Modeling approaches to understand odor-guided locomotion

In addition to behavioral and physiological studies on male and female moths, I have ongoing collaborations with Dr. Danny Grünbaum and Dr. Jim Belanger to study odor-guided navigation using simulation and analytical modeling approaches. These modeling experiments have resulted in new insights into how moths and other animals are able to locate distant unseen sources using plumes of odorant molecules, and have suggested several behavioral experiments using real moths. We are also working toward implementing the successful simulation models as control systems for odor-guided robots with the collaboration of Dr. Wayne Jouse from the Department of Aerospace and Mechanical Engineering at the University of Arizona.

Many animals, ranging from bacteria, through nematodes, to insects, fish and mammals, use air- or water-borne plumes of odor molecules to locate distant unseen resources. Remarkably, the movement tracks that many species produce when tracking odor plumes have a very similar side-to-side zigzag shape whether walking, swimming, or flying.

Simulating Moth Tracking Behavior

Such similarity across very different species and conditions suggests that there may be something fundamentally "good" about this approach to tracking chemical plumes (however, similar looking behavior does not necessarily indicate similar underlying mechanisms). Considerable amounts of information exist for how many types of animals track odor plumes (for example: nematode worms or crabs and lobsters). However, more is known about how male moths locate a source of female sex-attractant pheromones, from odor detection at the receptor level to the motor outputs that constitute behavior, than any other animal. This makes moth orientation to odor plumes an ideal system in which to study the fundamental "rules" that enable these animals to orient to and locate distant unseen sources of chemicals.

From the beginning our goal has not been to make a computer model that re-creates the moth behavior that we study. Rather the goal has been to construct models based on what we know about how moths control their behavior to see how good these models are at locating odor sources in our simulation's environment. This approach allows us to test our ideas and hypotheses about moth orientation in ways that would be impossible using real moths. Importantly, we learn at least as much when the simulated moths fail to locate the odor source as when they are successful. Because our simulation models generate "flight" tracks just like real moths we are also able to analyze how and why the simulations fail, and compare the model's performance to real moths. [Dr. Belanger's original model was constructed in collaboration with Drs. Ed Arbas and Mark Willis in the Arbas laboratory. The Belanger/Willis collaboration has continued fruitfully since Arbas' death in 1995.]

The model moth in the simulation was constructed based on what is known about the sensory and control systems of real moths so we could examine more fully the complex interaction between the odor stimulus, sensory processing, interacting control systems, and ongoing centrally organized behavior. The simulation environment is simple and flexible and can be made to reflect the changing conditions of real environments.

<<Previous   |   Next>>