Skip to content
Research · Oct 26, 2026

Study proposes multimodal framework to predict perceived conversation success in autistic and non-autistic dyads

Researchers combine articulatory, facial, acoustic-prosodic, and turn-taking measures to model perceived conversational success across three spontaneous interaction datasets.

Trust79
HypeLow hype

1 source · cross-referenced

ShareXLinkedInEmail
TL;DR
  • Predicts perceived conversation success using multimodal features from articulatory, facial, acoustic-prosodic, and turn-taking measures.
  • Analyzes three spontaneous dyadic-interaction datasets including autistic and non-autistic individuals.
  • Finds conversational success is multidimensional and requires a multi-faceted analytical framework.
  • Submitted to ICASSP 2027.

A new arXiv preprint proposes a multimodal approach to predict perceived conversation success in dyadic interactions involving autistic and non-autistic participants. The study, led by researchers including Thanushi Withanage and Elizabeth Redcay, analyzes three spontaneous interaction datasets using complementary features spanning articulatory movements, facial expressions, acoustic-prosody, and turn-taking behavior.

Prior work has largely focused on predicting perceived success using single modalities or in pediatric autism-enriched samples, or on nonverbal behavior in non-autistic groups. This study instead integrates multiple modalities to model perceived success as a multidimensional construct, arguing that no single channel captures the full picture of conversational quality.

The authors report that combining these modalities improves prediction of perceived success by capturing information unavailable to any single modality. They submit the work to ICASSP 2027 and note its potential to inform tools for social skills training and real-time feedback systems in mixed neurotype interactions.

Sources
  1. 01arXiv cs.CL — Multi modal perceived conversation success prediction for spontaneous autistic and non-autistic dyadic interactions
Also on Research

Stories may contain errors. Dispatch is assembled with AI assistance and curated by human editors; despite the trust-score filter, mistakes happen. We correct publicly — every article links to its revision history. Nothing here is financial, legal, or medical advice. Verify before relying on any claim.

© 2026 Dispatch. No ads. No sponsorships. No paid placement. Reader-supported via Ko-fi.

Built by a person who cares about honest AI news.