Comparison of muscle activity from upper trapezius and wrist extensors between dominant and non-dominant upper limbs during computer-based tasks

Abstract

BACKGROUND:

Sustained low-level muscle activity occurring during computer-based tasks is associated with the development of WMSDs (work-related musculoskeletal disorders) and this biomechanical exposure may be different between limbs.

OBJECTIVE:

To compare muscle activity from dominant and non-dominant upper trapezius (UT) and wrist extensors (WE) during computer-based tasks in real work settings.

METHODS:

Forty-five workers were monitored during two hours while performing their usual administrative tasks. Surface electromyography (sEMG) was recorded from UT and WE muscles in both sides. Rest and general exposure variables were calculated.

RESULTS:

The 50th percentile demonstrated little muscle activity demand, for both dominant and non-dominant UT and no difference between sides was observed. The dominant WE muscles had lower measures of rest and higher muscle activity when compared with the non-dominant side.

CONCLUSIONS:

Differences in sEMG between upper limbs were only found in WE muscles, probably due to the use of the mouse. The overall low-level muscle activity suggests a constant activation of the same motor units for the entire data-collection period, which can be considered harmful for musculoskeletal health.

Keywords

Office work occupational health EMG muscular rest gap frequency

1 Introduction

Computer-based tasks have increased substantially in the last decades [1, 2]. Work-related musculoskeletal disorders (WMSDs) are an important public health issue in industrialized countries, especially when considering the neck-shoulder and wrist regions [3 –5]. Kiss et al. (2012) assessed 7,017 computer workers and observed that more than 51.3% of them reported musculoskeletal symptoms in the neck-shoulder region in the last year. Moreover, the hand-wrist was identified as the region with the second highest prevalence of complaints (12.8%) [6].

The high prevalence of WMSDs in those regions is associated with the static posture adopted during computer-based tasks, which generates sustained low-level muscle activity [7, 8], imposing biomechanical strain on ligaments, joint capsules, muscle fibers, and other musculoskeletal structures [9]. It is also associated with sympathetic vasoconstriction, which leads to decreased muscle oxygenation, causing the muscle metabolism to tend towards an anaerobic state [10, 11]. This muscle impairment may be explained by the Cinderella hypothesis, in which the muscle fibers remain at a low-level muscle activity for a long period in static positions, without rest, leading to a selective overuse of low-threshold muscle fibers. When these conditions occur frequently during computer-based tasks, it may trigger the development of a WMSD [1].

Literature supports that some short interruptions in the muscle activity, also known as gaps, are beneficial to musculoskeletal health since they promote variations in the muscle pattern. When these gaps are rare during the performance of a task, it is considered a predictive factor for the development of WMSDs [12, 13]. Nevertheless, it is important to recognize that the workload is not symmetrically distributed between the upper limbs in many types of work, including computer-based work. Therefore, to better understand the mechanisms related to WMSD development, aiming at reducing its occurrence and preventing new episodes, it is important to study the muscle activity patterns during the performance of computer-based tasks with a particular eye on the differences between sides.

Marker and co-workers [14] evaluated pain free office-workers and found that the dominant upper limb had higher amplitudes of muscle activity and lower measures of muscle rest when compared to the non-dominant side. On the other hand, Nordander et al. and Yoshizaki et al. [15, 16] did not find significant differences between the dominant and non-dominant upper limbs. The authors believe that differences between upper limbs would appear in more complex tasks or if additional external load was imposed on the neck-shoulder muscles.

Thus, the purpose of this study was to quantify muscle activity of upper trapezius (UT) and wrist extensor (WE) muscles during the performance of computer-based tasks and then compare the dominant and non-dominant upper limbs. We hypothesized that the dominant UT and WE muscles would have significantly higher amplitude values and lower measures of rest when compared to their non-dominant pair, once the combined use of keyboard and mouse could require higher muscle activation when compared to the upper limb that exclusively uses the keyboard.

2 Methods

2.1 Participants

Fifty computer-based office workers (11 males and 39 females) from a public university were randomly selected to compose this sample. They performed administrative tasks related to coordination of undergraduate or post-graduate courses and management of areas such as finances, human resources and transportation. All participants had their workstations evaluated and adjusted according to the guidelines based on the model elaborated by the Finnish Institute of Occupational Health [17]. Those adjustments were made in consideration of the fact that inadequate workstation conditions are associated with awkward upper limb postures, which may be related to the development and worsening of WMSD [18].

