Search

BIO DESIGN

pISSN 1225-8962
eISSN 2287-982X

Article

Article

Original Article

Split Viewer

Phys. Ther. Korea 2022; 29(4): 262-268

Published online November 20, 2022

https://doi.org/10.12674/ptk.2022.29.4.262

© Korean Research Society of Physical Therapy

Correlation Between Knee Muscle Strength and Maximal Cycling Speed Measured Using 3D Depth Camera in Virtual Reality Environment

Ye Jin Kim1 , PT, BPT, Hye-seon Jeon2,3 , PT, PhD, Joo-hee Park2 , PT, PhD, Gyeong-Ah Moon1 , PT, BPT, Yixin Wang1 , PT, BPT

1Department of Physical Therapy, The Graduate School, Yonsei University, 2Department of Physical Therapy, College of Health Science, Yonsei University, 3Department of Ergonomic Therapy, The Graduate School of Health and Environment, Yonsei University, Wonju, Korea

Correspondence to: Hye-seon Jeon
E-mail: hyeseonj@yonsei.ac.kr
https://orcid.org/0000-0003-3986-2030

Received: October 14, 2022; Revised: November 4, 2022; Accepted: November 4, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background: Virtual reality (VR) programs based on motion capture camera are the most convenient and cost-effective approaches for remote rehabilitation. Assessment of physical function is critical for providing optimal VR rehabilitation training; however, direct muscle strength measurement using camera-based kinematic data is impracticable. Therefore, it is necessary to develop a method to indirectly estimate the muscle strength of users from the value obtained using a motion capture camera. Objects: The purpose of this study was to determine whether the pedaling speed converted using the VR engine from the captured foot position data in the VR environment can be used as an indirect way to evaluate knee muscle strength, and to investigate the validity and reliability of a camera-based VR program.
Methods: Thirty healthy adults were included in this study. Each subject performed a 15-second maximum pedaling test in the VR and built-in speedometer modes. In the VR speedometer mode, a motion capture camera was used to detect the position of the ankle joints and automatically calculate the pedaling speed. An isokinetic dynamometer was used to assess the isometric and isokinetic peak torques of knee flexion and extension.
Results: The pedaling speeds in VR and built-in speedometer modes revealed a significantly high positive correlation (r = 0.922). In addition, the intra-rater reliability of the pedaling speed in the VR speedometer mode was good (ICC [intraclass correlation coefficient] = 0.685). The results of the Pearson correlation analysis revealed a significant moderate positive correlation between the pedaling speed of the VR speedometer and the peak torque of knee isokinetic flexion (r = 0.639) and extension (r = 0.598).
Conclusion: This study suggests the potential benefits of measuring the maximum pedaling speed using 3D depth camera in a VR environment as an indirect assessment of muscle strength. However, technological improvements must be followed to obtain more accurate estimation of muscle strength from the VR cycling test.

Keywords: Bicycling, Muscle strength, Rehabilitation, Virtual reality

Virtual reality (VR) is the application of visual simulations created with a computer program that imitates the real world or imaginable environments, items, and experiences in real time, and it requires interactivity through ongoing movements of the user [1]. VR-based rehabilitation has emerged as a new treatment option in various aspects of modern rehabilitation settings because it can optimize motor learning in a safe environment and help improve functional activities of daily living by replicating real-life scenarios [2]. VR rehabilitation also maximizes a sense of accomplishment and motivation for successive attempts, and these VR programs frequently include various exercise components, such as strength, stretching, endurance, and cardiorespiratory fitness, to improve the level of functional activities [3-5].

VR rehabilitation programs can be utilized for the remote rehabilitation of patients with time and geographic restrictions. The importance of remote rehabilitation has been highlighted during the COVID pandemic. A VR-based remote rehabilitation program equipped with an appropriate assessment function has the potential to provide an individualized and customized rehabilitation program without the need of a rehabilitation specialist. In addition, the therapist can remotely check the physical functions measured by the VR program at home and recommend a home exercise program. Motion capture systems traditionally used for rehabilitation consist of multiple sensors and need a large laboratory space. It is also expensive and not portable, so it is not appropriate for tele-rehabilitation or home-use rehabilitation programs [6]. Therefore, motion capture system-based VR programs using depth cameras such as Kinect (Microsoft, Redmond, WA, USA) are among the most convenient and cost-effective approaches for remote VR rehabilitation systems.

VR rehabilitation provides a gamified experience by creating assorted virtual environments during rehabilitation sessions for patients through monitors or VR tools [7]. To provide a proper level of challenge to individual users, the task difficulty of the VR game exercises must be modifiable to optimize motor learning according to the findings from assessments [3]. However, most of the currently developed VR games in rehabilitation do not include individualized assessments of the function and health status of users within the VR program [8,9]. Therefore, we developed a camera-based VR rehabilitation program that is capable of both evaluating physical status and providing customized rehabilitation exercises for individuals with decreased physical function secondary to aging, neuro-muscular diseases, and general deconditioning. The prototype VR program consisted of cycling, sitting, and standing balance exercises and a game for cognitive training.

Decreased muscle strength can significantly affect a person's functional capacity and the ability to perform activities of daily living, which decreases their independence and quality of life. In addition, loss of muscle strength in the elderly or patients is an important health indicator that is associated with many health conditions, such as mortality and risk of cardiovascular disease, causing rise to extensive societal costs [10]. Thus, the muscle strength test is the most critical element in the evaluation of physical function. However, direct muscle strength measurement from camera-based kinematic data is impracticable. Therefore, we need to develop a method to indirectly measure muscle strength using camera-based kinematic data in VR rehabilitation programs. Previous study has reported that the maximum and average power in cyclists is highly correlated with maximum muscle strength around the knee and hip joints [11]. In addition, another study demonstrated that the muscle thickness of the vastus lateralis, the major contributor to cycling ability, is the strongest predictor of maximum anaerobic power and is strongly correlated with the 30 seconds mean power and pedaling speed [12]. We devised a method to calculate the pedaling speed from motion data obtained while riding a stationary bike with full effort in a VR environment on the premise that the maximum speed of the bicycle can indirectly estimate the muscle strength of the lower extremities.

The main purpose of this study was to determine whether the converted pedaling speed from joint position data captured by a motion camera via a VR application could be used as an indirect method to evaluate muscle strength. For this, the pedaling speed obtained from the VR program using a bike and the muscle torque measured by the Isokinetic equipment were compared. Second, the study aimed to compare the pedaling speed measured by a motion capture camera with the actual pedaling speed and to investigate the validity and reliability of the camera-based VR program.

