Abstract: To survive in dynamic environments animals have evolved the capacity to combine previous sensory experiences to guide decision-making. In mammals, the integration of past sensory experience is thought to involve the activity of large groups of neurons distributed throughout the neocortex and associated structures. The biological mechanisms of this form of memory, however, remain largely unknown. Over the past several years we have begun to probe the neural basis of sensory integration in rats, using a combination of quantitative behavioral models, neural circuit-based manipulations and, more recently, a set of new technologies that enable optical imaging of calcium dynamics in populations of neurons. Using this approach, we identified a group of neurons in the rat frontal cortex that encode, collectively, the memory of past sensory stimuli. In contrast to prevailing models of neural integration, we found that the responses of individual neurons to sensory events were transient, yet temporally diverse, with some neurons exhibiting peaks in activity at different latencies, up to seconds after a sensory event. These data support a model of neural integration, inspired by advances in theoretical computing, in which neurons collectively represent the memory of past sensory experience on behavioral timescales using diverse, time varying responses.