Andrew D. Straw

I study how neurons work together to create useful behavior. A highly quantitative analysis of animal behavior is a particularly powerful technique in neuroscience--not only is behavior a product of the evolution of an animal and its nervous system, it can provide the experimenter with a robust, if indirect, readout of the simultaneous activity of motor neurons and muscles.

I study the control of behavior in animals from 3 perspectives:

• The algorithmic perspective. I seek to discover what rules govern motor output based on sensory input and internal state.
• The implementation perspective. What neurons, neural circuits, muscles and other structures implement the algorithms and how?
• The engineering perspective. Because mathematical models and robotic platforms are much simpler than animals, they allow highly instructive analysis. Yet the robust, multi-level control of animals is beyond the capabilities of engineered systems -- engineers have much to learn from biologists, much as biologists have much to learn from engineers.