Relationship between the kinematics of wrist/finger joint movements and muscle loading during typing in people with different skill levels ========================================================================================================================================== * Takanori Ito * Akira Kobayashi * Yuka Matsumoto * Hayase Funakoshi * Mio Yoshida * Naohiko Kanemura * Takanori Kokubun ## Abstract **Background** Musculoskeletal disorders caused by computer use are associated with socio-economic problems. Reducing the biomechanical loading caused by hand movements could reduce the occurrence of such disorders. However, most participants recruited for past experimental studies had a certain level of typing skill, and differences in the biomechanical loading of those with different typing skills have not been examined. This study aimed to quantify the relationship between the movements of the wrist and finger joints and the loading on the forearm muscles and to clarify the differences between typists of different skill levels. **Methods** A 3D motion capture system measured wrist and index finger joint movements, and surface electromyography measured muscle activities for the right hand during a typing task. We compared wrist and finger joint movements and forearm muscle loading during typing, keystroke time, and key release time between skilled and unskilled typists. **Findings** Skilled typists move their wrists and fingers faster with less muscle activity during typing, the active tension of their finger flexor muscles during keystrokes are high, and they have higher mechanical stresses on the finger flexor tendons during keystrokes. Unskilled typists develop strategies to prevent excessive stiffness in their wrist joints when making keystrokes. They are expected to have a higher cumulative loading on the extensor muscles of the wrist during key release. **Interpretation** Biomechanical loading in typing skill is different, which may make it possible to predict disability. This could provide information on the changes in physical parameters and environments to prevent disability. Keywords * Typing * Upper extremity musculoskeletal disorders (UE-MSDs) * Motion capture system * Electromyography * wrist/finger joint movements * Muscle loading ## 1. Introduction Computers are essential tools in modern society and are frequently used for long periods of time. The intensive use of computer keyboards is associated with increased discomfort in the wrist and fingers (So et al., 2017), reducing worker productivity. Additionally, computer use for more than 4 hours per day or 20 hours per week increases the risk of upper extremity musculoskeletal disorders (UE-MSDs) of the wrist and fingers, such as tendonitis and carpal tunnel syndrome (Bergqvist et al. 1995; Gerr et al. 2002; Jenssen et al. 2003). People are expected to spend more time working at computers as digital transformation continues, possibly increasing the incidence of UE-MSDs. Most studies that have investigated the kinematics of keyboard typing have focused on wrist movements. Few studies (Sommerich et al. 1996; Baker et al. 2007) have focused on finger movements during typing, but there is no research on the relationship between finger movements and the muscle activities that facilitate these movements. The motion of the sagittal plane of the finger/wrist joints is interrelated as described by the tenodesis effect (Zhao et al. 2002; Su et al. 2005). Evaluating only the finger/wrist joints and muscle activities may lead to misinterpretation of mechanical loading on the body. Some UE-MSDs are thought to be related to high muscle loading on a single keystroke, which is repeated over a long period of time. However, the relationship between these characteristic movements and individual typists remains unclear: each typist has a unique pattern and style of typing. Since typing is a complex task controlled by many joints and muscles, the movement style may be affected by skill level; different styles of movement may then carry different risks of disability. This study aimed to quantify the relationship between the movements of the wrist and finger joints and the loading of the forearm muscles and to clarify the differences according to typing skill. This study hypothesized those skilled typists exhibit increased immobilization of the wrist joints and high-speed wrist and finger joint movements. This would increase the muscle loading on the forearm muscles, resulting in different levels of risk for developing musculoskeletal disorders of the hand. By clarifying these relationships, we believe that we can deepen our understanding of the influence of subject characteristics on