CalmReminder: A Design Probe for Parental Engagement with Children with Hyperactivity, Augmented by Real-Time Motion Sensing with a Watch

Paper Title

CalmReminder: A Design Probe for Parental Engagement with Children with Hyperactivity, Augmented by Real-Time Motion Sensing with a Watch

Publication Info

• Topic area: Wearable sensing and adaptive notifications for parental engagement in ADHD management.
• Keywords: ADHD, hyperactivity, parental engagement, wearable sensing, calm moments, just-in-time adaptive interventions, family-centered design, motion sensing, positive reinforcement, behavior management.

Background and Problem

• Problem / challenge: Existing digital interventions for ADHD often fail to reflect parents’ lived practices, focus on monitoring problematic behaviors rather than reinforcing positive ones, and lack personalization.
• Significance: Supporting parents in recognizing and reinforcing calm moments can improve parent-child interactions, reduce reactive parenting, and foster positive behavioral changes in children with ADHD.
• Motivation and related work: Prior research highlights the importance of timely positive reinforcement and family-centered interventions but has not explored real-time sensing to detect calm moments or its impact on parental engagement. This paper addresses this gap by introducing a system that detects calm moments and prompts parents to engage positively.

Solution

• Proposed approach: CalmReminder, a smartwatch-based system that detects children’s calm moments using motion sensing and delivers just-in-time prompts to parents for reflection and positive reinforcement.
• Novelty:
  1. Real-time calm-moment detection using wearable IMU data and personalized regression models.
  2. Exploration of diverse notification strategies (none, hourly, random, calm-only) in a four-week in-home study.
  3. Insights into how families flexibly adopt and adapt system features to align with their parenting strategies.
• Procedure and key techniques:
  • A smartwatch app collects IMU data and computes motion energy in real time.
  • A personalized regression model maps motion energy to parental perceptions of calmness.
  • Notifications are delivered to parents’ smartphones based on predefined strategies (e.g., calm-only).
  • Parents respond to intraday, end-of-day, and end-of-week surveys to provide feedback and train the model.

Results

• Concrete findings:
  • The calm-moment detection model achieved a perceived calm ratio of 78.0% in the calm-only condition.
  • Parents actively appropriated system features, using them for positive reinforcement, mindfulness, activity scheduling, and conversation starters.
  • High baseline stress levels correlated with system rejection or dropout, while lower stress levels supported adoption.
• Advantage over baselines:
  • Calm-only notifications aligned with parents’ perceptions of calm moments better than random notifications (78.0% vs. 38.7% perceived calm ratio).
  • Personalized regression models (R² = 0.44) outperformed global models (R² = 0.25).
• Experiments / evaluation:
  • Four-week in-home study with 16 families (12 completed) of children aged 6–12 with ADHD.
  • Within-subjects design comparing four notification strategies (none, hourly, random, calm-only).
  • Data collected through surveys and interviews to assess usability, engagement, and parental perceptions.
• Limitations and future work:
  • Small sample size and gender imbalance (mostly male children).
  • Limited ability to evaluate model accuracy due to sparse ground-truth labels.
  • Future work should explore integration with parent training programs, involve children’s perspectives, and improve technical aspects like modeling and connectivity.

Summary

CalmReminder is a smartwatch-based system designed to detect calm moments in children with ADHD and deliver real-time prompts to parents for reflection and positive reinforcement. A four-week study with 12 families demonstrated the feasibility of calm-moment detection and highlighted diverse ways parents adopted the system to align with their parenting strategies. The study revealed that high baseline stress hindered adoption, while flexible notification strategies supported varied family needs. These findings suggest that behavior-based prompting can enhance parent-child interactions and inform the design of personalized, family-centered interventions. Future work should address technical improvements, expand participant diversity, and explore dyadic engagement strategies.