The study included subjects who: (a) worked for at least 5 years using the computer as a tool to perform administrative tasks, (b) worked at least 4 hours per day with the computer, (c) had less than a month out of work, except for vacations, during the previous year, (d) agreed to participate in all stages of this study. We excluded all subjects who reported any rheumatic, circulatory or inflammatory systemic disease. All participants used the dominant hand to operate a traditional mouse, without any ergonomic adaptations. All participants signed an informed consent form that was approved by the local ethics committee.

2.2 Experimental procedure

The participants were analyzed during a period of two hours during a regular working day at their own workplace while they performed their usual tasks. During the analysis period, the participants performed computer-based tasks, answered phone calls, signed documents, helped students with administrative issues, walked to other departments, talked to their colleagues and took informal breaks. A researcher accompanied each participant and categorized the office work into four different tasks, registering on a tablet the type of task performed and time spent at each task within the period of data collection. However, for the purpose of this study, only the periods using the computer were considered while other activities such as answering phone calls and assisting students with administrative issues were excluded from the data analysis.

2.2.1 Electromyography

The EMG signals of the UT and WE muscles were recorded for both dominant and non-dominant sides using a portable device (Myomonitor IV, DelSys, Boston, USA). The EMG signals were acquired at 1000 Hz and conditioned by the main amplifier (Bagnoli-8 EMG System, Delsys), which provided a gain of 1000 Hz, frequency pass-band 20–450 Hz, 16-bit resolution and noise of 1.2 microvolts (RMS).

Before electrode placement, the skin was lightly rubbed with 70% ethyl alcohol and shaved to reduce the impedance and to eliminate possible interference. The UT electrodes were placed according to Mathiassen et al. [10], approximately two centimeters lateral to the midpoint between the dorsal process of the seventh cervical vertebra and the acromion with the rectangular contacts aligned in the direction of the muscle fibers. The WE electrodes were placed over the extensor carpi radialis longus and brevis, identified by palpation during forearm extension with the forearm pronated, at a distance of one-third of the forearm length from the lateral epicondyle [19]. The reference electrode (adhesive, 5 cm square) was affixed at the manubrium of the sternum.

Three maximal voluntary contractions (MVCs) were performed for each muscle. The MVCs were tested with subjects in the seated position and lasted five seconds each with an interval of one minute between them. The MVC for the UT was performed with the shoulders flexed at 90° of to the frontal plane with elbows extended while a manual resistance was applied downward at the elbow level. The MVC for the WE was performed with the forearms on a flat surface with elbow flexed at 90° while the participant was instructed to perform a wrist extension while a manual resistance was applied at the metacarpal level by the physiotherapist.

2.3 Data processing

The EMG signals were processed using Matlab (version 7.6, The Mathworks Inc., Natick, MA, USA). For data analysis, only the computer-based tasks were considered. A time concatenation of each one of the four tasks was done using the time measurements obtained from the tablet. The EMG signal was concatenated minute-by-minute based on the temporal vector of the tasks. If the last epoch spent at a specific task was less than a whole minute (time window below 60 seconds), the remaining seconds of that specific task were discarded [4 , 20].

All signals were corrected for offset and band-pass filtered using a zero-lag 6th order Butterworth filter at 20–450 Hz band. Signals were RMS-converted using consecutive 125 ms moving windows without overlap. They were then normalized as the percentage of the largest amplitude average RMS obtained during the three central seconds of the MVC’s [10].

Measurements of gap frequency (min^-1), relative rest time (RRT) and amplitude probability distribution function (APDF) were calculated as described below. Gap frequency was expressed as the number of gaps/min, where a single gap was defined as the period of muscle contraction when the muscle activation was below 1% MVC for at least 125 ms. The RRT, defined as the relative cumulative time of gaps within the task [1] was calculated as the overall duration of “rest” time below 1% MVC and expressed as a percentage of total task time (% RRT). Both gap frequency and RRT were considered measurements of muscular rest. The percentiles of the APDF were also calculated. The 10th APDF percentile (P10) was considered to be an indicator of static activity, the 50th APDF percentile (P50) to be an indicator of median muscle activity, and the 90th APDF percentile (P90) to be an indicator of peak contractions [21].

