|Year : 2022 | Volume
| Issue : 3 | Page : 115-119
Inertial measurement unit-based functional evaluation for adhesive capsulitis assessment
Chih-Ya Chang1, Yung-Tsan Wu2, Ching-Yueh Lin3, Te-Jung Liu4, Tsung-Yen Ho5, Yu-Ping Shen2, Kai-Chun Liu6, Ting-Yang Lu7, Li-Wei Chou8
1 Department of Physical Medicine and Rehabilitation, School of Medicine, National Defense Medical Center, Tri-Service General Hospital; Department of Physical Therapy and Assistive Technology, National Yang-Ming Chiao Tung University, Taipei, Taiwan
2 Department of Physical Medicine and Rehabilitation, School of Medicine, National Defense Medical Center, Tri-Service General Hospital, Taipei, Taiwan
3 Department of Physical Medicine and Rehabilitation, School of Medicine, National Defense Medical Center, Tri-Service General Hospital, Taipei; Department of Physical Medicine and Rehabilitation, Kaohsiung Armed Forces General Hospital, Kaohsiung, Taiwan
4 Department of Physical Medicine and Rehabilitation, School of Medicine, National Defense Medical Center, Tri-Service General Hospital, Taipei; Department of Physical Medicine and Rehabilitation, Taoyuan Armed Force General Hospital, Taoyuan, Taiwan
5 Department of Physical Medicine and Rehabilitation, School of Medicine, National Defense Medical Center, Tri-Service General Hospital, Taipei; Department of Physical Medicine and Rehabilitation, Taichung Armed Forces General Hospital, Taiping District, Taichung City, Taiwan
6 Research Center for Information Technology Innovation, Academia Sinica; Departments of Biomedical Engineering, National Yang Ming Chiao Tung University, Taipei, Taiwan
7 Departments of Biomedical Engineering, National Yang Ming Chiao Tung University, Taipei, Taiwan
8 Department of Physical Therapy and Assistive Technology, National Yang-Ming Chiao Tung University, Taipei, Taiwan
|Date of Submission||10-Mar-2021|
|Date of Decision||26-Mar-2021|
|Date of Acceptance||27-Mar-2021|
|Date of Web Publication||28-May-2021|
Dr. Li-Wei Chou
Department of Physical Therapy and Assistive Technology, National Yang Ming Chiao Tung University, Taipei
Source of Support: None, Conflict of Interest: None
Aims: The inertial measurement unit (IMU), as a sensor-based assessment tool, could provide objective and quantitative data for evaluating a patient with adhesive capsulitis (AC). The IMUs have advantages in simplification of implementation, cost, and computation complexity. We aimed to propose an IMU-based approach to extract statistical features for the assessment of AC in daily activity. Methods: Nine healthy subjects and nine AC patients participate in this experiment. The accelerometers are placed on the wrist and arm to measure the movement performance. Each subject is asked to perform three basic shoulder motions, including flexion, extension, and abduction. Eight types of features are extracted from the norm of accelerometer signals, including mean, standard deviation (SD), variation, maximum, minimum, range, kurtosis, and skewness. These features are explored to distinguish the differences in the movement performance between healthy subjects and AC patients. Statistical Analysis Used: Student's t-test and effect size (Cohen's d) are calculated to assess the reliability of the proposed evaluation approach. Results: The results show that the feature of SD extracted from the wrist can achieve significant differences and large effect sizes between healthy subjects and AC patients. Conclusion: We demonstrate the feasibility of the proposed IMU-based functional evaluation for the AC assessment using statistical features.
