In a bright-field measurement from a vertebrate brain stained by superfusing a solution of the dye over the surface, each pixel in a camera receives light from a substantial number (thousands) of neurons and neuronal processes (population signals). Because of scattering and out-of-focus light, this will be true even if the pixel size corresponds to a small area of the brain. In this situation, the voltage-sensitive dye signal will be a population average of the change in membrane potential of all of these neurons and processes. Many investigators have published voltage-sensitive dye imaging studies of population activities in brain slices. Their methods, including choice of dyes, illumination intensity, and imaging device, vary across a large spectrum. Here we present a protocol for visualizing spatiotemporal patterns in rodent neocortex in vitro. Detecting these patterns requires high-sensitivity imaging in single trials, because averaging will obscure the complex dynamics of the spatiotemporal patterns.