An investigation of speed of visual information processing in infants

Abstract

The study was designed to examine: (1) the relationship between an indirect measure of speed of processing, namely fixation length, and a more direct measure of speed of processing, namely learning rate, and (2) the relationship between speed of processing and the nature of the information obtained. 6 1/2-month-old infants were given a learning task, in which a familiar, cartoon-like face was paired repeatedly with a series of novel faces that differed from the familiar only in their internal features (brows, eyes, nose, and mouth). Trials continued until infants met criterion (ie., a stable preference for the novel) or until they completed the full sequence of 18 trials. In subsequent probe trials, the familiar configuration of facial features was paired with subtle variants of itself, to determine the nature of the information infants had extracted during learning. Peak and mean fixations were obtained from the learning task and a preliminary looking task. The results indicate a relation between rate of learning and performance on probe trials. Whereas infants who reached criterion very quickly performed at chance level, no better, in fact, than infants who failed to reach criterion at all, infants who took somewhat longer to reach criterion were able to discern subtle changes in the familiar stimulus. Fixation measures from the learning task showed some relation to rate of learning but no relation to probe performance. The findings are compatible with the notion that fast and slow learners differ with respect to the nature of the information obtained during learning. An alternative interpretation was also considered.