Connectionist theories of visual perception suggest that the brain processes visual information through a network of neurons that are sensitive to different visual features, such as edges, lines, and colours. These features combine to form complex visual representations, which can then be used to recognise objects and scenes.