biomechanical loading during typing and provide basic knowledge for physical and environmental improvement approaches aimed at maintaining healthy work environments and preventing the occurrence of UE-MSDs. ## 2. Methods ### 2.1. Subjects Twelve healthy volunteers (nine men and three women) participated in this experiment. All participants provided written informed consent, following a detailed explanation of the study’s purpose and risks according to the Declaration of Helsinki. This study was approved by the ethics review committee at Saitama Prefectural University, Saitama, Japan (approval number: 20,508). ### 2.2. Equipment We used a 3D motion capture system (Vicon Motion Systems Limited, Vicon, Oxford Metrics plc, London, UK; sampling frequency: 100 Hz) and wireless multipoint surface electromyograph (EMG) (DELSYS, Delsys Trigno Wireless System, Massachusetts, USA; sampling frequency: 1,000 Hz). The height of the table was 70 cm, and a monitor (Apple, MacBook Air) and QWERTY keyboard (Microsoft, Wired Keyboard 600 ANB-00040, Washington, USA) were placed on it. The chair was adjusted for each subject to their preferred configuration prior to beginning the tests. ### 2.3. Protocol First, the height and weight of the subjects and the physical information necessary for modeling were measured. Forty four infrared-reflecting markers (diameters 4 and 14 mm) were attached to points on the participants (Fig.1)— according to the Vicon plug-in-gait full-body (PIG) and wrist/finger (Baker et al. 2007; Cook et al. 2007) models—to calculate the sagittal plane of the wrist and the metacarpophalangeal (MCP) joint of the index finger angle and velocity. Surface EMGs were attached to the right side of the flexor carpi radialis (FCR), flexor carpi ulnaris (FCU), extensor carpi radialis (ECR), and extensor carpi ulnaris (ECU) muscles. Then, the maximum voluntary isometric contraction (MVC) of each muscle was measured. Each participant practiced typing for 5 min to get accustomed to the test environment. During this practice, the height of the chair and the angle of the monitor were adjusted according to each participant’s preference. As a test (pretest) for typing skills, we used an electronic keyboarding program (e-typing, e-typing Co., Ltd., Aichi, Japan; [http://www.e-typing.ne.jp](http://www.e-typing.ne.jp)) which has a text display that advances as the user types. All subjects used the same sentence. For the typing test, all subjects were asked to type the exact same sentences, which consisted of approximately 5,000 Japanese characters. No instructions were given regarding typing speed, rhythm, or time limit; all subjects typed according to their usual natural style. After confirming that the subject’s hands and wrists were placed on the keyboard, we instructed them to begin the task, whereupon they started typing the sentences displayed on the monitor. Each participant performed the task three times, resting as necessary. ### 2.4. Data processing We analyzed all data using Nexus 2.9 (Vicon Motion Systems Limited, Vicon, Oxford Metrics plc, London, UK) and Python 3.7 (Python Software Foundation, Beaverton, Orlando, USA). The wrist joint angle was calculated from the markers attached to the index finger metacarpal head, radial styloid process, ulnar styloid process, and forearm in the PIG model. For the wrist joint angle in this study, the positive value is expressed as the dorsal flexion, and the negative value as the palmar flexion angle. The MCP joint was used to define the hand as one rigid body, and the angle of the proximal phalanx segment with respect to the hand on the sagittal plane was calculated and used as the flexion/extension angle (Fig. 1). For the MCP angle in this study, the positive value is expressed as the flexion angle and the negative value as the extension angle. ![Fig.1](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2021/06/12/2021.06.09.21258367/F1.medium.gif) [Fig.1](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/F1) Fig.1 Explanation of experimental equipment. (A) The positions of the markers and EMGs B) Calculation of MCP joint angle The kinematics data were filtered with a fourth-order Butterworth low-pass filter with a cut-off frequency of 6 Hz. After filtering was completed, the mean joint angle, joint angle change as the root-mean-square (RMS), and joint velocity were calculated in each trial of the wrist and MCP joint. The EMG data were passed through a fourth-order Butterworth bandpass filter with a cutoff frequency of 20–480 Hz, full-wave rectified, and then smoothed at 10 Hz. The EMG data during the typing sessions were normalized to the percentage of MVC. ### 2.5. Data analyses #### 2.5.1. Grouping