A visual concept shapes image perception.

Six subjects--three radiologists and three laypersons--were shown a hidden figure picture, a sonogram, a tomogram, and two standard radiographic images while their eye movements were recorded. They were asked to summarize each of the images in one sentence. The distribution of visual attention was found to be related to the assignment of meaning to the image. When the interpretation of the image was correct, visual attention focused on the main subject or radiological abnormality. When the viewers found the images to be ambiguous, visual attention focused on dominant pictorial features such as borders or textured areas. These features were then used to generate a meaningful perception in response to the request for a summary statement. The data are interpreted in terms of a theory that holds that picture perception begins with a parallel global characterization of the scene that uses data from both the retina and memory to form a visual concept. According to this theory, it may be necessary to have seen radiological abnormalities in the past to be able to form an appropriate visual concept for a current abnormality.