1. Participants

Thirty healthy adults (18 males and 12 females) without any known history of physical injury or illness affecting the study within the past 6 months were included. This study was approved by the Institutional Review Board at Yonsei University Mirae campus (IRB no. 1041849-202111-BM-197-03). The purpose and detailed procedure of this study were explained to the subjects, and they provided consent for voluntary participation. The physical characteristics of the patients are presented in Table 1.

Table 1 . General characteristics of the participants (N = 30).

VariableMale (n = 18)Female (n = 12)
Age (y)26.61 ± 2.3425.75 ± 3.61
Height (cm)174.61 ± 6.91160.0 ± 3.67
Weight (kg)73.57 ± 8.6252.64 ± 3.92
Body mass index (kg/m2)24.09 ± 2.0720.59 ± 1.71

Values are presented as mean ± standard deviation..



2. Instruments

A prototype VR rehabilitation system (RehaFit; JCMedilab, Chuncheon, Korea) was used in this experiment. This system was designed to evaluate and train individuals with decreased physical function secondary to aging, neuro-muscular diseases, and general deconditioning. RehaFit comprises of a 3D VR environment in which avatars interact with virtual objects. The scenes were created using 3D Max (Autodesk Inc., San Rafael, CA, USA) and imported into Unity 3D (Unity 2021.1.16; Unity Technologies, San Francisco, CA, USA) to create applications for VR exercises as well as VR-based assessments of maximum pedaling speed, sitting and standing balance ability, and cognitive function. In this experiment, a VR assessment program for maximum pedaling speed was selectively used. An avatar created in 3D Max ‘Cycle Game Man’ 3D model was animated according to the real time lower extremity joint positions captured by a motion camera. The avatar provides a third-person view of the user while cycling.

The pedaling speed assessment included a 3D motion capture camera (Astra Pro; Orbbec Co., Troy, MI, USA), 60-inch TV monitor, PC, and stationary bike (Moduta; Virchybike, Seoul, Korea) that automatically reports the pedaling speed using an embedded speedometer. The Orbbec astra pro camera used in this study is a 3D depth camera capable of depth tracking up to a range of 8 m. The camera has a resolution of 640*480 (30 fps) and detects depth with infrared coded structured light. We created a program using the Unity engine to automatically calculate the pedaling speed by detecting the position of the ankle joints in the frontal plane using a camera while cycling. In this program, the number of pedal rotations was increased one by one each time the left ankle joint triggered the collider, and the pedaling speed was calculated using the time per pedal rotation

An isokinetic dynamometer (Biodex Medical, Shirley, NY, USA) was used to assess the isometric and isokinetic peak torques of knee flexion and extension.

3. Procedures

The stationary bike was placed about approximately 2 m in front of the monitor using a motion capture camera. The subjects were instructed to stare at the monitor in the front while sitting on the bike without bending or leaning forward their upper body against it (Figure 1).

Figure 1. Stationary bike used in this study.

Prior to the 15 seconds maximum cycling tests, they rode the bike for 1 minute to get used to the stationary bike within a VR environment. Each subject performed the 15 seconds maximum pedaling test under two different conditions: “VR speedometer mode” and “built-in speedometer mode.” The “VR speedometer mode” is a test condition to obtain the pedaling speed, calculated from the recognized position and movement path of the ankle joint detected by the motion capture camera (Figure 2). The “built-in speedometer mode” is the test condition for obtaining the pedaling speed provided by the embedded speedometer of the stationary bike. After the start signal was given, the subjects pedaled at their maximum speed for 15 seconds. To obtain the average value of the maximum pedaling speed for 15 seconds, the subjects were asked to repeat 15 seconds of riding three times per condition with a 2 to 5 minutes recovery period between trials. The order of the condition was randomized to prevent order effects.

Figure 2. Motion capture in virtual reality speedometer mode.

Then, the participant sat on the Biodex dynamometer chair, and their trunk, pelvis, and thighs were secured with straps to minimize compensatory movements. After aligning the dynamometer axis to the participant’s knee joint axis, the calf was fixed using straps. The dynamometer was repeated three times in isometric and isokinetic modes. Isometric peak torque was measured separately three times for knee flexion and extension for 5 seconds while the knee angle was fixed at 60°. To measure isokinetic peak torque for knee flexion and extension, the subjects performed three continuous flexions and extensions reciprocally at an angular velocity of 90 °/s. The knee flexion and extension ranges of the isokinetic mode were set to the full range of motion of each participant.

4. Statistical Analysis

The collected data were statistically analyzed using Windows IBM SPSS Statistics 25.0 (IBM Co., Armonk, NY, USA). Intra-rater reliability was determined using intraclass correlation coefficient (ICC) values for a single measure and the associated 95% confidence intervals (CI). Reliability was defined as poor (ICC < 0.4), fair (ICC: 0.4–0.6), good (ICC: 0.6–0.75), or excellent (ICC > 0.75). Pearson correlation analysis was used to examine the correlation between the pedaling speeds measured using a VR and a built-in speedometer, and between the pedaling speeds and knee flexion/extension peak torque. The statistical significance level was set at 0.05.

1. Validity of the VR Speedometer for Measuring Maximal Pedaling Speed

The average pedaling speed of the VR speedometer was 6.75 ± 1.22 rpm and the average pedaling speed of built-in speedometer was 12.59 ± 2.23 rpm (Table 2). The results of the Pearson correlation analysis revealed a significantly high positive correlation between the maximum pedaling speed values of the VR and built-in speedometer (r = 0.922, p < 0.001) (Figure 3).

Table 2 . Pedaling speed in maximum cycling test.

VR
speedometer
Built-in
speedometer
Validity (r)
Pedaling speed (rpm)a6.75 ± 1.2212.59 ± 2.230.922

Values are presented as mean ± standard deviation. VR, virtual reality..

aAverage pedaling speed during 15 seconds maximum cycling test..



Figure 3. Scatter plot of the correlation between the VR speedometer and the built-in speedometer. VR, virtual reality.

2. Intra-rater Reliability of Pedaling Speed

Table 3 shows the intra-rater reliability of the three repeated measures with ICC (2,1) values. The VR speedometer had good intra-rater reliability (ICC = 0.685), and the built-in speedometer had an excellent level of reliability (ICC = 0.935).

Table 3 . Intra-rater reliability of pedaling speed.

ICC (95% CI)p-value
VR speedometer0.685 (0.511–0.821)< 0.001
Built-in speedometer0.935 (0.884–0.958)< 0.001

ICC, intraclass correlation coefficient; CI, confidence interval; VR, virtual reality..



3. Correlation Between Pedaling Speed and Knee Flexion/Extension Peak Torque

The knee flexion/extension peak torque in both isometric and isokinetic modes are shown in Table 3.