2.4 Statistics

Statistical analysis was performed using the Statistical Package for Social Sciences (IBM corporation, SPSS version 14, Armonk, NY, USA) with alpha level at p < 0.05. The normality of distribution was tested through a Shapiro Wilk test. Once all variables failed the normality assumption, the non-parametric Mann-Whitney test was used to compare the muscle activity between the dominant and non-dominant limb.

3 Results

From the 50 office workers chosen to compose this sample, one participant had the diagnosis of fibromyalgia and four withdrew themselves from the study during its course. Therefore, the final sample was composed of 45 participants (42.4±9.1 years old, 166±7 cm, 71.8±14.1 kg), 11 males and 34 females. The mean time working as a computer-based office worker at the university was 17±10 years. Among this sample, four left-handed and 41 right-handed participants were identified.

In total, the computer-based task represented a proportion of 45±16.9% of the total time within the period of data collection (two hours). The median muscle activity (P50) was around 6.0% MVC for both the dominant and non-dominant UT but no significant differences were found for those measurements (Table 1). When the WE muscles were evaluated, a significant difference was found when comparing the dominant and non-dominant limbs. The non-dominant WE muscle showed more gap frequency (min^-1) and higher % RRT values when compared with the dominant limb, as shown in Table 1. Significant differences were observed between limbs in the percentiles of static and median load in the WE muscles. The dominant WE maintained higher P10 and P50 levels during the performance of computer-based tasks, when compared with the non-dominant limb.

Table 1
Mean values (and lower-upper quartiles) of number of gaps/minute, RRT and muscle activation levels in 10th, 50th and 90th percentiles

Upper Trapezius P-value Wrist Extensors P-value

dominant non-dominant dominant non-dominant

Gap frequency (min^-1) 14.2 19.3 0.22 3.6 8.4 <0.01*

(6.7–18.7) (7.4–25.8) (0.6–4.3) (3.0–13.5)

RRT (% time) 19.9 21.4 0.54 4.8 10.3 <0.01*

(6.2–30.3) (8.7–31.9) (0.3–3.3) (1.4–16.1)

10th percentile (% MVC) 1.6 1.4 0.18 3.9 1.5 <0.01*

(0.4–1.8) (0.2–1.2) (2.3–5.8) (0.7–2.6)

50th percentile (% MVC) 6.3 5.9 0.85 10.1 6.8 <0.01*

(2.3–8.1) (1.9–8.1) (7.3–13.4) (3.5–10.4)

90th percentile (% MVC) 15.4 15.3 0.76 19.8 19.9 0.70

(8.1–20.0) (8.9–19.1) (14.6–24.6) (14.1–23.6)

	Upper Trapezius	P-value	Wrist Extensors	P-value
Gap frequency (min^-1)	14.2	19.3	0.22	3.6	8.4	<0.01*
	(6.7–18.7)	(7.4–25.8)		(0.6–4.3)	(3.0–13.5)
RRT (% time)	19.9	21.4	0.54	4.8	10.3	<0.01*
	(6.2–30.3)	(8.7–31.9)		(0.3–3.3)	(1.4–16.1)
10th percentile (% MVC)	1.6	1.4	0.18	3.9	1.5	<0.01*
	(0.4–1.8)	(0.2–1.2)		(2.3–5.8)	(0.7–2.6)
50th percentile (% MVC)	6.3	5.9	0.85	10.1	6.8	<0.01*
	(2.3–8.1)	(1.9–8.1)		(7.3–13.4)	(3.5–10.4)
90th percentile (% MVC)	15.4	15.3	0.76	19.8	19.9	0.70
	(8.1–20.0)	(8.9–19.1)		(14.6–24.6)	(14.1–23.6)

*Indicates P < 0.05 through Mann-Whitney test.

4 Discussion

In the present study, the EMG of the UT and WE from both dominant and non-dominant limbs were investigated during the performance of computer-based tasks in a real work environment. The results demonstrated that the initial hypothesis was only partially confirmed. Even though a significant difference between upper limbs was observed through the EMG of the WE muscles, no difference was found for the UT muscles. We believe that the significant difference occurred due to the use of the mouse by the dominant upper limb, requiring higher muscle activity than the non-dominant limb. The lack of differences between the UT muscle activity may be attributed to other factors that affect muscle activation, such as psychosocial factors (for instance, any stressor may affect both right and left UT’s) and the low-level muscle activity pattern that is characteristic of the computer-based tasks.

