The Future of Education

Children affected by the war in Ukraine face significant emotional regulation challenges that directly impact their ability to learn and participate in an inclusive school environment [1, 2]. Teachers who host displaced and vulnerable students often lack practical tools to recognize subtle signs of stress, especially when children have developed protective behaviors such as emotional masking [3]. In our previous work, we introduced BioMirror, a research platform for iOS that integrates real-time facial emotion recognition via the iPhone TrueDepth camera with physiological monitoring via the Apple Watch [4]. The current system uses ARKit to track 52 facial shape combination parameters and classify seven core emotions using a rules-based algorithm with temporal smoothing; It simultaneously collects heart rate, heart rate variability (RMSSD, SDNN, LF/HF ratio), wrist temperature, and motion data via HealthKit and CoreMotion, calculating an emotional-physiological coherence score every five seconds by smoothing the facial and autonomic arousal streams. Technical validation with 20 adult volunteers under controlled conditions demonstrated 87.3% accuracy in basic emotion recognition, data synchronization between devices in less than 50 ms, and a biometric signal quality of 94% [4]. In this paper, we present a significant extension of the platform: a companion macOS app, BioMirror Researcher, designed as a real-time experiment control panel. The macOS app automatically discovers iOS devices running BioMirror on the local network via Bonjour/mDNS and establishes persistent WebSocket connections, allowing researchers to monitor multiple participants simultaneously from a single workstation. The dashboard displays a live grid of connected devices with real-time emotion classification, heart rate graphs, coherence metrics trends, battery levels, and Apple Watch connection status; provides a notification bar that indicates low battery, degraded face tracking, or watch disconnection; and allows you to remotely manage a session - start, stop, and assign participants across all registered devices. This multi-device architecture transforms BioMirror from a single-participant tool to a classroom-scale research tool, allowing for research with groups of 10-30 students at a time, while maintaining full control over data quality and session integrity. We discuss how this expanded platform can be adapted for inclusive school practices: implementing short, guided microsessions with a child-friendly digital nature, transforming coherence data into actionable teacher-centered metrics, and using minimal data mode to protect children’s privacy. We outline a pilot mixed-methods study design for assessment with students aged 8-14 in inclusive schools in Ukraine [5, 6].