The perception of globally coherent motion

Area of Science:

• Visual perception
• Motion perception
• Computational neuroscience

Background:

• Human observers integrate local visual cues to perceive global motion.
• Understanding the spatial integration of ambiguous motion signals is crucial for visual perception research.
• Previous models, like intersection-of-constraints, have been proposed to explain motion perception.

Purpose of the Study:

• To investigate how human observers perceive coherent motion from disparate local motion measures.
• To examine the spatial integration of ambiguous motion signals along contours for global motion perception.
• To determine the underlying pooling process in global motion perception.

Main Methods:

• Observers viewed displays with numerous apertures, each containing contours with specified orientations and velocities.
• Contour trajectories were manipulated to create globally coherent motions (rotations, expansions, translations).
• Analysis of observer reports on global motion direction, considering contour orientation sampling and feature presence.

Main Results:

• Observer reports of global motion direction were biased with asymmetric contour orientation sampling.
• Performance improved with the presence of identifiable features (line ends, crossings) for trajectory tracking.
• Results align with a vector average pooling process, not intersection-of-constraints analysis.

Conclusions:

• Human visual system integrates local motion signals through a vector averaging mechanism.
• Asymmetric sampling of contour orientations introduces bias in global motion perception.
• Identifiable features enhance the accuracy of global motion perception by aiding trajectory tracking.