Reinforced diffusions as models of memory-mediated animal movement

Area of Science:

• Ecology and mathematical modeling of animal behavior.

Background:

• Understanding how memory shapes animal movement is crucial for ecological studies, conservation planning, and climate change adaptation.
• Existing research indicates memory in animal movement has temporal and spatial dimensions, incorporating both attractive and aversive elements.

Purpose of the Study:

• To introduce reinforced diffusions as a novel modeling framework to investigate the role of memory in animal movement patterns.
• To explore how memory, particularly central place attraction, influences space usage and generates diverse movement behaviors.

Main Methods:

• Developed a reinforced diffusion model, a continuous-time extension of reinforced random walks.
• Incorporated memory reinforcement based on time and distance from the current location.
• Analyzed movement patterns resulting from the interplay of memory and central place attraction.

Main Results:

• Identified three distinct movement behaviors: non-collapsing bounded wandering, spatial collapse to a small area, and convergence to cyclical paths.
• Demonstrated the model's ability to emulate real-world animal movement learning, such as homing pigeon route learning and ant path consolidation.

Conclusions:

• Reinforced diffusions provide a mathematically explicit framework for studying memory structures in animal movement.
• The model links abstract memory assumptions to biological concepts like immediate surroundings and memory decay, offering insights into animal navigation and spatial memory.