Within sensory structures, even the simplest stimulus engages thousands of neurons that have widely varying stimulus selectivity and are spatially distributed in sensory brain maps. This organization raises the essential question of the rules governing the integration of the activity of such a large dispersed population of neurons to produce uniform percepts and reliable behaviors. Do behavioral responses to a sensory stimulus rely on a weighted sum of all active neurons that represent it, or a weighted sum of particular subpopulations of neurons (for example, defined by genetic identity, stimulus selectivity, location, or projection targets)? Moreover, how are neurons in different regions of an active population weighted? While these issues have been computationally investigated using a variety of decoding approaches, the causal link between population activity and behavior has been lacking. This project will provide such links by using patterned stimulation of targeted and characterized neuronal populations with in vivo holographic stimulation to bias and drive behavioral responses. The proposed experiments will determine how the spatial relationships between neurons relate to their relative impact on behavior, and assess how this spatial weighting is affected by changes in stimulus intensity, signal-to-noise ratio, and stimulus complexity. Comparisons between sensory systems will reveal which rules are general, and which are related to particular sensory demands. Because sensory neurons are broadly tuned, every stimulus activates neurons with different stimulus preferences. The experiments will test whether neurons that share a particular stimulus preference have cooperative effects and how this weighting is affected by variation in signal-to-noise ratio. Finally, because neurons in each brain area have different identities that reflect their different functional contributions, the project will test if behavioral roles vary between neurons across cortical layers, genetic identities, or projection targets. Collectively, these experiments will provide new and previously unattainable information about the encoding and readout of stimuli as well as how those results generalize to more natural sensory stimuli.