The results of the Pearson correlation analysis revealed a significant moderate positive correlation between the pedaling speed of the VR speedometer, and the peak torque of knee isometric flexion (r = 0.419, p = 0.021) and extension (r = 0.478, p = 0.008). The pedaling speed of the VR speedometer also had a moderately significant positive correlation with the peak torques of knee isokinetic flexion (r = 0.639, p < 0.001) and extension (r = 0.598, p < 0.001) (Table 4).

Table 4 . Knee flexion and extension peak torque measured using Biodex (Shirley) dynamometer.

Isometric (Nm)Isokinetic (Nm)


FlexionExtensionFlexionExtension
VR speedometer0.419 (0.021*)0.478 (0.008**)0.639 (0.000***)0.598 (0.000***)
Built-in speedometer0.444 (0.014*)0.424 (0.019*)0.652 (0.000***)0.558 (0.001**)

Values are presented as Pearson correlation coefficient [r] (p-value). *p < 0.05, **p < 0.01, ***p < 0.001..



The pedaling speed of the built-in speedometer also revealed a significant moderate positive correlation with the peak torque of isometric knee flexion (r = 0.444, p = 0.014) and extension (r = 0.424, p = 0.019). The pedaling speed of the built-in speedometer revealed a significant moderate positive correlation with the peak torque of knee isokinetic flexion (r = 0.652, p < 0.001) and extension (r = 0.558, p = 0.001) (Table 4).

The primary purpose of this study was to determine whether the converted pedaling speed from the ankle joint position data captured by a motion camera via a VR application can be used as an indirect way to assess muscle strength. The second purpose was to examine the validity and reliability of the converted pedaling speed from the joint position data captured by a motion camera via a VR application compared to the pedaling speed provided by the imbedded speedometer of the stationary bike. Thirty healthy young subjects without musculoskeletal disorders participated in the study.

First, we found that the validity of the 15 seconds maximum cycling test using VR speedometer compared to the value using the built-in speedometer was significantly high (r = 0.922, p < 0.001). The reliability and validity of the Orbbec camera have not been reported. However, the validity of the kinematic value measured using a depth sense camera, such as Kinect series, and the Vicon (Vicon Motion System, Oxford, UK) measured value, i.e., the gold standard of 3D motion analysis, is reported to be relatively high for various movements [13]. A previous study tested the validity of measured gait using Kinect v2 and evaluated the validity of the camera measurement compared to the gold standard Vicon; the correlation is very high (r = 0.970), and the camera-based measurement method has been proven to be effective [13,14]. High-end 3D motion analysis equipment with multiple optical cameras, such as Vicon and Qualisys (Qualisys, Göteborg, Sweden), provides accurate kinematic data during movement in laboratory environments; however, this equipment is expensive, and requires a long setup time and a specialist for operation and analysis. Conversely, low-cost depth cameras do not require sensors to be attached to the body. Therefore, it is more convenient to use camera-based evaluations for future entry-level VR rehabilitation programs [15].

Second, the intra-rater reliability of the VR speedometer mode was good (ICC = 0.685); however, this value was lower than that of the built-in speedometer mode (ICC = 0.935). These results suggest the feasibility of VR stationary cycling speed measurement using a low-cost depth camera. However, it also suggests the need for improvement and elaboration of the algorithm for estimating pedaling speed by recognizing the camera-sensed ankle joint and calculating the number of revolutions using the unity engine. A possible explanation for the lower reliability of the VR speedometer mode is that the camera perceives the effects of various environmental factors on the joint position. According to Grigg et al. [16], virtual marker tracking was affected by the subject's clothing color, lighting, and the contrasting colors of the bike and shoes, especially when wearing uni-colored clothing, and the standard deviation from marker recognition of the elbow was found to increase.

In this study, we also assessed the correlation between pedaling speed measured using the VR and built-in speedometer and peak torque value measured using Biodex, the gold standard for muscle torque measurement. We found that the pedaling speeds of both the VR and built-in speedometers were moderately correlated with the peak torque of isometric and isokinetic knee flexion/extension. It has been known that there is high correlation between gait speed and muscle strength [17-19]. Furthermore, Peñailillo et al. [20] found a correlation between maximum leg extension strength and sprinting performance in youth soccer players. As such, movements that used the lower extremity muscles as a whole were highly correlated with maximum muscle strength. In addition, previous study has shown that maximum and average power in cyclists is highly correlated with maximum muscle strength around the knee and hip joints [11]. Along with earlier research showing that the power and motion speed are highly correlated with maximum muscle strength the findings of our study support the idea that pedaling speed can be used as an evaluation method that indirectly reflects actual muscle strength. In addition, we found that the correlation coefficients were higher in isokinetic mode than in isometric mode. These results are because the repeated alternating flexion and extension of the lower extremity joints during cycling are more similar to the isokinetic mode, which requires alternating knee flexion and extension than the static isometric mode in Biodex.

This study had several limitations. First, because the subjects were all young and healthy people, it is difficult to generalize the correlation between the measured pedaling speed and muscle strength in different age groups. Therefore, future studies should collect data from subjects of different ages to confirm the correlation between pedaling speed and muscle strength. Second, a regression study is needed to predict muscle strength based on pedaling speed as well as other physical and demographic characteristics of individuals, such as, gender, age, height, weight, health, and disability status.

There was a high correlation between the pedaling speed measured by the camera in the VR environment and the actual pedaling speed. Pedaling speed and muscle peak torque showed moderate correlations. This study suggests the potential and benefits of measuring maximum pedaling speed using a 3D depth camera in a VR environment as an indirect assessment of lower extremity strength. However, technological improvements must be followed for the accurate estimation of muscle strength from the VR cycling test.

This work was supported by the National Research Fondation of Korea (NRF) grant funded by the Korea goverment (MSIT) (No.2021R1F1A1051369).

No potential conflict of interest relevant to this article was reported.

Conceptualization: YJK, HJ, JP. Data curation: YJK, JP, GAM, YW. Formal analysis: YJK, HJ, JP, GAM, YW. Funding acquisition: HJ. Investigation: YJK, JP, GAM, YW. Methodology: YJK, HJ, JP, GAM, YW. Project administration: YJK, HJ, JP. Resources: YJK, GAM, YW. Supervision: YJK, HJ, JP, GAM, YW. Validation: YJK, HJ. Visualization: YJK. Writing - original draft: YJK, HJ. Writing - review & editing: YJK, HJ.

