Modelling fast forms of visual neural plasticity using a modified second-order motion energy model.

Area of Science:

• Computational neuroscience
• Human visual perception
• Sensory adaptation modeling

Background:

• The Adelson-Bergen motion energy sensor is a leading model for low-level visual motion sensing.
• The standard model fails to predict adaptation effects in motion perception.
• Previous extensions addressed long-timescale adaptation but not rapid adaptation.

Purpose of the Study:

• To extend the Adelson-Bergen motion energy sensor model to incorporate rapid adaptation.
• To account for psychophysical data on short-timescale motion perception phenomena.
• To provide a computational framework for multi-timescale adaptation in visual motion sensing.

Main Methods:

• Incorporated a second-order RC circuit into the Adelson-Bergen motion energy sensor model.
• Simulated the model's response to sudden changes in visual stimulation.
• Compared model outputs with psychophysical data for rapid visual motion priming (rVMP) and rapid motion after-effect (rMAE).

Main Results:

• The extended model accurately predicts psychophysical data for rapid adaptation effects.
• The second-order RC circuit introduces a finite reaction time, crucial for modeling rapid adaptation.
• Model outputs align with observed facilitation and suppression in rapid motion perception.

Conclusions:

• The enhanced motion energy sensor model successfully captures rapid adaptation phenomena in human vision.
• Multi-stage leaky integrator circuits provide a viable computational scheme for modeling adaptation across multiple timescales.
• This work advances computational models of sensory adaptation in visual motion processing.