The infant is faced with a barrage of complex information in the world, being conveyed through every sense modality. The infant must isolate streams of information specifying unified objects and events (e.g., object with impact sound, moving face with voice) in order to perceive an organized flow of information. However, research has suggested that information from different modalities may interact with one another to alter perception of an object or event. The objective of this project is to investigate audiovisual interactions in categorization and identify their developmental course. The central hypothesis for the proposed research is that auditory information plays an important role in categorization and that this role is modified during development, based on the extent of the individual's category knowledge. The proposed investigations are significant because they will provide a link between research in intersensory perception and categorization to demonstrate the role of multimodal information in the development of categories. Gender is a natural category that varies along a continuum and has been shown to have a long developmental trajectory that continues well into childhood. Previous research has demonstrated a typicality effect in gender categorization in which more masculine or feminine faces are categorized more accurately and quickly by both adults and children. Infants and adults will be tested to determine if categorization of facial gender can be biased based on the gender of a synchronous voice. The range, and development, of this effect will be tested by varying the typicality of the face which individuals are asked to categorize. Five- and ten-month-old infants will be tested in an infant-controlled habituation procedure in which gender-ambiguous or less typical faces of a single gender are paired with very typical voices and tested with very typical male and female faces. Preference for the opposite gender of the habituation voices indicates that infants categorized the habituation faces based on the gender of the synchronous voices. Adult participants will be asked to identify the gender of a series of gender-ambiguous and less typical faces synchronized with very typical voices. Accuracy and reaction time data will be analyzed for trials in which the gender of face and voice match versus trials in which the gender of face and voice are mismatched. Future research will include investigations of other categories and interactions of other modality combinations.
