Modeling and Simulation of Kairomone-Mediated Host Search in a Parasitic Isopod
Mentor 1
Peter Hinow
Start Date
10-5-2022 10:00 AM
Description
Many animals live in complex olfactory landscapes, exploiting chemical cues to locate food, mates and avoid predators. An example are small aquatic organisms with limited vision. When searching for a potential prey they can employ a gradient or plume of a chemical that is a signature of the prey. We are interested in the behavior of a parasitic isopod (1 mm long) in search of a coral reef fish, from which it needs a blood meal. We create and simulate an agent in order to investigate their respective advantages or disadvantages. Our thinking is inspired by the theory of Braitenberg vehicles, simple agents that can react to input of sensors. We use the Open-Source software Simbrain, which is a powerful tool for design and simulation of neural networks and agents equipped with such networks moving in virtual 2d landscapes. Simbrain can simulate an agent and a smell source. The simulated agent is equipped with two sensors that it uses to detect the chemical gradient and pursue the prey. The agent consists of a simple neural network. Several scenarios are simulated with variable parameters. The simulation consists of a simple neural network which guides the Braitenberg vehicle according to input from two sensors. Simbrain runs large numbers of trials and collects data on the success rate, path trajectory, and the time taken to reach the prey. This will provide insight into how isopods find prey. Further research could include optimizing more complex neural networks for which adaptations produce the greatest success in the host search hunt.
Modeling and Simulation of Kairomone-Mediated Host Search in a Parasitic Isopod
Many animals live in complex olfactory landscapes, exploiting chemical cues to locate food, mates and avoid predators. An example are small aquatic organisms with limited vision. When searching for a potential prey they can employ a gradient or plume of a chemical that is a signature of the prey. We are interested in the behavior of a parasitic isopod (1 mm long) in search of a coral reef fish, from which it needs a blood meal. We create and simulate an agent in order to investigate their respective advantages or disadvantages. Our thinking is inspired by the theory of Braitenberg vehicles, simple agents that can react to input of sensors. We use the Open-Source software Simbrain, which is a powerful tool for design and simulation of neural networks and agents equipped with such networks moving in virtual 2d landscapes. Simbrain can simulate an agent and a smell source. The simulated agent is equipped with two sensors that it uses to detect the chemical gradient and pursue the prey. The agent consists of a simple neural network. Several scenarios are simulated with variable parameters. The simulation consists of a simple neural network which guides the Braitenberg vehicle according to input from two sensors. Simbrain runs large numbers of trials and collects data on the success rate, path trajectory, and the time taken to reach the prey. This will provide insight into how isopods find prey. Further research could include optimizing more complex neural networks for which adaptations produce the greatest success in the host search hunt.
