Sotaro Kita

Abstract The Shared Book Reading Corpus is a collection of audiovisual recordings of English-speaking caregivers and 13–14-month-old infants (N = 44 dyads) reading together in a controlled lab setting. Caregivers were instructed to read a First 100 Words picture-book with their infant as they naturally would at home. An overview camera recorded the full interaction, while two head-mounted cameras worn by all caregivers and most infants captured their individual perspectives.

When we listen to someone speak, seeing their face can make it easier to understand their words, especially in noisy environments. This study examined how autistic and non-autistic adults process speech using both sound and visual cues, such as watching the speaker's mouth move. While both groups benefited equally from combining sound and visual information to understand speech, their brain activity patterns were different.

We studied multisensory integration (MSI) of audio-visual speech accompanied by iconic gestures in an ecologically valid, naturalistic situation with a noisy context. Participants showed behavioural (improved speech recognition) and neural (alpha suppression) correlates of MSI in audio-visual trials. We demonstrate that non-linear MSI effects can be detected in the oscillatory power responses to complex speech stimuli. 1 INTRODUCTION Speech is a paramount element of human social life.