by the pretest results The accuracy (%) and typing speed (words per minute, WPM) in Japanese for all subjects in the pretest were obtained. The accuracy is the percentage of keys that can be input without mistakes. The reason for calculating this is to make sure that we are not doing the task unsystematically in an attempt to accomplish it quickly. The typing speed was calculated for all participants, and those whose speed was higher than the average value were classified as skilled (in the skilled typist [ST] group), and those whose speed was lower than the average value were classified as unskilled (the unskilled typist [UT] group). In a study that investigated the keystroke characteristics of typing novices and experts (Takaoka et al., 2014), both were classified as having a correct response rate of 95% or higher, or close to it, as a condition for employment. #### 2.5.2. Analysis interval For all data, the start of the analysis interval was defined as the time when the first Enter key was pressed down after the examiner gave the subject the signal to start. The end of the analysis interval was defined as the time when the Enter key presented at the end of the task text was pressed down. The keystroke for the Enter key on the keyboard was determined using the 3D motion capture system: the change in position information of the Enter key when the position of the 4 mm marker attached to it was lowered was measured. Using this input, the timing of the input signal of the key and the start of measurement by the 3D motion capture system and the EMG were matched, and all kinematic data and EMG data were synchronized in time. #### 2.5.3. Analysis of wrist joint and metacarpophalangeal joint motion The initial posture of the wrist joint at the beginning of typing was investigated by calculating the mean and standard deviation (SD) values of the initial dorsiflexion angle of the wrist joint at the start time. The mean and SD values of the palmar flexion/dorsiflexion angles of the wrist and MCP joint angles of the index finger at the analysis interval calculated to examine the average posture of the wrist and the MCP joint during typing. Similarly, the mean and SD values of the RMS of the palmar flexion/dorsiflexion angles of the wrist and index finger in the analysis section were calculated. The angle velocity of the wrist and index finger MCP joints was investigated by calculating the mean and SD values of the angular change (angular velocity) per unit time. #### 2.5.4. Analysis of muscle loadings in the forearm To investigate the muscle loading in each muscle of the forearm, we used the amplitude probability distribution function analysis for the time series data of %MVC. Amplitude probability distribution function analysis is a method that expresses the percentage of the total time that the output was below a certain level as the probability (*P*) of occurrence for that output during the entire time that the EMG was measured. After plotting the function of power output and appearance probability of each muscle in each trial for each subject, we calculated the mean and SD values of the muscle loadings at *P* = 0.1, 0.5, and 0.9. #### 2.5.5. Identification of press and release times Based on the subject’s key input signals, we obtained the time when the input of each key started and ended. The time at which the “U” key was pressed and input started was defined as the press time, and the time when the fingertip left the “U” key and no input was made, as the release time. The mean and SD values of each data point were calculated and compared between the two groups. ### 2.6. Statistics After confirming the normality of all kinematics and muscle activity data using the Shapiro-Wilk test, we investigated the homoscedasticity of the data using the F-test. Since there was no difference in kinematics data and EMG data in each trial (confirmed by one-way ANOVA; *P* < 0.05), the average values of the three trials were used as representative values. The data of the two groups were compared using the unpaired t-test (*P* < 0.05) or Wilcoxon signed-rank test. ## 3. Results ### 3.1. Subjects and environments The subjects of this study were 12 healthy adults, nine men and three women. The mean (SD) value of correct answers in the typing test showed in Table.1. The mean value of WPM of all subjects was used as the benchmark for skill: subjects who achieved a score higher than the mean were classified into the ST group, and those who achieved a lower score than the mean into the UT group. View this table: [Table.1](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/T1) Table.1 Subjects and Environments The mean value of WPM of all subjects was