As expected, we found significant differences between dominant and non-dominant upper limbs for the forearm EMG when considering the variables of gaps, RRT, 10th and 50th percentile. This shows the influence of the computer mouse on muscle activity. The upper limb that uses the mouse had fewer periods of muscle contraction below 1% of MVC, therefore had a smaller percentage of muscle rest than the non-dominant side, as shown by the RRT (%). In a study conducted with 15 right-handed computer workers, it was observed that the dominant WE muscle showed higher muscle activity at the 10th percentile and lower percentage of muscle rest when compared with the non-dominant side, in line with our study [22].

When evaluating the UT muscle activity, no significant differences were found for the dominant and non-dominant upper limb, contrary to expectations. Marker and coworkers [14] observed significant differences between upper limbs when evaluating 77 pain-free office workers in their natural work environment. They observed higher UT muscle activity at the 10th, 50th and 90th percentile and lower percentage of muscle rest and gap frequency when compared with the UT from the non-dominant side [14].

The same tendency was observed when subjects with self-reported neck-shoulder complaints were evaluated [1]. The muscle activity at the 10th and 50th percentile was higher on the dominant UT compared to non-dominant side, but no significant difference was found regarding the percentage of muscle rest and gap frequency during the performance of standardized computer-based tasks in a controlled laboratory environment [1]. Differences in muscle activity and muscle rest between the dominant and non-dominant muscles indicate that the non-dominant UT is used less during the performance of computer-based tasks, likely decreasing global loads placed on this muscle throughout the workday [14].

Even though the same tendency was observed in this study (higher muscle activity and lower muscle rest on the dominant UT), none of the variables were significantly different when comparing dominant and non-dominant limbs. In line with our results, Nornander et al. [15] only found significant differences between the UT muscle activity from dominant and non-dominant upper limbs when considering the 10th percentile of the APDF during the evaluation of right-handed subjects. When left-handed subjects were considered, no consistent pattern of higher load on the left side was identified [15]. Also, when considering simpler tasks such as elevation and lowering of the arm, no significant differences were found between dominant and non-dominant upper limb muscle activity [16].

The lack of differences between upper limbs may indicate that the UT muscle activity suffers influence of other factors than just physical load. During the performance of a computer-based stress task performed with the dominant upper limb, it was observed the range of RMS increased in the non-dominant side of subjects with neck-shoulder disorders [23]. The authors attributed the increase on the non-dominant UT muscle activity to an overflow from afferents from ipsilateral muscles to contralateral motor neurons that may be affected in subjects with neck-shoulder pain. A review of the epidemiological literature reported a positive association between psychosocial factors, such as stress, and upper extremity symptoms or development of disorders [24]. It was also observed that imposed mental stress increases the muscle activity of the trapezius muscle, suggesting an association between work-related stress and neck-shoulder pain due to sustained low-level muscle activity [25, 26].

Another important factor to consider when analyzing the lack of difference between the dominant and non-dominant UT is that the computer-based task evaluated in this study showed very low values of muscle activity when compared with other tasks, such as cleaning and assembly tasks [15 , 28]. This low-level muscle activity pattern leads to little variation along the time exposure, which also could lead to the development of WMSDs or worsen ongoing symptoms [29, 30]. Even though a significant difference was found between the dominant and non-dominant upper limbs for the forearm muscle activity, it is important to consider that regardless of this difference, both sides are exposed to risk factors for the development of WMSD considering the sustained low-level muscle activity, regardless of the upper limbs’ dominance [7, 8].

A study performed by Szeto et al. [7] compared 43 female office workers with and without neck-shoulder pain. When a discomfort measurement was included to quantify differences within the pain group, it was observed that women reported discomfort not only on the dominant muscle, but also on the non-dominant upper trapezius, the second most discomforting region. It was also observed that the group that reported more discomfort had higher muscle activity levels than women with neck-shoulder pain at low levels of discomfort or women without symptoms [7].

Several studies proposed methods to increase the variation along the time exposure in terms of posture and muscle activity during the performance of computer-based tasks. For example, the following have been proposed: alternation between computer and non-computer tasks [31]; alternation between sitting and standing position during office work [32, 33]; the inclusion of rest breaks [34] or even active pauses [35, 36]; and, the performance of physical exercise [37]. These initiatives can be implemented in office work context but should be done carefully to improve physical health [38]. Before redesigning office work, it is necessary to understand occupational risk factors in different scenarios and the potential benefits in terms of increasing the variation along the time exposure, which can be considered a protective factor against the development of WMSDs.

