During the past few decades, optical methods for imaging activity in networks composed of thousands of neurons have been developed. These techniques rely mainly on organic-chemistry-based dyes as indicators of Ca(2+) and membrane potential. However, recently a new generation of probes, genetically encoded fluorescent protein sensors, has emerged for use by physiologists studying the operation of neuronal circuits. We critically review the development of these new probes, and analyze objectives and experimental conditions in which classical probes are likely to prevail and where the fluorescent protein sensors will open paths to previously unexplored territories of functional neuroimaging.