used as the benchmark for skill: subjects who achieved a score higher than the mean were classified into the ST group, and those who achieved a lower score than the mean into the UT group. The mean (SD) age, height, weight, hand width, finger length and the male-to-female ratio showed in Table.1. All subjects were right handed. The mean (SD) chair height and screen angle showed in Table.1. The ratio of PC used on a daily basis (desktop PC to notebook PC) was 1:11. The mean (SD) values for frequency and duration of computer use were 5.2 (1.2) times per week and 3.3 (1.8) hours per day, respectively. There were no significant differences between the two groups in any of the data. ### 3.2. Wrist and metacarpophalangeal joint motion There was no significant difference in the mean dorsiflexion angle of the wrist joint at the beginning of the typing motion between the two groups. The mean value (SD) of the wrist joint dorsiflexion angle, RMS, and angular velocity showed in Fig.2. The angular velocity of flexion/extension was predominantly higher in the ST group (*P* = 0.003). ![Fig.2](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2021/06/12/2021.06.09.21258367/F2.medium.gif) [Fig.2](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/F2) Fig.2 Wrist and metacarpophalangeal joint motion The error bar represents the standard deviation between subjects. * P < 0.05 The mean (SD) values of the index finger MCP joint of flexion angle, RMS, and angular velocity showed in Fig.2. The mean angular velocity of the MCP in the ST group was significantly higher than that in the UT group (*P* = 0.002). ### 3.3. Muscle loading The mean values (SD) of %MVC of FCR, FCU, ECR, and ECU at the static activity level (*P* = 0.1), the mean activity level (*P* = 0.5), the maximum activity level (*P* = 0.9) showed in Fig.3. The mean of %MVC of FCR at *P* = 0.9 was significantly higher than that in the ST group (*P* = 0.023). The mean of %MVC of ECR at *P* = 0.1, 0.5, 0.9 were significantly higher than that in the UT group (*P* = 0.040, 0.020, 0.020). ![Fig.3](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2021/06/12/2021.06.09.21258367/F3.medium.gif) [Fig.3](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/F3) Fig.3 Muscle loading From left to right, the static activity level (P = 0.1), the mean activity level (P = 0.5), the maximum activity level (P = 0.9) are listed. The error bar represents the standard deviation between subjects. * P < 0.05 ### 3.4. Press and release time data The mean (SD) of the release time was 0.12 (0.02) and 0.20 (0.14) s for the ST group, 0.13 (0.01) and 0.11 (0.01) s for the UT group, and 0.11 (0.02) and 0.26 (0.15) s for the ST group. The comparison showed no significant difference between the two groups. The mean values (SD) of the wrist joint dorsiflexion angle, angular velocity, and the index finger MCP joint flexion angle, angular velocity during the press time and release time showed in Fig.4, Fig.5. The mean value of the wrist joint dorsiflexion angular velocity during the release time in the ST group was significantly higher than that in the UT group (*P* = 0.005). ![Fig.4](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2021/06/12/2021.06.09.21258367/F4.medium.gif) [Fig.4](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/F4) Fig.4 Press time data From left to right, the wrist joint dorsiflexion angle, angular velocity, the index finger MCP joint flexion angle, angular velocity, %MVC of FCR, FCU, ECR, and ECU are listed. The error bar represents the standard deviation between subjects. * P < 0.05 ![Fig.5](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2021/06/12/2021.06.09.21258367/F5.medium.gif) [Fig.5](http://medrxiv.org/content/early/2021/06/12/2021.06.09.21258367/F5) Fig.5 Release time data From left to right, the wrist joint dorsiflexion angle, angular velocity, the index finger MCP joint flexion angle, angular velocity, %MVC of FCR, FCU, ECR, and ECU are listed. The error bar represents the standard deviation between subjects. * P < 0.05 The mean value of the index finger MCP joint flexion/extension angular velocity during the press time and release time in the ST group was significantly higher than that in the UT group (*P* = 0.002, 0.006). The mean values (SD) of %MVC of FCR, FCU, ECR, and ECU during press time and release time showed in Fig.4, Fig.5. The mean values (SD) of %MVC of ECR during press time of the ST group was significantly lower than that of the UT group (*P* = 0.020). ## 4. Discussion In this study, we examined the typing skills, joint movements and muscle loading of 12 volunteers to evaluate the differences between skilled and unskilled typists. ### 4.1. Subjects and