  1. Canning CG, Allen NE, Nackaerts E, Paul SS, Nieuwboer A, Gilat M. Virtual reality in research and rehabilitation of gait and balance in Parkinson disease. Nat Rev Neurol 2020;16(8):409-25.
    Pubmed CrossRef
  2. Pazzaglia C, Imbimbo I, Tranchita E, Minganti C, Ricciardi D, Lo Monaco R, et al. Comparison of virtual reality rehabilitation and conventional rehabilitation in Parkinson's disease: a randomised controlled trial. Physiotherapy 2020;106:36-42.
    Pubmed CrossRef
  3. Triandafilou KM, Tsoupikova D, Barry AJ, Thielbar KN, Stoykov N, Kamper DG. Development of a 3D, networked multi-user virtual reality environment for home therapy after stroke. J Neuroeng Rehabil 2018;15(1):88.
    Pubmed KoreaMed CrossRef
  4. Asadzadeh A, Samad-Soltani T, Salahzadeh Z, Rezaei-Hachesu P. Effectiveness of virtual reality-based exercise therapy in rehabilitation: a scoping review. Inform Med Unlocked 2021;24:100562.
    CrossRef
  5. Prata MG, Scheicher ME. Correlation between balance and the level of functional independence among elderly people. Sao Paulo Med J 2012;130(2):97-101.
    Pubmed CrossRef
  6. Ye M, Yang C, Stankovic V, Stankovic L, Kerr K. A depth camera motion analysis framework for tele-rehabilitation: motion capture and person-centric kinematics analysis. IEEE J Sel Top Signal Process 2016;10(5):877-87.
    CrossRef
  7. Kritikos J, Mehmeti A, Nikolaou G, Koutsouris D. Fully portable low-cost motion capture system with real-time feedback for rehabilitation treatment. Paper presented at: 2019 International Conference on Virtual Rehabilitation (ICVR). Tel Aviv, Israel. Piscataway, NJ: IEEE, 2019; 2019;1-8.
    Pubmed CrossRef
  8. Vourvopoulos A, Jorge C, Abreu R, Figueiredo P, Fernandes JC, Bermúdez I Badia S. Efficacy and brain imaging correlates of an immersive motor imagery BCI-driven VR system for upper limb motor rehabilitation: a clinical case report. Front Hum Neurosci 2019;13:244.
    Pubmed KoreaMed CrossRef
  9. Omon K, Hara M, Ishikawa H. Virtual reality-guided, dual-task, body trunk balance training in the sitting position improved walking ability without improving leg strength. Prog Rehabil Med 2019;4:20190011.
    Pubmed KoreaMed CrossRef
  10. Kristensen OH, Stenager E, Dalgas U. Muscle strength and poststroke hemiplegia: a systematic review of muscle strength assessment and muscle strength impairment. Arch Phys Med Rehabil 2017;98(2):368-80. Erratum in: Arch Phys Med Rehabil 2017;98(6):1276.
    Pubmed CrossRef
  11. Park JH, Kim JE, Yoo JI, Kim YP, Kim EH, Seo TB. Comparison of maximum muscle strength and isokinetic knee and core muscle functions according to pedaling power difference of racing cyclist candidates. J Exerc Rehabil 2019;15(3):401-6.
    Pubmed KoreaMed CrossRef
  12. Lee HJ, Lee KW, Lee YW, Kim HJ. Correlation between cycling power and muscle thickness in cyclists. Clin Anat 2018;31(6):899-906.
    Pubmed CrossRef
  13. Kobsar D, Osis ST, Jacob C, Ferber R. Validity of a novel method to measure vertical oscillation during running using a depth camera. J Biomech 2019;85:182-6.
    Pubmed CrossRef
  14. Macpherson TW, Taylor J, McBain T, Weston M, Spears IR. Real-time measurement of pelvis and trunk kinematics during treadmill locomotion using a low-cost depth-sensing camera: a concurrent validity study. J Biomech 2016;49(3):474-8.
    Pubmed CrossRef
  15. Ota M, Tateuchi H, Hashiguchi T, Kato T, Ogino Y, Yamagata M, et al. Verification of reliability and validity of motion analysis systems during bilateral squat using human pose tracking algorithm. Gait Posture 2020;80:62-7.
    Pubmed CrossRef
  16. Grigg J, Haakonssen E, Rathbone E, Orr R, Keogh JWL. The validity and intra-tester reliability of markerless motion capture to analyse kinematics of the BMX Supercross gate start. Sports Biomech 2018;17(3):383-401.
    Pubmed CrossRef
  17. Hayashida I, Tanimoto Y, Takahashi Y, Kusabiraki T, Tamaki J. Correlation between muscle strength and muscle mass, and their association with walking speed, in community-dwelling elderly Japanese individuals. PLoS One 2014;9(11):e111810.
    Pubmed KoreaMed CrossRef
  18. Holviala J, Kraemer WJ, Sillanpää E, Karppinen H, Avela J, Kauhanen A, et al. Effects of strength, endurance and combined training on muscle strength, walking speed and dynamic balance in aging men. Eur J Appl Physiol 2012;112(4):1335-47.
    Pubmed CrossRef
  19. Tiainen K, Pajala S, Sipilä S, Kaprio J, Koskenvuo M, Alén M, et al. Genetic effects in common on maximal walking speed and muscle performance in older women. Scand J Med Sci Sports 2007;17(3):274-80.
    Pubmed CrossRef
  20. Peñailillo L, Espíldora F, Jannas-Vela S, Mujika I, Zbinden-Foncea H. Muscle strength and speed performance in youth soccer players. J Hum Kinet 2016;50:203-10.
    Pubmed KoreaMed CrossRef

Article

Original Article

Phys. Ther. Korea 2022; 29(4): 262-268

Published online November 20, 2022 https://doi.org/10.12674/ptk.2022.29.4.262

Copyright © Korean Research Society of Physical Therapy.

