Towards an Engagement-Driven Rehabilitation Framework: a Pilot Study

Paper Title

Towards an Engagement-Driven Rehabilitation Framework: a Pilot Study

Publication Info

• Topic area: Pediatric neurorehabilitation using engagement-driven adaptive frameworks.
• Keywords: Pediatric neurorehabilitation, engagement assessment, gamification, EEG monitoring, cognitive engagement, emotional engagement, behavioral engagement, adaptive therapy, dyspraxia, multimodal sensors.

Background and Problem

• Problem / challenge: Traditional pediatric neurorehabilitation methods lack personalized adaptability and fail to consider patients' mental states, leading to low motivation and adherence.
• Significance: Engagement is critical for effective rehabilitation, especially in children with neuromotor impairments. Addressing this gap can improve therapeutic outcomes and adherence.
• Motivation and related work: Prior work on gamification and serious games has shown promise in pediatric rehabilitation but often focuses on physical performance rather than engagement. Existing methods are limited by small sample sizes, short intervention durations, and lack of integration of emotional and cognitive metrics.

Solution

• Proposed approach: An engagement-driven rehabilitation framework integrating multimodal sensors (EEG, cameras) to assess emotional, cognitive, and behavioral engagement in real time, coupled with an adaptive game-based therapy system.
• Novelty:
  1. Conceptualization of engagement as a therapeutic signal for real-time adaptation of rehabilitation tasks.
  2. Co-design with therapists and engineers to ensure clinical relevance and adaptability.
  3. Pilot evaluation of the framework with children having neuromotor impairments, focusing on engagement and rehabilitation outcomes.
• Procedure and key techniques:
  • Multimodal engagement assessment using EEG (cognitive and emotional engagement) and cameras (behavioral engagement).
  • Adaptive game-based therapy where task difficulty and stimuli are adjusted based on engagement metrics.
  • Participatory design process involving therapists, psychologists, and engineers to tailor the system to clinical needs.
  • Pilot study with four children, analyzing engagement trends and motor coordination improvements.

Results

• Concrete findings:
  • Cognitive engagement classification accuracy: 74.5%; emotional engagement classification accuracy: 71.2%.
  • Improvements in motor coordination scores (e.g., Child 1 progressed from >10th to <50th percentile).
  • High emotional and cognitive engagement levels observed in most sessions.
• Advantage over baselines:
  • Real-time adaptation of therapy tasks based on engagement metrics.
  • Sustained attention and motivation without inducing stress or boredom.
• Experiments / evaluation:
  • Participants: Four children aged 6–7 with neuromotor impairments.
  • Sessions: 3–7 per child over two months.
  • Metrics: Engagement levels (emotional, cognitive, behavioral), dyspraxia test scores, practitioner evaluations.
• Limitations and future work:
  • Small sample size limits generalizability.
  • EEG headset caused discomfort for some participants, reducing compliance.
  • Future work includes replacing EEG with less intrusive sensors, integrating additional data sources (e.g., pressure sensors), and conducting larger-scale studies.

Summary

This study introduces an engagement-driven rehabilitation framework for children with neuromotor impairments, integrating multimodal sensors and adaptive game-based therapy. The framework dynamically adjusts tasks based on real-time engagement assessments, promoting sustained motivation and participation. A pilot study with four children demonstrated improvements in motor coordination and high engagement levels, though challenges with EEG compliance were noted. Future work aims to refine sensor technologies and validate the approach with larger, more diverse cohorts. This framework represents a promising step toward personalized, engagement-centered pediatric rehabilitation.