Keywords: Inertial measurement units, statistical features, adhesive capsulitis, Student's t-test, effect size
|How to cite this article:|
Chang CY, Wu YT, Lin CY, Liu TJ, Ho TY, Shen YP, Liu KC, Lu TY, Chou LW. Inertial measurement unit-based functional evaluation for adhesive capsulitis assessment. J Med Sci 2022;42:115-9
|How to cite this URL:|
Chang CY, Wu YT, Lin CY, Liu TJ, Ho TY, Shen YP, Liu KC, Lu TY, Chou LW. Inertial measurement unit-based functional evaluation for adhesive capsulitis assessment. J Med Sci [serial online] 2022 [cited 2022 Jul 3];42:115-9. Available from: https://www.jmedscindmc.com/text.asp?2022/42/3/115/317115
| Introduction|| |
Adhesive capsulitis (AC), also known as frozen shoulder, is a common joint condition that causes stiffness and pain among people aged 40–65 years, high peak at 50 years and especially for women., The stiffness and pain of shoulder joints lead the limitation to the range of motion (ROM) in all movement planes of shoulder joints. AC has great impacts on the quality of daily life and activities of daily living (ADL) performance., The common treatments in AC patients include medicine, physical therapy, local injection, or surgical approach which aim to relieve pain, improve joint mobility, and increase the independent ability.,, To support clinical evaluations and measurements, there is a requirement of objective and continuous assessment for helping clinical management and follow-up progresses.,
In recent years, various sensor technologies have been applied to assess the function of the shoulder, including vision-based and wearable-based approaches. These approaches have the potential to provide continuous, objective, and reliable motion and movement information for clinical professionals.,,, The vision-based approach used cameras to catch movement and skeleton information for shoulder joint calculation. Fernández-Baena et al. proposed shoulder joint calculation systems using Kinect sensors. This study exanimated the accuracy of the vision-based systems on the measurement of shoulder ROM. The results showed that the vision-based sensor can provide reliable ROM information for clinical usage. However, the vision-based approach was sensitive to the change of the ambient light and required higher computation resources. Another wearable-based approach was to place inertial measurement units (IMUs) on the human body to measure the joint angle and functional performance. Several studies used multiple IMUs to calculate clinical scores, including P score, RAV score, and M score on patients with shoulder disease.,,,,, The results showed that the calculated kinematic scores had significant differences between healthy subjects and patients. The wearable-based approach is beneficial for designing lightweight, small-size, and low-cost systems.
In this study, we aimed to propose an IMU-based approach to extract statistical features for AC function assessment. The evaluated clinical parameters involving statistical features (e.g., mean and standard deviation [SD]) were utilized to differentiate healthy adults and AC patients. The IMU-based approach had the potential to provide clinical professionals, with an objective approach for AC assessment.
| Materials and Methods|| |
Participants were outpatients at a rehabilitation department of Tri-Service General Hospital who were diagnosed with AC. The patients were included if they have shoulder pain with a limited ROM more than 3 months and aged 30–65 years. Participants were diagnosed with primary AC according to the standardized history, physical examination, and ultrasonographic evaluation by an experiment physiatrist. Patients were excluded if they had any of the following: full or massive thickness tear of the rotator cuff on ultrasonography or magnetic resonance imaging; secondary AC (secondary to other causes, including metabolic, rheumatic, or infectious arthritis, stroke, cancer, or fracture); and acute cervical radiculopathy.
The study was approved by the institutional review board (TSGHIRB No.: A202005024) at the university hospital, and all participants gave written informed consent. Our research procedure followed the Declaration of Helsinki. All participants were assured that their participation was entirely voluntary and that they could withdraw at any time.
Nine healthy subjects and nine AC patients participated in this experiment. Each subject was asked to perform three basic shoulder motions, including flexion (F), extension (E), and abduction (A). An illustration of three basic tasks is shown in [Figure 1]. During the test, subjects were free to perform the task at their preferred speed and in the maximum range of the movement. One camera with a sampling rate of 30 fps was placed in front of the subjects during the experiment, which is utilized as the reference system to assess the movement performance.
|Figure 1: An illustration of three basic shoulder motions. (a) Flexion, (b) extension, (c) abduction|
Click here to view
This work placed two IMUs on the wrist and arm while the subjects performed the shoulder tasks, which is shown in [Figure 2]. The IMU named OPAL published by APDM Inc. including accelerometers is applied to measure the shoulder movement in this work. The sampling rate of IMUs is set as 128 Hz.
In this work, different statistical parameters are selected to estimate the differences between healthy subjects and AC patients, including mean, SD, variation, maximum, minimum, range, kurtosis, and skewness. These features can extract the features of movement and reflect characteristics for clinical evaluation. They have been utilized in different clinical assessments for movement analysis and classification.