Correlation Between Knee Muscle Strength and Maximal Cycling Speed Measured Using 3D Depth Camera in Virtual Reality Environment

Ye Jin Kim1 , PT, BPT, Hye-seon Jeon2,3 , PT, PhD, Joo-hee Park2 , PT, PhD, Gyeong-Ah Moon1 , PT, BPT, Yixin Wang1 , PT, BPT

1Department of Physical Therapy, The Graduate School, Yonsei University, 2Department of Physical Therapy, College of Health Science, Yonsei University, 3Department of Ergonomic Therapy, The Graduate School of Health and Environment, Yonsei University, Wonju, Korea

Correspondence to:Hye-seon Jeon
E-mail: hyeseonj@yonsei.ac.kr
https://orcid.org/0000-0003-3986-2030

Received: October 14, 2022; Revised: November 4, 2022; Accepted: November 4, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background: Virtual reality (VR) programs based on motion capture camera are the most convenient and cost-effective approaches for remote rehabilitation. Assessment of physical function is critical for providing optimal VR rehabilitation training; however, direct muscle strength measurement using camera-based kinematic data is impracticable. Therefore, it is necessary to develop a method to indirectly estimate the muscle strength of users from the value obtained using a motion capture camera. Objects: The purpose of this study was to determine whether the pedaling speed converted using the VR engine from the captured foot position data in the VR environment can be used as an indirect way to evaluate knee muscle strength, and to investigate the validity and reliability of a camera-based VR program.
Methods: Thirty healthy adults were included in this study. Each subject performed a 15-second maximum pedaling test in the VR and built-in speedometer modes. In the VR speedometer mode, a motion capture camera was used to detect the position of the ankle joints and automatically calculate the pedaling speed. An isokinetic dynamometer was used to assess the isometric and isokinetic peak torques of knee flexion and extension.
Results: The pedaling speeds in VR and built-in speedometer modes revealed a significantly high positive correlation (r = 0.922). In addition, the intra-rater reliability of the pedaling speed in the VR speedometer mode was good (ICC [intraclass correlation coefficient] = 0.685). The results of the Pearson correlation analysis revealed a significant moderate positive correlation between the pedaling speed of the VR speedometer and the peak torque of knee isokinetic flexion (r = 0.639) and extension (r = 0.598).
Conclusion: This study suggests the potential benefits of measuring the maximum pedaling speed using 3D depth camera in a VR environment as an indirect assessment of muscle strength. However, technological improvements must be followed to obtain more accurate estimation of muscle strength from the VR cycling test.

Keywords: Bicycling, Muscle strength, Rehabilitation, Virtual reality

INTRODUCTION

Virtual reality (VR) is the application of visual simulations created with a computer program that imitates the real world or imaginable environments, items, and experiences in real time, and it requires interactivity through ongoing movements of the user [1]. VR-based rehabilitation has emerged as a new treatment option in various aspects of modern rehabilitation settings because it can optimize motor learning in a safe environment and help improve functional activities of daily living by replicating real-life scenarios [2]. VR rehabilitation also maximizes a sense of accomplishment and motivation for successive attempts, and these VR programs frequently include various exercise components, such as strength, stretching, endurance, and cardiorespiratory fitness, to improve the level of functional activities [3-5].

VR rehabilitation programs can be utilized for the remote rehabilitation of patients with time and geographic restrictions. The importance of remote rehabilitation has been highlighted during the COVID pandemic. A VR-based remote rehabilitation program equipped with an appropriate assessment function has the potential to provide an individualized and customized rehabilitation program without the need of a rehabilitation specialist. In addition, the therapist can remotely check the physical functions measured by the VR program at home and recommend a home exercise program. Motion capture systems traditionally used for rehabilitation consist of multiple sensors and need a large laboratory space. It is also expensive and not portable, so it is not appropriate for tele-rehabilitation or home-use rehabilitation programs [6]. Therefore, motion capture system-based VR programs using depth cameras such as Kinect (Microsoft, Redmond, WA, USA) are among the most convenient and cost-effective approaches for remote VR rehabilitation systems.

VR rehabilitation provides a gamified experience by creating assorted virtual environments during rehabilitation sessions for patients through monitors or VR tools [7]. To provide a proper level of challenge to individual users, the task difficulty of the VR game exercises must be modifiable to optimize motor learning according to the findings from assessments [3]. However, most of the currently developed VR games in rehabilitation do not include individualized assessments of the function and health status of users within the VR program [8,9]. Therefore, we developed a camera-based VR rehabilitation program that is capable of both evaluating physical status and providing customized rehabilitation exercises for individuals with decreased physical function secondary to aging, neuro-muscular diseases, and general deconditioning. The prototype VR program consisted of cycling, sitting, and standing balance exercises and a game for cognitive training.

Decreased muscle strength can significantly affect a person's functional capacity and the ability to perform activities of daily living, which decreases their independence and quality of life. In addition, loss of muscle strength in the elderly or patients is an important health indicator that is associated with many health conditions, such as mortality and risk of cardiovascular disease, causing rise to extensive societal costs [10]. Thus, the muscle strength test is the most critical element in the evaluation of physical function. However, direct muscle strength measurement from camera-based kinematic data is impracticable. Therefore, we need to develop a method to indirectly measure muscle strength using camera-based kinematic data in VR rehabilitation programs. Previous study has reported that the maximum and average power in cyclists is highly correlated with maximum muscle strength around the knee and hip joints [11]. In addition, another study demonstrated that the muscle thickness of the vastus lateralis, the major contributor to cycling ability, is the strongest predictor of maximum anaerobic power and is strongly correlated with the 30 seconds mean power and pedaling speed [12]. We devised a method to calculate the pedaling speed from motion data obtained while riding a stationary bike with full effort in a VR environment on the premise that the maximum speed of the bicycle can indirectly estimate the muscle strength of the lower extremities.

The main purpose of this study was to determine whether the converted pedaling speed from joint position data captured by a motion camera via a VR application could be used as an indirect method to evaluate muscle strength. For this, the pedaling speed obtained from the VR program using a bike and the muscle torque measured by the Isokinetic equipment were compared. Second, the study aimed to compare the pedaling speed measured by a motion capture camera with the actual pedaling speed and to investigate the validity and reliability of the camera-based VR program.

MATERIALS AND METHODS

1. Participants

Thirty healthy adults (18 males and 12 females) without any known history of physical injury or illness affecting the study within the past 6 months were included. This study was approved by the Institutional Review Board at Yonsei University Mirae campus (IRB no. 1041849-202111-BM-197-03). The purpose and detailed procedure of this study were explained to the subjects, and they provided consent for voluntary participation. The physical characteristics of the patients are presented in Table 1.

Table 1 . General characteristics of the participants (N = 30).

VariableMale (n = 18)Female (n = 12)
Age (y)26.61 ± 2.3425.75 ± 3.61
Height (cm)174.61 ± 6.91160.0 ± 3.67
Weight (kg)73.57 ± 8.6252.64 ± 3.92
Body mass index (kg/m2)24.09 ± 2.0720.59 ± 1.71

Values are presented as mean ± standard deviation..



2. Instruments

A prototype VR rehabilitation system (RehaFit; JCMedilab, Chuncheon, Korea) was used in this experiment. This system was designed to evaluate and train individuals with decreased physical function secondary to aging, neuro-muscular diseases, and general deconditioning. RehaFit comprises of a 3D VR environment in which avatars interact with virtual objects. The scenes were created using 3D Max (Autodesk Inc., San Rafael, CA, USA) and imported into Unity 3D (Unity 2021.1.16; Unity Technologies, San Francisco, CA, USA) to create applications for VR exercises as well as VR-based assessments of maximum pedaling speed, sitting and standing balance ability, and cognitive function. In this experiment, a VR assessment program for maximum pedaling speed was selectively used. An avatar created in 3D Max ‘Cycle Game Man’ 3D model was animated according to the real time lower extremity joint positions captured by a motion camera. The avatar provides a third-person view of the user while cycling.