environment Although the ST and UT groups showed similar typing accuracy, the typing speed between the two groups was significantly different (*P* < 0.05). In both groups, the accuracy was around 95%, indicating that the subjects were able to adapt to the test environment and reproduce their natural typing motion in it. In typing experiments with English sentences (Rempel et al. 2007; Szeto et al. 2000), the average typing speed of the subjects ranged from 20 to 50 WPM. On average, one word in English consists of five keystrokes: four keys and one space key (Crump et al. 2011; Nagasawa 2017). The WPM for Japanese in this study was equivalent to approximately 34 to 44 WPM in English, indicating that the typing speeds of all participants in this study were within the range of subjects in previous studies. Therefore, we believe that the validity of the design for assessing typing skills is ensured. ### 4.2. Kinematics of finger and wrist joint There were no significant differences in the dorsiflexion angle and RMS of the wrist between the ST and UT groups, but the angular velocity of flexion/extension was predominantly higher in the ST group. The kinematic parameters of the MCP joint of the index finger showed no significant differences in the joint angle and RMS, but the angular velocity of flexion/extension was predominantly higher in the ST group. The higher typing speed of the ST group was due to fast movements of the wrist joint and fingers, as well as coordinated fast joint action in the joints. Previous studies have shown that the dorsiflexion angle of the wrist joint changes depending on the gable angle, split angle, and tilt angle of the keyboard (Simoneau 2003; Rempel 2007). Since all participants in the present study performed the task with the same keyboard at the same tilt angle, it seems that typists adjust their posture and range of wrist motion to the characteristics of the keyboard regardless of the level of skill. ### 4.3. Muscle loading and movement The two groups showed similar muscle loading of the wrist flexors (FCR and FCU) during the task and muscle activity during press and release times. It has been reported that the flexion movement of the MCP joint during the typing motion is synergistic with the flexion movement of the wrist joint (Dennerlein 2007). In the present study, the ST group performed faster finger and wrist joint movements: the passive tension of the wrist flexors due to the dorsiflexion motion of the wrist acted on the flexion motion of the MCP and wrist joints, enabling efficient and energetically economical movement. However, the ST group may have used excessive muscle activity of the wrist flexors to accelerate the joint movement of the index finger during keystrokes. Additionally, because the fingertips were in contact with the keys at a high speed during key pressing, the finger flexors may have acted excessively to maintain the posture of the hand against the reaction force from the keys. The muscle activity of the wrist extensor muscle decreases as the dorsiflexion angle of the wrist decreases during typing (Simoneau 2003), possibly due to changes in muscle length and moment arm. In this study, there was no difference in the angle of the wrist joint and the degree of change of angle between the two groups, but the muscle loading of the static activity level of the ECR—an extensor muscle of the wrist joint—was significantly higher in the UT group than in the ST group. This may be due to excessive activity of the ECR in the UT group to stiffen the wrist joint, even when the moment arm of the muscle was long. This excessive contraction may reduce the sliding ability of the ECR tendon in its sheath. Additionally, the ECU muscle acts during lifting of the hand after a key press (Dennerlein 2007). In the UT group, when fingertips were released from the keystroke, excessive lifting of the hand was observed during the movement of the wrist joint, possibly increasing the muscle loading of the ECU muscle during key release. ### 4.4. Limitations The measurement conditions in this study were not restricted, as the posture of the participants was not constrained during the test. Participants could place their forearms on the desk or keep them floating. Since placing the forearm on the desk may decrease the muscle loading of the extensor muscles of the wrist joint, the distance between the upper body and the keyboard and the presence or absence of space on the desk in front of the keyboard must be checked at the beginning of the measurement, and the effects of differences in the contact time and contact area between the desk and the forearm on the body during the task should be determined. The majority of