Descriptive statistics, including mean and SD, were evaluated for healthy subjects and AC patients in different parameters. The Student's t-test is used to investigate the meaningful differences between two groups, and the significant level α is set at 0.05. Furthermore, the effect size (d) proposed by Cohen is calculated to assess the discrimination power of the parameters if d ≥ 0.8 means large effect size.
| Results|| |
The statistical results of features between the healthy subject group and patient group are shown in [Table 1], [Table 2], [Table 3], including mean, SD, P values, and effect size. The results show that more than 13 features can achieve significant difference and large effect size in E group while only three features achieve the similar statistical performance in F and A groups. The features extracted from the arm sensors did not have any significant differences in F and A groups. Furthermore, we could find that only one feature (SD) extracted from the norm signals of the wrist sensor was significantly meaningful on three basic shoulder tasks.
|Table 1: Comparison of the extracted features between adhesive capsulitis patients and healthy controls in flexion task|
Click here to view
|Table 2: Comparison of the extracted features between adhesive capsulitis patients and healthy controls in extension task|
Click here to view
|Table 3: Comparison of the extracted features between adhesive capsulitis patients and healthy controls in abduction task|
Click here to view
| Discussion|| |
AC is a one of the common shoulder joint conditions that causes stiffness and pain. It also effects the range of motion of shoulder joint. AC has great impacts on the quality of daily life and ADL. The common treatments in AC patients include medicine, physical therapy, local injection, or surgical approach, which aim to relieve pain, improve joint mobility, and increase the independent ability.,, To support clinical evaluations and measurements, there is a requirement of objective and continuous assessment for helping clinical management and follow-up progresses.,
The IMU-based approaches have been widely developed for clinical evaluation in various applications.,,,, Such an IMU-based approach has the potential to provide objective, continuous, and quantitative information for clinical diagnosis and assessment. The similar IMU-based measurement systems have been successfully applied to the ROM measurement., Furthermore, previous works have applied several movement features to evaluate the performance of shoulder tasks in AC patients., They have shown the feasibility of the IMU-based approach to AC performance evaluation. However, the limited features are discussed in these studies. In this work, we first proposed an IMU-based approach to evaluate the performance of the shoulder function using statistical features. Two sensor units placed on the wrist and arm are utilized to measure the movement performance. Several typical statistical features and parameters are calculated to distinguish the differences between healthy subjects and AC patients.
The results showed that some features achieve the significant differences between healthy subjects and AC patients. The results also presented that SD could reveal the meaningful differences on three basic shoulder tasks. This is because the performance of the tasks from the most AC was not smooth. They were limited to some specific joint angles or movement while testing. However, healthy subjects perform them at high and low speed according to their individual habits. Furthermore, we found that the healthy subjects could perform basic shoulder in fast speed, while the AC patients presented them in slower speed. Obviously, most AC patients prefer to perform these tasks carefully due to pain or limited ROM.
The results also presented that more than eight parameters could achieve significant differences in the flexion task, while less than three features had significant meaning in the extension and abduction tasks. This is because most AC patients have more severe difficulties in the flexion task, while previous works have shown that AC patients had limitations in the ADL related to flexion task. Furthermore, it requires to investigate other suitable features and parameters to identify the differences between AC patients and healthy subjects in the extension and abduction tasks.
At this moment, we analyzed the performance of the shoulder while the pain is not explored in this work. The relation and analysis between these parameters and pain index should be implemented in the future work. Further, the more elderly subjects and AC patients are going to be recruited.
| Conclusion|| |
In the work, we demonstrated the feasibility of an IMU-based approach to extract statistical features for AC function assessment. The experimental results showed that this approach can provide objective functional information for clinical evaluation. The proposed approach has the potential to support clinical professionals for treatment strategies and follow-up programs.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Neviaser AS, Neviaser RJ. Adhesive capsulitis of the shoulder. J Am Acad Orthop Surg 2011;19:536-42.
Bunker TD, Frcs PJ, Schranz PJ. Clinical Challenges in Orthopaedics: The Shoulder. Taylor & Francis; 1998.
Le HV, Lee SJ, Nazarian A, Rodriguez EK. Adhesive capsulitis of the shoulder: Review of pathophysiology and current clinical treatments. Shoulder Elbow 2017;9:75-84.
