Multimodal Prediction of Conversation Success in Autistic Dyadic Interactions

This study uses articulatory, facial, acoustic-prosodic, and turn-taking measures to predict perceived conversation success in spontaneous autistic and non-autistic dyads. The research demonstrates that conversational success is a multidimensional metric requiring integrated multimodal features rather than single-modality analysis.