Self-motion direction estimation from optic flow is a result of capacity-free and implicit ensemble coding.

Numerous studies have explored the mechanisms of heading estimation from optic flow and ensemble coding in other features, yet none have examined ensemble coding's role in heading estimation. This study addressed this gap through two experiments. Participants sequentially viewed three (experiment 1) or five/seven (experiment 2) optic flow-simulated headings, then reported specific directions. Results revealed that individual heading accuracy declined with increasing numbers, while estimates closely matched ensemble representations, demonstrating ensemble coding in heading estimation. Notably, ensemble coding accuracy remained unaffected by heading quantity, indicating its capacity-free nature-unlike capacity-limited individual heading processing. The discovered summary statistics of motion may help us to better understand the navigation in complex environments (e.g., how pedestrians/drivers judge their self-motion directions), which could potentially contribute to real-world implications.