Multiplicative nonlinearity in the perception of apparent motion

Area of Science:

• Visual perception
• Computational neuroscience
• Motion detection

Background:

• Apparent motion perception is crucial for understanding visual processing.
• Existing models of motion detection, like the Reichardt detector, often rely on asymmetrical delays.
• The precise mechanisms underlying motion perception are still under investigation.

Purpose of the Study:

• To investigate the computational mechanisms underlying apparent motion perception.
• To compare different models of motion perception, including summation, squared-sum, motion energy, and multiplicative combination of luminance changes.
• To propose a modified motion detector model that accounts for new findings.

Main Methods:

• Stimuli with simultaneous luminance changes at two element locations were used.
• The predictive power of different computational models (sum, squared-sum, motion energy, multiplicative combination) was evaluated against perceived motion.
• A modified Reichardt-style motion detector model was developed.

Main Results:

• The multiplicative combination of luminance changes better predicted apparent motion perception than summation, squared-sum, or motion energy.
• The findings support a Reichardt-style motion detector but suggest modifications to the original model.
• Motion direction is determined by asymmetrical stimulus information rather than asymmetrical delay.

Conclusions:

• A multiplicative model of luminance changes provides a superior account of apparent motion perception.
• A modified Reichardt detector, relying on asymmetrical stimulus information for direction, is proposed.
• This research offers new insights into the neural computations underlying visual motion detection.