THE DETECTION OF CHANGE AND THE PERCEPTUAL MOMENT HYPOTHESIS

This paper examines Stroud's theory that the perceptual input is quantized in time (the perceptual moment hypothesis). A model of the detection of change based on the theory is compared with other models derived from statistical quality control methods; in particular those using simple and geometric moving averages. They are compared theoretically to see which entails the least computational load for the brain. Empirical comparisons are made with data from loudness thresholds, brightness thresholds, and the perception of causality: all cases in which there is little change in stimulus intensity. The perceptual moment hypothesis is found to be the most satisfactory. Finally, some of the assumptions of the perceptual moment model are examined in more detail and the model is extended to deal with situations in which the intensity of stimulation alters greatly.