the participants recruited for this study were male. Physical parameters, such as hand size, range of motion of joints, and absolute muscular strength, vary depending on sex, and this may have affected the physicality of the typing test. In this study, we examined the differences in physical parameters with regard to skill level among healthy adult students. We used a younger age grouping, assuming people of this age would form the core labor force. However, the subjects in this study were not experienced at desk work, and we were not able to verify the results for older typists with greater experience. Therefore, the findings of this study should be generalized with caution because the characteristics of the study population may differ from those who are at risk of developing musculoskeletal disorders. In this study, the press time of the “U” keystroke and the time from the release of the “U” key to the start of the next keystroke were used as the analysis intervals, and the movements of the index finger were analyzed in detail. However, we were not able to confirm for all participants which finger was actually used to input the “U” key. Thus, we may not have been able to evaluate the movement of only the index finger. Additionally, because we did not measure the fingertip force during keystrokes, we could not show a direct relationship between the angular velocity of the MCP joint and the actual force generated during keystrokes. For future investigations, we intend evaluating the movements of fingers other than the index finger and examining the relationships between changes in posture, muscle loading, and fingertip force during specific keystrokes. ## 5. Conclusions The posture of the wrist and MCP joints of the index finger during the typing motion was similar, regardless of skill level. Skilled typists type faster with less muscle activity because they can utilize the passive tension of the muscle tendons for keystrokes. However, to achieve this high keystroke speed, they exhibit high active tension of the finger flexors. The active tension of the finger flexors may also be higher in response to the reaction force received from the keys as the fingertips press the keys, possibly increasing the mechanical stress on the finger flexor tendons and becoming a risk factor for the occurrence of disorders. Unskilled typists require more time to switch between flexion and extension of the hand because of the high load on the extensor muscles caused by the excessive immobilization of the wrist joint and their strategy to prevent keystrokes. This style of movement has a high risk for developing extensor muscle-tendon disorders of the wrist. ## Data Availability The data are not publicly available due to their containing information that could compromise the privacy of research participants ## Sources of funding There is no applicable funding. ## Acknowledgements The authors would like to acknowledge the members of the SPU Fundamental Research group for their assistance. ## Footnotes * **Author’s E-mail** Takanori Ito; 2391002b{at}spu.ac.jp, Akira Kobayashi; 1991003n{at}spu.ac.jp, Yuka Matsumoto; 2191005s{at}spu.ac.jp, Hayase Funakoshi; 2381302r{at}spu.ac.jp, Mio Yoshida; 2381309c{at}spu.ac.jp, Naohiko Kanemura; kanamura-naohiko{at}spu.ac.jp, Kokubun Takanori; kokubun-takanori{at}spu.ac.jp * Received June 9, 2021. * Revision received June 9, 2021. * Accepted June 12, 2021. * © 2021, Posted by Cold Spring Harbor Laboratory The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission. ## References 1. 1. Billy CL So, Andy SK Cheng, Grace PY Szeto. Cumulative IT Use Is Associated with Psychosocial Stress Factors and Musculoskeletal Symptoms. Int. J. Environ. Res. Public Health 2017; 14: 1541; doi: 10.3390/ijerph14z121541. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3390/ijerph14z121541&link_type=DOI) 2. 2. U Bergqvist, E Wolgast, B Nilsson, M Voss. Musculoskeletal disorders among visual display terminal workers: individual, ergonomic, and work organizational factors. Ergonomics 1995; 38(4): 763–776; doi: 10.1080/00140139508925148. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1080/00140139508925148&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=7729403&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1995QN20100014&link_type=ISI) 3. 3. F Gerr, M Marcus, C Ensor, D Kleinbaum, S Cohen, A Edwards, E Gentry, DJ Ortizl, C Monteilh. A prospective study of computer users: I. Study design and incidence of musculoskeletal symptoms and disorders. Am J Ind Med. 2002; 41(4): 221–235; doi:10.1002/ajim.10066. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/ajim.10066&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=11920966&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000174596100001&link_type=ISI) 4. 