# Distinct temporal dynamics of judging scene-relative object motion and estimating heading from optic flow

## TL;DR

The study shows that judging object motion and estimating self-movement direction from visual flow change differently over time.

## Contribution

The paper reveals distinct temporal dynamics between flow parsing and heading estimation using matched visual displays.

## Key findings

- Flow parsing accuracy decreases with longer stimulus durations.
- Heading estimation accuracy increases with longer stimulus durations.
- The processes involve distinct mechanisms with different temporal characteristics.

## Abstract

During self-movement, the visual system can identify scene-relative object motion via flow parsing and estimate the direction of self-movement (heading) from optic flow. However, the temporal dynamics of these two processes have not been examined and compared using matched displays. In this study, we examined how the accuracy of flow parsing and heading estimation changed with stimulus duration. Participants viewed a stereo optic flow display simulating forward translational self-movement through a cloud composed of wireframe objects with stimulus durations at 100, 200, 400, 700, and 1000 ms. In Experiment 1, a yellow dot probe moved vertically for 100 ms in the scene near the end of the trial. A nulling motion component was added through an adaptive staircase to the probe's image motion to determine when the probe was perceived to move vertically in the scene, which was then used to compute the accuracy of flow parsing. In Experiment 2, participants viewed the same optic flow display without the moving probe object. The simulated heading was randomly varied in each trial, and participants were asked to estimate heading at the end of the trial. As stimulus duration increased, the accuracy of flow parsing decreased, whereas the accuracy of heading estimation increased. These contrasting temporal dynamics suggest that despite both processes relying on optic flow, flow parsing and heading estimation involve distinct processing mechanisms with different temporal characteristics. This divergence, together with previous neurophysiological findings, led us to propose two potential neural mechanisms subserving these two processes to inspire future research.