Introduction. Interpersonal dynamics are central to autism intervention and child development, yet difficult to measure objectively. Prosodic synchrony and affective features are increasingly recognized as drivers of therapeutic and developmental change, but research has been limited by methodological challenges and lack of scalable tools.


Purpose. This series of studies introduces a computational framework grounded in affective computing and dynamic systems theory to quantify interpersonal and intrapersonal prosodic synchrony across multiple domains. We examined how synchrony relates to therapy outcomes, parental stress, gender, and clinical evaluations.


Method. We developed a fully automated pipeline combining deep learning-based speech segmentation with second-by-second modeling of vocal interactions. Synchrony was quantified via Cross-Recurrence Quantification Analysis applied to prosodic features over entire clinical sessions. Analyses included (i) longitudinal therapy sessions of 25 autistic preschoolers (75 sessions, 17 therapists) over one year; (ii) early child-therapist synchrony to predict long-term response; (iii) 62 longitudinal parent-child dyads (mothers and fathers separately) alongside parental stress; and (iv) N=111 ADOS-2 assessments to explore affective synchrony in diagnostics.


Results. Early child-therapist synchrony predicted one-year intervention gains, with better outcomes linked to variability, predictability, and prosodic features associated with emotional engagement. Longitudinal analyses revealed a shift from structural to dynamic transition synchrony markers across therapy. In parent-child interactions, gender and stress were associated with less stable and predictable synchrony patterns. Across all contexts, these results reflect the moment-to-moment coordination of interaction, capturing the temporal dynamics underlying interpersonal and intrapersonal synchrony.


Conclusions. Across intervention, caregiver-child, and diagnostic contexts, prosodic synchrony emerges as a key affective mechanism shaping developmental trajectories in autism. Automated computational pipelines enable scalable, ecologically valid insights, highlighting synchrony as both a clinical marker and potential intervention target, bridging research and practice while enhancing translational applications by providing clinically interpretable guidance.