Project Summary Animals interact with the world through dynamic, iterative sensory-motor processes that guide their ongoing movement. Odor-guided navigation is the basis for fundamental natural behaviors such as finding food sources, but little is known about the nature of the sensory signals that inform adaptive changes in locomotion. Here we propose to test how spatial information is encoded by distributed activity in the olfactory bulb, and how this information is decoded by higher-order brain areas. Olfactory sampling occurs through ‘sniffs’ – discrete sensory samples thought to provide a snapshot of the current odor environment. One possibility is that animals ascend concentration gradients by comparing intensity across successive sniffs. Another is that directional information may be present in left-right ‘stereo’ comparisons between hemispheres. To address this issue, we use new imaging tools to visualize large- scale activity dynamics over both hemispheres of olfactory bulb in mice engaged in odor-guided search. We also characterize functional computations in the downstream circuits that decode this spatial information, using whole cell recordings and optogenetic stimulation to establish rules for bilateral synaptic integration, as well as high-resolution in vivo photostimulation mapping to measure cross- hemisphere interactions in the intact circuit. Finally, we use optogenetic activity manipulations during odor-guided search to test the causal role of cross-sniff vs. left-right ‘stereo’ comparisons in driving rapid locomotor adjustments. Together, these data will help establish the sensory-motor algorithms that underlie a key ethological behavior, and provide a foundation for a wider exploration of how animals exploit sensory cues to navigate complex natural environments.