Some representative experiments that span over fifty years of research on visual selective attention are reviewed here. They parametrically describe spatial attention to locations in space, temporal attention to intervals in time, and feature attention to particular visual features. The resulting data are accurately described by a class of lean computation models (i.e., few estimated parameters) in which attention is a represented as a gating (i.e., filtering) process. The models represent plausible brain functions utilizing components that represent basic neural transformations. Spatial, temporal, and feature attention are supported by parallel, independent brain processes that combine multiplicatively in the combinations so far tested.