Multimodal physiological monitoring in augmented reality teaching environments for children with neurodevelopmental disorders

Zhang, Shuyi; Cho, Sukyoung; Duan, Fengle; Feng, Hao; Zhang, Qiaoyan; Ma, Muqing · 2026 · Crossref

DOI: 10.3389/fnhum.2025.1712662

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the lack of integrated systems combining augmented reality (AR) with multimodal physiological monitoring for children with neurodevelopmental disorders. Motivated by the rising prevalence of conditions such as autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), and specific learning disabilities (SLD), the research aims to create adaptive learning environments that objectively assess and respond to individual cognitive and emotional states. The authors hypothesized that AR would reduce cognitive load, that multimodal data fusion would improve disorder classification accuracy, and that real-time physiological feedback would enhance learning outcomes compared to static interventions. The researchers conducted a prospective study involving 115 children (45 with ASD, 38 with ADHD, 32 with SLD) aged 3–10 years. Participants engaged in AR-enhanced learning tasks using Microsoft HoloLens 2 while wearing a 64-channel EEG cap, a 12-lead ECG system, and a high-speed eye-tracking device. The study design included a 12-month intervention phase where a reinforcement learning algorithm adjusted AR content difficulty based on real-time physiological feedback. Data processing involved independent component analysis for EEG artifact removal, Kubios HRV software for cardiac metrics, and Tobii Pro Lab for gaze behavior. Machine learning models fused these multimodal streams to classify disorder-specific patterns and monitor cognitive load. The multimodal fusion approach achieved 89.3% classification accuracy in distinguishing between ASD, ADHD, SLD, and typical development, significantly outperforming single-modality methods. Key biomarkers included frontal theta power variations, heart rate variability indices (specifically the LF/HF ratio), and fixation duration patterns. The AR environment reduced cognitive load by 27% compared to traditional screen-based settings, evidenced by decreased frontal theta power and more efficient visual scanning. Over the 12-month period, the personalized intervention improved attention performance by 31.2% and social interaction scores by 24.8%. These findings demonstrate the efficacy of combining AR technology with physiological monitoring for adaptive special education. The study establishes that multimodal physiological data can reliably identify disorder-specific signatures and that real-time adaptation based on these metrics significantly enhances educational outcomes. The results support the potential for AR-based systems to provide structured, engaging, and individually tailored learning experiences, offering a promising technological solution for addressing the heterogeneous needs of children with neurodevelopmental disorders.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• neuro workload indices

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data