ASD Cabin: A systematic and multiscale evaluation framework for quantitative behavioral analysis in autism screening

Early and accurate screening for autism spectrum disorder (ASD) continues to represent a significant unmet clinical challenge within pediatric healthcare. Quantitative behavioral assessment is fundamentally limited by the absence of standardized, ecologically valid evaluation paradigms that elicit consistent, comparable social responses across children. Conventional behavioral assessments rely on loosely controlled tasks and heterogeneous protocols, producing fragmented behavioral readouts that hinder mechanistic interpretation, cross-subject comparability, and reproducible benchmarking across independent studies. To address this critical gap, we propose ASD Cabin, a standardized, ecologically grounded, and highly reproducible scenario-based evaluation framework designed for objective quantitative behavioral screening of ASD. The framework unifies multiple real-world social interaction scenarios into a single, rigorously controlled experimental pipeline that enforces uniform behavioral elicitation and measurement across all participants. This design enables coherent multi-scenario behavioral profiling rather than isolated, task-specific measurements. To establish a benchmark for unified multi-scenario ASD behavioral research, we curate a dedicated behavioral dataset collected strictly under the ASD Cabin protocol and conduct a systematic validation study to evaluate its screening efficacy. Experimental results demonstrate that the proposed framework achieves robust and reliable classification performance, with an overall accuracy of 95.26% and an F1-score of 95.23%. The results verify that ASD Cabin can stably extract interpretable, discriminative behavioral patterns with strong cross-sample comparability, demonstrating its promise as a scalable evaluation framework supporting quantitative behavioral analysis for early ASD screening.