5 Methodological considerations

It is important to consider the fact that the population was not homogeneous, which contributes to increasing the external validity of the study but, on the other hand, it may have hidden significant differences what would have been apparent if specific populations had been considered by themselves. Even though subjects with systemic disorders were excluded, some subjects included in the sample had self-reported complaints on the upper limbs. Additionally, both sexes were evaluated together and recent evidence suggests differences in the trapezius activation of men and women [39, 40]. It is also important to consider that the computer-based tasks were considered without discrimination between mouse, keyboard, and idle tasks, which have different demands on the musculoskeletal system when comparing dominant and non-dominant upper limbs, especially when considering the wrist extensor muscles.

6 Conclusion

This paper aimed at quantifying the muscle activity of upper trapezius and wrist extensors muscles during the performance of computer-based tasks in a real occupational environment and comparing dominant and non-dominant upper limbs. In general, the overall muscle activity was low during the computer-based tasks for all four muscles. Higher muscle activity and lower muscle rest were found in the dominant upper limb compared to the non-dominant one. This difference was only significant for the wrist extensors muscles, presumably due to the mouse use. The overall low-level muscle activity that was identified suggests a constant activation of the same motor units for the entire data-collection period, which can be considered harmful for musculoskeletal health. Methods to increase variation and prevent the development of WMSDs in upper limbs should be tested in this population.

Footnotes

Acknowledgments

The authors would like to thank the São Paulo Research Foundation for financial support and the Federal University of São Carlos for allowing the performance of this experimental protocol on a real work environment. We would also like to thank all the computer work participants for their support and kindness.

Conflict of interest

None to report.

References

Thorn

, Søgaard

, Kallenberg

LAC

, Sandsjö

, Sjøgaard

, Hermens

, et al. Trapezius muscle rest time during standardised computer work-a comparison of female computer users with and without self-reported neck/shoulder complaints. J Electromyogr Kinesiol. 2007;17:420–7. doi:10.1016/j.jelekin.2006.04.010.

IJmker

, Blatter

, van der Beek

, van Mechelen

, Bongers

. Prospective research on musculoskeletal disorders in office workers (PROMO): Study protocol. BMC Musculoskelet Disord. 2006;7:55. doi:10.1186/1471-2474-7-55.

Lassen

, Mikkelsen

, Kryger

, Andersen

. Risk factors for persistent elbow, forearm and hand pain among computer workers. Scand J Work Environ Health. 2005;31:122–31. doi:10.5271/sjweh.859.

Hansson

G-Å

, Nordander

, Asterland

, Ohlsson

, Strömberg

, Skerfving

, et al. Sensitivity of trapezius electromyography to differences between work tasks - Influence of gap definition and normalisation methods. J Electromyogr Kinesiol. 2000;10:103–15. doi:10.1016/S1050-6411(99)00030-9.

Waersted

, Hanvold

, Veiersted

. Computer work and musculoskeletal disorders of the neck and upper extremity: A systematic review. BMC Musculoskelet Disord. 2010;11:79. doi:10.1186/1471-2474-11-79.

Kiss

, De Meester

, Kruse

, Chavée

, Braeckman

. Neck and shoulder complaints in computer workers and associated easy to assess occupational factors–a large-scale cross-sectional multivariate study. Int Arch Occup Environ Health. 2012;85:197–206. doi:10.1007/s00420-011-0645-4.

Szeto

GPY

, Straker

, O’Sullivan

. A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work–Neck and shoulder muscle recruitment patterns. Man Ther. 2005;10:270–80. doi:10.1016/j.math.2005.01.004.

Kleine

, Schumann

, Bradl

, Grieshaber

, Scholle

. Surface EMG of shoulder and back muscles and posture analysis in secretaries typing at visual display units. Int Arch Occup Environ Health. 1999;72:387–94.

Arvidsson

, Hansson

G-Å

, Mathiassen

, Skerfving

. Changes in physical workload with implementation of mouse-based information technology in air traffic control. Int J Ind Ergon. 2006;36:613–22. doi:10.1016/j.ergon.2006.03.002.

10.

Mathiassen

, Winkel

, Hägg

. Normalization of surface EMG amplitude from the upper trapezius muscle in ergonomic studies - A review. J Electromyogr Kinesiol. 1995;5:197–226.

11.

Hansen

. Sympathetic neural control of skeletal muscle blood flow and oxygenation. Dan Med Bull. 2002;49:109–29.