The pedaling speed assessment included a 3D motion capture camera (Astra Pro; Orbbec Co., Troy, MI, USA), 60-inch TV monitor, PC, and stationary bike (Moduta; Virchybike, Seoul, Korea) that automatically reports the pedaling speed using an embedded speedometer. The Orbbec astra pro camera used in this study is a 3D depth camera capable of depth tracking up to a range of 8 m. The camera has a resolution of 640*480 (30 fps) and detects depth with infrared coded structured light. We created a program using the Unity engine to automatically calculate the pedaling speed by detecting the position of the ankle joints in the frontal plane using a camera while cycling. In this program, the number of pedal rotations was increased one by one each time the left ankle joint triggered the collider, and the pedaling speed was calculated using the time per pedal rotation

An isokinetic dynamometer (Biodex Medical, Shirley, NY, USA) was used to assess the isometric and isokinetic peak torques of knee flexion and extension.

3. Procedures

The stationary bike was placed about approximately 2 m in front of the monitor using a motion capture camera. The subjects were instructed to stare at the monitor in the front while sitting on the bike without bending or leaning forward their upper body against it (Figure 1).

Figure 1. Stationary bike used in this study.

Prior to the 15 seconds maximum cycling tests, they rode the bike for 1 minute to get used to the stationary bike within a VR environment. Each subject performed the 15 seconds maximum pedaling test under two different conditions: “VR speedometer mode” and “built-in speedometer mode.” The “VR speedometer mode” is a test condition to obtain the pedaling speed, calculated from the recognized position and movement path of the ankle joint detected by the motion capture camera (Figure 2). The “built-in speedometer mode” is the test condition for obtaining the pedaling speed provided by the embedded speedometer of the stationary bike. After the start signal was given, the subjects pedaled at their maximum speed for 15 seconds. To obtain the average value of the maximum pedaling speed for 15 seconds, the subjects were asked to repeat 15 seconds of riding three times per condition with a 2 to 5 minutes recovery period between trials. The order of the condition was randomized to prevent order effects.

Figure 2. Motion capture in virtual reality speedometer mode.

Then, the participant sat on the Biodex dynamometer chair, and their trunk, pelvis, and thighs were secured with straps to minimize compensatory movements. After aligning the dynamometer axis to the participant’s knee joint axis, the calf was fixed using straps. The dynamometer was repeated three times in isometric and isokinetic modes. Isometric peak torque was measured separately three times for knee flexion and extension for 5 seconds while the knee angle was fixed at 60°. To measure isokinetic peak torque for knee flexion and extension, the subjects performed three continuous flexions and extensions reciprocally at an angular velocity of 90 °/s. The knee flexion and extension ranges of the isokinetic mode were set to the full range of motion of each participant.

4. Statistical Analysis

The collected data were statistically analyzed using Windows IBM SPSS Statistics 25.0 (IBM Co., Armonk, NY, USA). Intra-rater reliability was determined using intraclass correlation coefficient (ICC) values for a single measure and the associated 95% confidence intervals (CI). Reliability was defined as poor (ICC < 0.4), fair (ICC: 0.4–0.6), good (ICC: 0.6–0.75), or excellent (ICC > 0.75). Pearson correlation analysis was used to examine the correlation between the pedaling speeds measured using a VR and a built-in speedometer, and between the pedaling speeds and knee flexion/extension peak torque. The statistical significance level was set at 0.05.

RESULTS

1. Validity of the VR Speedometer for Measuring Maximal Pedaling Speed

The average pedaling speed of the VR speedometer was 6.75 ± 1.22 rpm and the average pedaling speed of built-in speedometer was 12.59 ± 2.23 rpm (Table 2). The results of the Pearson correlation analysis revealed a significantly high positive correlation between the maximum pedaling speed values of the VR and built-in speedometer (r = 0.922, p < 0.001) (Figure 3).

Table 2 . Pedaling speed in maximum cycling test.

VR
speedometer
Built-in
speedometer
Validity (r)
Pedaling speed (rpm)a6.75 ± 1.2212.59 ± 2.230.922

Values are presented as mean ± standard deviation. VR, virtual reality..

aAverage pedaling speed during 15 seconds maximum cycling test..



Figure 3. Scatter plot of the correlation between the VR speedometer and the built-in speedometer. VR, virtual reality.

2. Intra-rater Reliability of Pedaling Speed

Table 3 shows the intra-rater reliability of the three repeated measures with ICC (2,1) values. The VR speedometer had good intra-rater reliability (ICC = 0.685), and the built-in speedometer had an excellent level of reliability (ICC = 0.935).

Table 3 . Intra-rater reliability of pedaling speed.

ICC (95% CI)p-value
VR speedometer0.685 (0.511–0.821)< 0.001
Built-in speedometer0.935 (0.884–0.958)< 0.001

ICC, intraclass correlation coefficient; CI, confidence interval; VR, virtual reality..



3. Correlation Between Pedaling Speed and Knee Flexion/Extension Peak Torque

The knee flexion/extension peak torque in both isometric and isokinetic modes are shown in Table 3.

The results of the Pearson correlation analysis revealed a significant moderate positive correlation between the pedaling speed of the VR speedometer, and the peak torque of knee isometric flexion (r = 0.419, p = 0.021) and extension (r = 0.478, p = 0.008). The pedaling speed of the VR speedometer also had a moderately significant positive correlation with the peak torques of knee isokinetic flexion (r = 0.639, p < 0.001) and extension (r = 0.598, p < 0.001) (Table 4).

Table 4 . Knee flexion and extension peak torque measured using Biodex (Shirley) dynamometer.

Isometric (Nm)Isokinetic (Nm)


FlexionExtensionFlexionExtension
VR speedometer0.419 (0.021*)0.478 (0.008**)0.639 (0.000***)0.598 (0.000***)
Built-in speedometer0.444 (0.014*)0.424 (0.019*)0.652 (0.000***)0.558 (0.001**)

Values are presented as Pearson correlation coefficient [r] (p-value). *p < 0.05, **p < 0.01, ***p < 0.001..



