Enabling Adaptive Cardio-Respiratory Biofeedback Training on Ubiquitous Hand-Worn Devices

Paper Title

Enabling Adaptive Cardio-Respiratory Biofeedback Training on Ubiquitous Hand-Worn Devices

Publication Info

• Topic area: Adaptive cardio-respiratory biofeedback using wearable devices for personalized mental health interventions.
• Keywords: Adaptive biofeedback, cardio-respiratory coupling, HRV, wearable devices, smart rings, smartwatches, embodied interaction, real-time feedback, physiological sensing, stress regulation.

Background and Problem

• Problem / challenge: Conventional HRV biofeedback systems rely on fixed breathing guidance, which is often uncomfortable, misaligned with users' dynamic physiological states, and dependent on specialized equipment. Existing adaptive systems are limited by single-signal feedback and lack scalability for everyday use.
• Significance: Enhancing cardio-respiratory coupling (CRC) through biofeedback has proven benefits for stress reduction, anxiety management, and chronic condition treatment. Making such systems accessible on everyday wearable devices could democratize mental health interventions.
• Motivation and related work: Prior research has shown the benefits of adaptive HRV biofeedback and embodied interaction but has not integrated dual-factor feedback (cardiac and respiratory) into scalable, ubiquitous platforms. This paper addresses these gaps by introducing a fully adaptive system using hand-worn devices.

Solution

• Proposed approach: A dual-factor adaptive cardio-respiratory biofeedback system implemented on hand-worn devices (smart rings and smartwatches) that dynamically adjusts breathing guidance based on real-time HRV and respiration signals.
• Novelty:
  1. First hand-worn adaptive biofeedback system integrating cardiac and respiratory indicators for real-time guidance.
  2. Introduction of a hand-on-abdomen embodied interaction posture to enhance diaphragmatic breathing and signal quality.
  3. Lotus-based visualization design for intuitive, real-time feedback on HRV and breathing alignment.
  4. Empirical evidence of the system's effectiveness in improving HRV, breathing alignment, and user experience.
• Procedure and key techniques:
  • Users place a hand on their abdomen to stabilize sensors and enhance interoceptive awareness.
  • IMU and PPG sensors capture respiration and HRV signals, processed via a dual-factor adaptive algorithm.
  • Breathing guidance dynamically adjusts based on short-term respiration rate stability and long-term HRV trends.
  • Feedback is delivered through a lotus-based visualization interface, with petals reflecting HRV and leaves guiding breathing.

Results

• Concrete findings:
  • The adaptive system achieved a mean HRV rise of 30.2 ms, with HRV sustained above 120% of baseline for 263.3 seconds (76.6% longer than baseline).
  • Breathing alignment was highest in the adaptive condition, with 82% of time at ≥80% alignment and 36% at ≥95%.
  • Aesthetic appeal ratings were significantly higher for the lotus visualization compared to baseline.
• Advantage over baselines:
  • Adaptive guidance outperformed fixed guidance in HRV improvement, breathing alignment, and user engagement.
  • Embodied interaction (hand-on-abdomen) enhanced respiratory alignment and accelerated physiological engagement.
  • The fully integrated system (Embodied-Adaptive) showed the best performance across all metrics.
• Experiments / evaluation:
  • A 48-participant study tested four conditions (Baseline, Fixed, Embodied-Fixed, Embodied-Adaptive) using physiological (HRV, breathing alignment) and subjective (engagement, workload, usability) metrics.
  • Signal quality assessments validated the accuracy of smart ring and smartwatch sensors against reference devices.
• Limitations and future work:
  • Short-term evaluation; long-term effects remain untested.
  • Limited to smart ring implementation; smartwatch and smartphone adaptations need exploration.
  • Respiration depth was not directly measured; future systems could integrate depth-based feedback.

Summary

This paper introduces a dual-factor adaptive cardio-respiratory biofeedback system using hand-worn devices, combining real-time HRV and respiration signals with an embodied interaction paradigm. The system demonstrated significant improvements in HRV, breathing alignment, and user engagement compared to fixed guidance methods. The lotus-based visualization and hand-on-abdomen posture enhanced user experience and physiological outcomes. While the system shows promise for scalable mental health interventions, future work should explore long-term effects and broader device compatibility.