Monique Wochatz1, Nina Tilgner2, Steffen Mueller3, Sophie Rabe4, Sarah Eichler4, Michael John5, Heinz Völler4, Frank Mayer2. 1. University of Potsdam, University Outpatient Clinic, Sports Medicine & Sports Orthopaedics, Potsdam, Germany. Electronic address: wochatz@uni-potsdam.de. 2. University of Potsdam, University Outpatient Clinic, Sports Medicine & Sports Orthopaedics, Potsdam, Germany. 3. Trier University of Applied Science, Computer Science/Therapy Science, Trier, Germany. 4. University of Potsdam, Center of Rehabilitation Research, Potsdam, Germany. 5. Fraunhofer Institute for Open Communication Systems, Berlin, Germany.
Abstract
BACKGROUND: Besides its initial use as a video gaming system the Kinect might also be suitable to capture human movements in the clinical context. However, the system's reliability and validity to capture rehabilitation exercises is unclear. RESEARCH QUESTION: The purpose of this study was to evaluate the test-retest reliability of lower extremity kinematics during squat, hip abduction and lunge exercises captured by the Kinect and to evaluate the agreement to a reference 3D camera-based motion system. METHODS: Twenty-one healthy individuals performed five repetitions of each lower limb exercise on two different days. Movements were simultaneously assessed by the Kinect and the reference 3D motion system. Joint angles and positions of the lower limb were calculated for sagittal and frontal plane. For the inter-session reliability and the agreement between the two systems standard error of measurement (SEM), bias with limits of agreement (LoA) and Pearson Correlation Coefficient (r) were calculated. RESULTS: Parameters indicated varying reliability for the assessed joint angles and positions and decreasing reliability with increasing task complexity. Across all exercises, measurement deviations were shown especially for small movement amplitudes. Variability was acceptable for joint angles and positions during the squat, partially acceptable during the hip abduction and predominately inacceptable during the lunge. The agreement between systems was characterized by systematic errors. Overestimations by the Kinect were apparent for hip flexion during the squat and hip abduction/adduction during the hip abduction exercise as well as for the knee positions during the lunge. Knee and hip flexion during hip abduction and lunge were underestimated by the Kinect. SIGNIFICANCE: The Kinect system can reliably assess lower limb joint angles and positions during simple exercises. The validity of the system is however restricted. An application in the field of early orthopedic rehabilitation without further development of post-processing techniques seems so far limited.
BACKGROUND: Besides its initial use as a video gaming system the Kinect might also be suitable to capture human movements in the clinical context. However, the system's reliability and validity to capture rehabilitation exercises is unclear. RESEARCH QUESTION: The purpose of this study was to evaluate the test-retest reliability of lower extremity kinematics during squat, hip abduction and lunge exercises captured by the Kinect and to evaluate the agreement to a reference 3D camera-based motion system. METHODS: Twenty-one healthy individuals performed five repetitions of each lower limb exercise on two different days. Movements were simultaneously assessed by the Kinect and the reference 3D motion system. Joint angles and positions of the lower limb were calculated for sagittal and frontal plane. For the inter-session reliability and the agreement between the two systems standard error of measurement (SEM), bias with limits of agreement (LoA) and Pearson Correlation Coefficient (r) were calculated. RESULTS: Parameters indicated varying reliability for the assessed joint angles and positions and decreasing reliability with increasing task complexity. Across all exercises, measurement deviations were shown especially for small movement amplitudes. Variability was acceptable for joint angles and positions during the squat, partially acceptable during the hip abduction and predominately inacceptable during the lunge. The agreement between systems was characterized by systematic errors. Overestimations by the Kinect were apparent for hip flexion during the squat and hip abduction/adduction during the hip abduction exercise as well as for the knee positions during the lunge. Knee and hip flexion during hip abduction and lunge were underestimated by the Kinect. SIGNIFICANCE: The Kinect system can reliably assess lower limb joint angles and positions during simple exercises. The validity of the system is however restricted. An application in the field of early orthopedic rehabilitation without further development of post-processing techniques seems so far limited.
Authors: Francis Lovecchio; Grant J Riew; Dino Samartzis; Philip K Louie; Niccole Germscheid; Howard S An; Jason Pui Yin Cheung; Norman Chutkan; Gary Michael Mallow; Marko H Neva; Frank M Phillips; Daniel M Sciubba; Mohammad El-Sharkawi; Marcelo Valacco; Michael H McCarthy; Melvin C Makhni; Sravisht Iyer Journal: Eur Spine J Date: 2020-11-22 Impact factor: 3.134
Authors: Evelyne Wiskerke; Jan Kool; Roger Hilfiker; Karl Martin Sattelmayer; Geert Verheyden Journal: JMIR Serious Games Date: 2022-03-22 Impact factor: 3.364