The pedaling speed of the built-in speedometer also revealed a significant moderate positive correlation with the peak torque of isometric knee flexion (r = 0.444, p = 0.014) and extension (r = 0.424, p = 0.019). The pedaling speed of the built-in speedometer revealed a significant moderate positive correlation with the peak torque of knee isokinetic flexion (r = 0.652, p < 0.001) and extension (r = 0.558, p = 0.001) (Table 4).

DISCUSSION

The primary purpose of this study was to determine whether the converted pedaling speed from the ankle joint position data captured by a motion camera via a VR application can be used as an indirect way to assess muscle strength. The second purpose was to examine the validity and reliability of the converted pedaling speed from the joint position data captured by a motion camera via a VR application compared to the pedaling speed provided by the imbedded speedometer of the stationary bike. Thirty healthy young subjects without musculoskeletal disorders participated in the study.

First, we found that the validity of the 15 seconds maximum cycling test using VR speedometer compared to the value using the built-in speedometer was significantly high (r = 0.922, p < 0.001). The reliability and validity of the Orbbec camera have not been reported. However, the validity of the kinematic value measured using a depth sense camera, such as Kinect series, and the Vicon (Vicon Motion System, Oxford, UK) measured value, i.e., the gold standard of 3D motion analysis, is reported to be relatively high for various movements [13]. A previous study tested the validity of measured gait using Kinect v2 and evaluated the validity of the camera measurement compared to the gold standard Vicon; the correlation is very high (r = 0.970), and the camera-based measurement method has been proven to be effective [13,14]. High-end 3D motion analysis equipment with multiple optical cameras, such as Vicon and Qualisys (Qualisys, Göteborg, Sweden), provides accurate kinematic data during movement in laboratory environments; however, this equipment is expensive, and requires a long setup time and a specialist for operation and analysis. Conversely, low-cost depth cameras do not require sensors to be attached to the body. Therefore, it is more convenient to use camera-based evaluations for future entry-level VR rehabilitation programs [15].

Second, the intra-rater reliability of the VR speedometer mode was good (ICC = 0.685); however, this value was lower than that of the built-in speedometer mode (ICC = 0.935). These results suggest the feasibility of VR stationary cycling speed measurement using a low-cost depth camera. However, it also suggests the need for improvement and elaboration of the algorithm for estimating pedaling speed by recognizing the camera-sensed ankle joint and calculating the number of revolutions using the unity engine. A possible explanation for the lower reliability of the VR speedometer mode is that the camera perceives the effects of various environmental factors on the joint position. According to Grigg et al. [16], virtual marker tracking was affected by the subject's clothing color, lighting, and the contrasting colors of the bike and shoes, especially when wearing uni-colored clothing, and the standard deviation from marker recognition of the elbow was found to increase.

In this study, we also assessed the correlation between pedaling speed measured using the VR and built-in speedometer and peak torque value measured using Biodex, the gold standard for muscle torque measurement. We found that the pedaling speeds of both the VR and built-in speedometers were moderately correlated with the peak torque of isometric and isokinetic knee flexion/extension. It has been known that there is high correlation between gait speed and muscle strength [17-19]. Furthermore, Peñailillo et al. [20] found a correlation between maximum leg extension strength and sprinting performance in youth soccer players. As such, movements that used the lower extremity muscles as a whole were highly correlated with maximum muscle strength. In addition, previous study has shown that maximum and average power in cyclists is highly correlated with maximum muscle strength around the knee and hip joints [11]. Along with earlier research showing that the power and motion speed are highly correlated with maximum muscle strength the findings of our study support the idea that pedaling speed can be used as an evaluation method that indirectly reflects actual muscle strength. In addition, we found that the correlation coefficients were higher in isokinetic mode than in isometric mode. These results are because the repeated alternating flexion and extension of the lower extremity joints during cycling are more similar to the isokinetic mode, which requires alternating knee flexion and extension than the static isometric mode in Biodex.

This study had several limitations. First, because the subjects were all young and healthy people, it is difficult to generalize the correlation between the measured pedaling speed and muscle strength in different age groups. Therefore, future studies should collect data from subjects of different ages to confirm the correlation between pedaling speed and muscle strength. Second, a regression study is needed to predict muscle strength based on pedaling speed as well as other physical and demographic characteristics of individuals, such as, gender, age, height, weight, health, and disability status.

CONCLUSIONS

There was a high correlation between the pedaling speed measured by the camera in the VR environment and the actual pedaling speed. Pedaling speed and muscle peak torque showed moderate correlations. This study suggests the potential and benefits of measuring maximum pedaling speed using a 3D depth camera in a VR environment as an indirect assessment of lower extremity strength. However, technological improvements must be followed for the accurate estimation of muscle strength from the VR cycling test.

ACKNOWLEDGEMENTS

None.

FUNDING

This work was supported by the National Research Fondation of Korea (NRF) grant funded by the Korea goverment (MSIT) (No.2021R1F1A1051369).

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

AUTHOR CONTRIBUTIONS

Conceptualization: YJK, HJ, JP. Data curation: YJK, JP, GAM, YW. Formal analysis: YJK, HJ, JP, GAM, YW. Funding acquisition: HJ. Investigation: YJK, JP, GAM, YW. Methodology: YJK, HJ, JP, GAM, YW. Project administration: YJK, HJ, JP. Resources: YJK, GAM, YW. Supervision: YJK, HJ, JP, GAM, YW. Validation: YJK, HJ. Visualization: YJK. Writing - original draft: YJK, HJ. Writing - review & editing: YJK, HJ.

Fig 1.

Figure 1.Stationary bike used in this study.
Physical Therapy Korea 2022; 29: 262-268https://doi.org/10.12674/ptk.2022.29.4.262

Fig 2.

Figure 2.Motion capture in virtual reality speedometer mode.
Physical Therapy Korea 2022; 29: 262-268https://doi.org/10.12674/ptk.2022.29.4.262

Fig 3.

Figure 3.Scatter plot of the correlation between the VR speedometer and the built-in speedometer. VR, virtual reality.
Physical Therapy Korea 2022; 29: 262-268https://doi.org/10.12674/ptk.2022.29.4.262

Table 1 . General characteristics of the participants (N = 30).

VariableMale (n = 18)Female (n = 12)
Age (y)26.61 ± 2.3425.75 ± 3.61
Height (cm)174.61 ± 6.91160.0 ± 3.67
Weight (kg)73.57 ± 8.6252.64 ± 3.92
Body mass index (kg/m2)24.09 ± 2.0720.59 ± 1.71

Values are presented as mean ± standard deviation..


Table 2 . Pedaling speed in maximum cycling test.

VR
speedometer
Built-in
speedometer
Validity (r)
Pedaling speed (rpm)a6.75 ± 1.2212.59 ± 2.230.922

Values are presented as mean ± standard deviation. VR, virtual reality..

aAverage pedaling speed during 15 seconds maximum cycling test..


