A Normalization Mechanism for Estimating Visual Motion across Speeds and Scales

Area of Science:

• Neuroscience
• Computational Neuroscience
• Visual Perception

Background:

• Visual speed estimation is crucial for survival in many species.
• Mammals and primates use spatiotemporal channels to process optic flow.
• The brain's mechanism for inferring speed from complex natural scenes remains unclear.

Purpose of the Study:

• Investigate how the brain estimates visual speed from complex naturalistic stimuli.
• Explore the role of spatiotemporal channels in speed perception.
• Develop a computational model to explain observed speed perception phenomena.

Main Methods:

• Utilized "compound motion clouds" (CMCs), naturalistic textures with controlled motion.
• Presented CMCs to participants and recorded perceived speed variations.
• Developed a computational model simulating channel interactions and normalization.

Main Results:

• Perceived speed of CMCs varied based on activated channel combinations, despite identical physical speeds.
• The computational model demonstrated a speed-scale normalization mechanism.
• This mechanism involves boosting and weakening channel activity based on speed and scale, including lateral inhibition.

Conclusions:

• The visual system employs a speed-scale normalization mechanism to interpret complex visual input.
• This mechanism integrates information across speed and scale channels, potentially involving lateral inhibition.
• This normalization aids in creating a coherent visual perception of speed in natural environments.