Quantification and Visualization of Interpersonal Synchrony Using Wearable Sensors: A Case Study on Autistic and...

Area of Science:

• Social Sciences
• Developmental Psychology
• Rehabilitation Science

Background:

• Interpersonal synchrony (IS) is crucial for social interaction but traditionally measured via time-consuming, subjective video analysis.
• Existing methods lack scalability and objectivity, hindering comprehensive research in social and developmental fields.

Purpose of the Study:

• To develop and validate an automated, sensor-based framework for quantifying and visualizing interpersonal synchrony (IS).
• To compare different time series similarity measures and machine learning approaches for IS classification and pseudosynchrony detection.
• To provide a scalable, objective, and reproducible alternative to manual coding for IS research.

Main Methods:

• Utilized wearable sensor data to capture time series interactions.
• Evaluated Cross-correlation (CC), Dynamic Time Warping (DTW), and Cross-Wavelet Analysis (XWA) as features for machine learning models.
• Compared surrogate data analysis and supervised learning for pseudosynchrony identification.

Main Results:

• Similarity-based features significantly outperformed conventional statistical features in classifying high and low IS levels using ensemble classifiers.
• The study systematically evaluated methodological trade-offs in pseudosynchrony detection.
• Developed visualization tools for dynamic tracking and filtering of pseudosynchrony.

Conclusions:

• The proposed sensor-based framework provides a scalable, objective, and reproducible method for IS assessment.
• This approach addresses a critical gap in synchrony research, enabling broader applications in social, developmental, and rehabilitation studies.
• Automated IS quantification enhances the efficiency and reliability of social interaction analysis.