Table 3 . Intra-rater reliability of pedaling speed.

ICC (95% CI)p-value
VR speedometer0.685 (0.511–0.821)< 0.001
Built-in speedometer0.935 (0.884–0.958)< 0.001

ICC, intraclass correlation coefficient; CI, confidence interval; VR, virtual reality..


Table 4 . Knee flexion and extension peak torque measured using Biodex (Shirley) dynamometer.

Isometric (Nm)Isokinetic (Nm)


FlexionExtensionFlexionExtension
VR speedometer0.419 (0.021*)0.478 (0.008**)0.639 (0.000***)0.598 (0.000***)
Built-in speedometer0.444 (0.014*)0.424 (0.019*)0.652 (0.000***)0.558 (0.001**)

Values are presented as Pearson correlation coefficient [r] (p-value). *p < 0.05, **p < 0.01, ***p < 0.001..


References

  1. Canning CG, Allen NE, Nackaerts E, Paul SS, Nieuwboer A, Gilat M. Virtual reality in research and rehabilitation of gait and balance in Parkinson disease. Nat Rev Neurol 2020;16(8):409-25.
    Pubmed CrossRef
  2. Pazzaglia C, Imbimbo I, Tranchita E, Minganti C, Ricciardi D, Lo Monaco R, et al. Comparison of virtual reality rehabilitation and conventional rehabilitation in Parkinson's disease: a randomised controlled trial. Physiotherapy 2020;106:36-42.
    Pubmed CrossRef
  3. Triandafilou KM, Tsoupikova D, Barry AJ, Thielbar KN, Stoykov N, Kamper DG. Development of a 3D, networked multi-user virtual reality environment for home therapy after stroke. J Neuroeng Rehabil 2018;15(1):88.
    Pubmed KoreaMed CrossRef
  4. Asadzadeh A, Samad-Soltani T, Salahzadeh Z, Rezaei-Hachesu P. Effectiveness of virtual reality-based exercise therapy in rehabilitation: a scoping review. Inform Med Unlocked 2021;24:100562.
    CrossRef
  5. Prata MG, Scheicher ME. Correlation between balance and the level of functional independence among elderly people. Sao Paulo Med J 2012;130(2):97-101.
    Pubmed CrossRef
  6. Ye M, Yang C, Stankovic V, Stankovic L, Kerr K. A depth camera motion analysis framework for tele-rehabilitation: motion capture and person-centric kinematics analysis. IEEE J Sel Top Signal Process 2016;10(5):877-87.
    CrossRef
  7. Kritikos J, Mehmeti A, Nikolaou G, Koutsouris D. Fully portable low-cost motion capture system with real-time feedback for rehabilitation treatment. Paper presented at: 2019 International Conference on Virtual Rehabilitation (ICVR). Tel Aviv, Israel. Piscataway, NJ: IEEE, 2019; 2019;1-8.
    Pubmed CrossRef
  8. Vourvopoulos A, Jorge C, Abreu R, Figueiredo P, Fernandes JC, Bermúdez I Badia S. Efficacy and brain imaging correlates of an immersive motor imagery BCI-driven VR system for upper limb motor rehabilitation: a clinical case report. Front Hum Neurosci 2019;13:244.
    Pubmed KoreaMed CrossRef
  9. Omon K, Hara M, Ishikawa H. Virtual reality-guided, dual-task, body trunk balance training in the sitting position improved walking ability without improving leg strength. Prog Rehabil Med 2019;4:20190011.
    Pubmed KoreaMed CrossRef
  10. Kristensen OH, Stenager E, Dalgas U. Muscle strength and poststroke hemiplegia: a systematic review of muscle strength assessment and muscle strength impairment. Arch Phys Med Rehabil 2017;98(2):368-80. Erratum in: Arch Phys Med Rehabil 2017;98(6):1276.
    Pubmed CrossRef
  11. Park JH, Kim JE, Yoo JI, Kim YP, Kim EH, Seo TB. Comparison of maximum muscle strength and isokinetic knee and core muscle functions according to pedaling power difference of racing cyclist candidates. J Exerc Rehabil 2019;15(3):401-6.
    Pubmed KoreaMed CrossRef
  12. Lee HJ, Lee KW, Lee YW, Kim HJ. Correlation between cycling power and muscle thickness in cyclists. Clin Anat 2018;31(6):899-906.
    Pubmed CrossRef
  13. Kobsar D, Osis ST, Jacob C, Ferber R. Validity of a novel method to measure vertical oscillation during running using a depth camera. J Biomech 2019;85:182-6.
    Pubmed CrossRef
  14. Macpherson TW, Taylor J, McBain T, Weston M, Spears IR. Real-time measurement of pelvis and trunk kinematics during treadmill locomotion using a low-cost depth-sensing camera: a concurrent validity study. J Biomech 2016;49(3):474-8.
    Pubmed CrossRef
  15. Ota M, Tateuchi H, Hashiguchi T, Kato T, Ogino Y, Yamagata M, et al. Verification of reliability and validity of motion analysis systems during bilateral squat using human pose tracking algorithm. Gait Posture 2020;80:62-7.
    Pubmed CrossRef
  16. Grigg J, Haakonssen E, Rathbone E, Orr R, Keogh JWL. The validity and intra-tester reliability of markerless motion capture to analyse kinematics of the BMX Supercross gate start. Sports Biomech 2018;17(3):383-401.
    Pubmed CrossRef
  17. Hayashida I, Tanimoto Y, Takahashi Y, Kusabiraki T, Tamaki J. Correlation between muscle strength and muscle mass, and their association with walking speed, in community-dwelling elderly Japanese individuals. PLoS One 2014;9(11):e111810.
    Pubmed KoreaMed CrossRef
  18. Holviala J, Kraemer WJ, Sillanpää E, Karppinen H, Avela J, Kauhanen A, et al. Effects of strength, endurance and combined training on muscle strength, walking speed and dynamic balance in aging men. Eur J Appl Physiol 2012;112(4):1335-47.
    Pubmed CrossRef
  19. Tiainen K, Pajala S, Sipilä S, Kaprio J, Koskenvuo M, Alén M, et al. Genetic effects in common on maximal walking speed and muscle performance in older women. Scand J Med Sci Sports 2007;17(3):274-80.
    Pubmed CrossRef
  20. Peñailillo L, Espíldora F, Jannas-Vela S, Mujika I, Zbinden-Foncea H. Muscle strength and speed performance in youth soccer players. J Hum Kinet 2016;50:203-10.
    Pubmed KoreaMed CrossRef