or the bird's-eye view. The researchers focused on the moment between the cue and the answer, a time when the brain was preparing to transform the visual information.

The brain scans revealed a clear difference in activity depending on the task. When participants prepared to answer in the bird's-eye view, their brains showed a significant increase in rhythmic electrical activity in the theta and alpha frequency bands, specifically over the back and top of the head. This pattern of activity is often associated with brain regions involved in navigation and spatial memory, such as the retrosplenial cortex and the hippocampus. In contrast, when they prepared to answer from their own perspective, this extra surge of activity was absent. This suggests that the brain has to work harder to perform the mental gymnastics required to switch from a self-centered view to a world-centered map, and that this extra processing step is likely what introduces the outward bias.

Finally, the researchers used a computer model based on how the brain combines sensory information with past expectations to explain these findings. The model showed that the brain's internal "map" of what is likely to happen changes depending on the task. When people think about their movement from their own perspective, they rely heavily on the specific range of movements they just saw, which pulls their estimate toward the center. But when they switch to the bird's-eye view, the brain seems to adopt a broader, flatter expectation, assuming that movement could be happening in any direction across the full circle. This shift in expectation, combined with the brain's natural sensitivity to straight-ahead movement, creates the outward push. The study concludes that the conflicting results in the history of heading research were not due to a flaw in human perception, but because different experiments had unknowingly asked the brain to solve the problem in two different coordinate systems. By recognizing that the reference frame itself drives the bias, the team has resolved a long-standing mystery in vision science, showing that how we ask a question can fundamentally alter the answer our brain gives.