The cartilage-generated bioelectric potentials induced by dynamic joint movement; an exploratory study

• Published in: BMC musculoskeletal disorders Vol. 26; no. 1; pp. 669 - 14
• Main Authors: Lee, Jae-Hyun, Jang, Ye-Seul, Chang, Won-Du
• Format: Journal Article
• Language: English
• Published: London BioMed Central 09.07.2025; BioMed Central Ltd; BMC
• Subjects: Adult; Articular cartilage; Biomechanics; Cartilage; Cartilage, Articular - physiology; Decomposition; Deep learning; Electrodes; Electromyography; Electrophysiology; Epidemiology; Ethnicity; Exercise; Female; Health aspects; Humans; Injuries; Internal Medicine; Isometric; Kinematics and kinetics; Knee; Knee joint; Knee Joint - physiology; Load; Male; Mechanical loading; Mechanical properties; Medical research; Medicine; Medicine & Public Health; Medicine, Experimental; Movement - physiology; Muscle contraction; Neural networks; Orthopedics; Physiological aspects; Range of Motion, Articular - physiology; Rehabilitation; Rheumatology; Sports Medicine; Standard scores; Weight-Bearing - physiology; Young Adult; South Korea; Knee; Biomechanics; Cartilage; Rehabilitation; Kinematics and kinetics
• ISSN: 1471-2474; 1471-2474
• DOI: 10.1186/s12891-025-08939-8

Background Excessive loading can damage knee cartilage, making it essential to assess and measure joint load effectively. Despite its importance, real-time monitoring of cartilage load in clinical settings remains challenging due to significant technical constraints. Electroarthrography, a recently introduced non-invasive technique, offers a promising solution by detecting load-generated potentials in joint cartilage through surface electrodes. While previous studies have primarily focused on static load applications, such as standing weight shift task or simple isometric contraction, our study explores its potential in dynamic loading scenarios. Methods We analyzed data from 20 knees in 20 subjects, using eight surface electrodes placed around each knee to capture electrical signals during three activities: active knee extension in a seated position, passive range of motion exercise in a decubitus position, and restricted squats. The recorded signals were processed into potential-time graphs, decomposed according to movement states, and analyzed through a deep neural network. Results The results showed that cartilage-generated potentials were significantly higher during active extension compared to passive extension (1.62 mV vs. 0.87 mV; p < 0.05), with the deep neural network achieving an average classification accuracy of 98.77%. Conclusion These findings highlight the feasibility of measuring and classifying cartilage-generated potentials during dynamic physical activities, providing valuable insights into load-related differences. This approach establishes a solid foundation for applications in rehabilitation medicine by facilitating the determination of appropriate exercise intensities, assessing risks associated with daily activities, and classifying physical activities. Further studies focusing on diverse biomechanical conditions will enhance its clinical utility.