4. C Jensen. Development of neck and hand-wrist symptoms in relation to duration of computer use at work. Scand J Work Environ Health. 2003; 29(3): 197–205; doi: 10.5271/sjweh.722. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.5271/sjweh.722&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=12828389&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000183452800004&link_type=ISI) 5. 5. CM Sommerich, WS Marras, M Parnianpour. A quantitative description of typing biomechanics. J Occup Rehabil. 1996; 6(1): 33–56; doi: 10.1007/BF02110393. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/BF02110393&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24234809&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 6. 6. NA Baker, R Cham, EH Cidboye, J Cook, MS Redfern. Kinematics of the fingers and hands during computer keyboard use. Clin Biomech. 2007; 22(1): 34–43; doi: 10.1016/j.clinbiomech.2006.08.008. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.clinbiomech.2006.08.008&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17052825&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 7. 7. C Zhao, PC Amadio, ME Zobitz, T Momose, P Couvreur, KN An. Effect of synergistic motion on flexor digitorum profundus tendon excursion. Clinical Orthop Relat Res. 2002; (396): 223–230; doi: 10.1097/00003086-200203000-00033. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/00003086-200203000-00033&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=11859247&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 8. 8. F-C Su, YL Chou, CS Yang, GT Lin, KN An. Movement of finger joints induced by synergistic wrist motion. Clin Biomech. 2005; 20(5): 491–497; doi: 10.1016/j.clinbiomech.2005.01.002. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.clinbiomech.2005.01.002&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15836936&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 9. 9. JR Cook, NA Baker, R Cham, E Hale, MS Redfern. Measurements of wrist and finger postures: A comparison of goniometric and motion capture techniques. J Appl Biomech. 2007; 23(1): 70–78; doi: 10.1123/jab.23.1.70. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1123/jab.23.1.70&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17585179&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 10. 10. E Takaoka, M Sugiura, H Komiya. A study on typist characterization. IPSJ SIG Technical Report. 2014; 125(9): 1–14. 11. 11. D Rempel, A Barr, D Brafman, Ed Young. The effect of six keyboard designs on wrist and forearm postures. Appl Ergon. 2007; 38(3): 293–298; doi: 10.1016/j.apergo.2006.05.001. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.apergo.2006.05.001&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16806042&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom) 12. 12. GPY Szeto, J Kim FN. A comparison of wrist posture and forearm muscle activities while using an alternative keyboard and a standard keyboard. J Occup Rehabili. 2000; 10(3): 189–197; doi: 10.1023/A:1026662302131. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1023/A:1026662302131&link_type=DOI) 13. 13.Gordon D, Logan MJC Crump, Hierarchical Control of Cognitive Processes: The Case for Skilled Typewriting. Psychology of Learning and Motivation. 2011; (54): 1–27; doi: 10.1016/B978-0-12-385527-5.00001-2. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/B978-0-12-385527-5.00001-2&link_type=DOI) 14. 14. N Nagasawa. How Japanese university students type on smartphone and PC. Computer and Education. 2017; (43): 67–72; doi: 10.14949/konpyutariyoukyouiku.43.67. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.14949/konpyutariyoukyouiku.43.67&link_type=DOI) 15. 15. GG Simoneau, RW Marklin, JE Berman. Effect of Computer Keyboard Slope on Wrist Position and Forearm Electromyography of Typists Without Musculoskeletal Disorders. Phys Ther. 2003; 83(9): 816–830; doi: 10.1093/ptj/83.9.816. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6OToicHRqb3VybmFsIjtzOjU6InJlc2lkIjtzOjg6IjgzLzkvODE2IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjEvMDYvMTIvMjAyMS4wNi4wOS4yMTI1ODM2Ny5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 16. 16. JT Dennerlein, I Kingma, B Visser, J.H van Dieën. The contribution of the wrist, elbow and shoulder joints to single-finger tapping. J Biomech. 2007; 40(13): 3013–3022; doi: 10.1016/j.jbiomech.2007.01.025. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.jbiomech.2007.01.025&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17467717&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F06%2F12%2F2021.06.09.21258367.atom)