Animals extract information from the world to direct behavior. The computations by which neural circuits collect sensory information with imperfect sensors and convert this information into behavioral programs remain unknown. My lab seeks to reveal building blocks of neural computation underlying sensory perception and adaptive decision making. We study a sensorimotor integration task common to all living organisms: chemotaxis (orientation behavior in response to chemicals). We tackle this problem in the Drosophila melanogaster larva, one of simplest genetic model organisms that demonstrate robust navigation under the control of a brain composed of less than 10,000 neurons. Our research aims to achieve data-driven models of the neural mechanisms that convert dynamic sensory signals into orientation responses. Combining neuronal imaging and perturbation analysis through optogenetics, we generate mechanistic hypothesis about the neural implementation of navigational decisions. By closing the loop between theory and experimentation, we intend to iteratively refine our models, thereby unraveling mechanisms of neural-circuit computations that direct active sensing and decision making. We are also interested in exploring the diversity of chemotactic algorithms in different species of the Drosophila group to establish general design principles between the structure and the function of sensorimotor circuits.