Dias R, Cutts S, Massoud S. Frozen shoulder. BMJ 2005;331:1453-6.
Shaffer B, Tibone JE, Kerlan RK. Frozen shoulder. A long-term follow-up. JBJS 1992;74:738-46.
D'Orsi GM, Via AG, Frizziero A, Oliva F. Treatment of adhesive capsulitis: A review. Muscles Ligaments Tendons J 2012;2:70-8.
Yip M, Francis AM, Roberts T, Rokito A, Zuckerman JD, Virk MS. Treatment of adhesive capsulitis of the shoulder: A critical analysis review. JBJS Rev 2018;6:e5.
de los Reyes-Guzmán A, Dimbwadyo-Terrer I, Trincado-Alonso F, Monasterio-Huelin F, Torricelli D, Gil-Agudo A. Quantitative assessment based on kinematic measures of functional impairments during upper extremity movements: A review. Clin Biomech (Bristol, Avon) 2014;29:719-27.
Chang CY, Hsieh CY, Huang HY, Wu YT, Chen LC, Chan CT, et al
. Automatic functional shoulder task identification and sub-task segmentation using wearable inertial measurement units for frozen shoulder assessment. Sensors (Basel) 2020;21:106.
De Baets L, Vanbrabant S, Dierickx C, van der Straaten R, Timmermans A. Assessment of scapulothoracic, glenohumeral, and elbow motion in adhesive capsulitis by means of inertial sensor technology: A within-session, intra-operator and inter-operator reliability and agreement study. Sensors (Basel) 2020;20:876.
Coley B, Jolles BM, Farron A, Pichonnaz C, Bassin JP, Aminian K. Estimating dominant upper-limb segments during daily activity. Gait Posture 2008;27:368-75.
Luinge HJ, Veltink PH, Baten CT. Ambulatory measurement of arm orientation. J Biomech 2007;40:78-85.
Rundquist PJ, Anderson DD, Guanche CA, Ludewig PM. Shoulder kinematics in subjects with frozen shoulder. Arch Phys Med Rehabil 2003;84:1473-9.
Huber ME, Seitz AL, Leeser M, Sternad D. Validity and reliability of Kinect skeleton for measuring shoulder joint angles: A feasibility study. Physiotherapy 2015;101:389-93.
Coley B, Jolles BM, Farron A, Bourgeois A, Nussbaumer F, Pichonnaz C, et al
. Outcome evaluation in shoulder surgery using 3D kinematics sensors. Gait Posture 2007;25:523-32.
Pichonnaz C, Aminian K, Ancey C, Jaccard H, Lécureux E, Duc C, et al
. Heightened clinical utility of smartphone versus body-worn inertial system for shoulder function B-B score. PLoS One 2017;12:e0174365.
Körver RJ, Heyligers IC, Samijo SK, Grimm B. Inertia based functional scoring of the shoulder in clinical practice. Physiol Meas 2014;35:167-76.
Bavan L, Wood J, Surmacz K, Beard D, Rees J. Instrumented assessment of shoulder function: A study of inertial sensor based methods. Clin Biomech 2020;72:164-71.
Palmerini L, Mellone S, Avanzolini G, Valzania F, Chiari L. Quantification of motor impairment in Parkinson's disease using an instrumented timed up and go test. IEEE Trans Neural Syst Rehabil Eng 2013;21:664-73.
Cohen J. Statistical Power Analysis for the Behavioral Sciences. Academic Press; 2013.
Sprint G, Cook DJ, Weeks DL. Toward automating clinical assessments: A survey of the timed up and go. IEEE Rev Biomed Eng 2015;8:64-77.
Greene BR, O'Donovan A, Romero-Ortuno R, Cogan L, Scanaill CN, Kenny RA. Quantitative falls risk assessment using the timed up and go test. IEEE Trans Biomed Eng 2010;57:2918-26.
Greene BR, Kenny RA. Assessment of cognitive decline through quantitative analysis of the timed up and go test. IEEE Trans Biomed Eng 2012;59:988-95.
Biswas D, Cranny A, Gupta N, Maharatna K, Achner J, Klemke J, et al
. Recognizing upper limb movements with wrist worn inertial sensors using k-means clustering classification. Hum Mov Sci 2015;40:59-76.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]