The role of audiovisual object features and scene context on children's ability to categorise familiar objects

Children's ability to visually categorise objects emerges early in development. However, the use of relevant object features from other modalities, or contextual scene cues, on categorisation across development is poorly understood. In particular, children's categorisation may benefit from any associated cue to categorisation or may be specific to the availability of object-relevant sensory information. We tested children between the ages of 5-13 years on their categorisation of familiar objects, which were presented either as images-only (V), sounds-only (A) or bimodal AV objects. In addition, the objects were presented either in isolation or within a scene context that was either semantically congruent or incongruent to the object category. Whilst categorisation performance improved with age, children were better able to categorise object images than sounds, and there was no overall advantage for combined AV information over images only on either accuracy or response times. A semantically congruent context did not further improve categorisation accuracy, although response times were fastest to objects presented within a congruent scene context, particularly when an object's sound was presented. Our findings suggest that visual information dominated the categorisation of familiar objects over sounds in children, with no evidence that these cues are combined at any age. However, children's categorisation of objects benefitted from the associative scene context, particularly when the object information was impoverished. Our findings have important implications for understanding how children adaptively use available information in the categorisation of familiar objects.