Design, development, and clinical validation of KaraMetrics: A sensor-based hand function assessment tool for stroke rehabilitation

Abstract

Stroke is a leading cause of adult disability worldwide, with persistent hand dysfunction severely limiting independence and quality of life for survivors. Existing clinical assessments of hand function such as the Fugl-Meyer Assessment (FMA), Action Research Arm Test (ARAT), and Jebsen-Taylor Hand Function Test (JTHFT) are time-consuming, subjective, expensive, or require specialized training, making them impractical in low-resource settings and overburdened clinical environments. This thesis presents KaraMetrics: a novel, modular, sensor-based system for rapid, objective, and fine-grained assessment of post-stroke hand function. The system integrates five low-cost, Arduino-based subsystems to evaluate grip strength, pinch strength, finger-tapping speed, range of motion, and manual dexterity, tested using a 10-15-minute protocol. A custom web application supports seamless data acquisition and visualization, including individualized radial plots to guide therapy planning and monitor progress. KaraMetrics was iteratively developed and validated through pilot testing with five stroke survivors, followed by a comparative study involving seventeen patients benchmarked against the JTHFT, conducted at the National Institute of Mental Health and Neuro Sciences (NIMHANS), Bengaluru. Results show that KaraMetrics is significantly more sensitive to functional changes over time than standard clinical tools. The system demonstrated excellent interrater and intrarater reliability, with variation below 5.5% in over 99% of repeated measurements. Moreover, it provided individualized functional profiles, enabling targeted, patient-specific rehabilitation strategies. By combining clinical relevance with interpretability and objectivity, KaraMetrics addresses key limitations of existing assessment approaches. It offers a scalable solution for therapists, clinicians, and researchers seeking precise, objective monitoring of hand function particularly in resource-limited settings. This work contributes a validated hardware-software integrated assessment tool to the field of neurological rehabilitation and lays the groundwork for broader adoption of sensor-driven assessment in stroke care.