Abstract
Understanding text and speech requires several bindings, integrations of information. Our eyes combine letters to recognize a printed word. Our ears combine phonemes to recognize a spoken word. Sometimes we combine information from eye and ear, especially when learning to read. Dubois, Poeppel, and Pelli (submitted to Nature and VSS) assessed this audio-visual binding, showing that although adults inefficiently combine within a modality (e.g., letters or phonemes), they perfectly combine information from eye and ear. How does our ability to integrate features in text and speech develop? Here we probe how the bindings within and across modalities change as a function of age. Child and adult participants are asked to identify words presented concurrently through eye and ear, in both audio and visual noise. We vary the overall signal-to-noise ratio and the ratio of audio to visual signal energy. From age 6 to adulthood, we find that auditory and visual efficiencies increase by factors of 2.4 and 2, respectively. Assessing bimodal binding reveals that a 6-year-old, despite having much lower efficiency within each modality, already combines information from eye and ear perfectly, like an adult.