12.

Ciccarelli

, Straker

, Mathiassen

, Pollock

. Diversity of tasks and information technologies used by office workers at and away from work. Ergonomics. 2011;54:1017–28. doi:10.1080/00140139.2011.609913.

13.

Hägg

, Aström

. Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders. Int Arch Occup Environ Health. 1997;69:423–32. doi:10.1007/s004200050170.

14.

Marker

, Balter

, Nofsinger

, Anton

, Fethke

, Maluf

. Upper trapezius muscle activity in healthy office workers: Reliability and sensitivity of occupational exposure measures to differences in sex and hand dominance. Ergonomics. 2016;139:1–10. doi:10.1080/00140139.2015.1130860.

15.

Nordander

, Hansson

G-Å

, Rylander

, Asterland

, Byström

, Ohlsson

, et al. Muscular rest and gap frequency as EMG measures of physical exposure: The impact of work tasks and individual related factors. Ergonomics. 2000;43:1904–19. doi:10.1080/00140130050174536.

16.

Yoshizaki

, Hamada

, Tamai

, Sahara

, Fujiwara

, Fujimoto

. Analysis of the scapulohumeral rhythm and electromyography of the shoulder muscles during elevation and lowering: Comparison of dominant and nondominant shoulders. J Shoulder Elb Surg. 2009;18:756–63. doi:10.1016/j.jse.2009.02.021.

17.

Ahonen

, Launis

, Kuorinka

. Ergonomic Workplace Analysis. Helsink: Finnish Institute of Occupational Health; 1989.

18.

Rodrigues

MSA

, Leite

RDV

, Lelis

, Chaves

. Differences in ergonomic and workstation factors between computer office workers with and without reported musculoskeletal pain. Work. 2017;57:563–72. doi:10.3233/WOR-172582.

19.

Hansson

G-åke

, Balogh

, Ohlsson

, Granqvist

, Nordander

, Arvidsson

, et al. Physical workload in various types of work: Part I. Wrist and forearm. Int J Ind Ergon. 2009;39:221–33. doi:10.1016/j.ergon.2008.04.003.

20.

Mathiassen

, Möller

, Forsman

. Variability in mechanical exposure within and between individuals performing a highly constrained industrial work task. Ergonomics. 2003;46:800–24. doi:10.1080/0014013031000090125.

21.

Jonsson

. Measurement and evaluation of local muscular strain in the shoulder during constrained work. J Hum Ergol (Tokyo). 1982;11:73–88.

22.

Byström

, Hansson

G-A

, Rylander

, Ohlsson

, Källrot

, Skerfving

. Physical workload on neck and upper limb using two CAD applications. Appl Ergon. 2002;33:63–74. doi:10.1016/S0003-6870(01)00044-8.

23.

Voerman

, Vollenbroek-Hutten

, Hermens

. Upper trapezius muscle activation patterns in neck-shoulder pain patients and healthy controls. Eur J Appl Physiol. 2007;102:1–9. doi:10.1007/s00421-006-0215-8.

24.

Bongers

, Kremer

, ter Laak

. Are psychosocial factors, risk factors for symptoms and signs of the shoulder, elbow, or hand/wrist? A review of the epidemiological literature. Am J Ind Med. 2002;41:315–42. doi:10.1002/ajim.10050.

25.

Sjøgaard

, Rosendal

, Kristiansen

, Blangsted

, Skotte

, Larsson

, et al. Muscle oxygenation and glycolysis in females with trapezius myalgia during stress and repetitive work using microdialysis and NIRS. Eur J Appl Physiol. 2010;108:657–69. doi:10.1007/s00421-009-1268-2.

26.

Taib

MFM

, Bahn

, Yun

. The effect of psychosocial stress on muscle activity during computer work: Comparative study between desktop computer and mobile computing products. Work. 2016;54:543–55. doi:10.3233/WOR-162334.

27.

Mathiassen

, Nordander

, Svendsen

, Wellman

, Dempsey

. Task-based estimation of mechanical job exposure in occupational groups. Scand J Work Environ Health. 2005;31:138–51. doi:10.5271/sjweh.861.

28.

Mathiassen

, Burdorf

, van der Beek

. Statistical power and measurement allocation in ergonomic intervention studies assessing upper trapezius EMG amplitude. A case study of assembly work. J Electromyogr Kinesiol. 2002;12:45–57.

29.

Wahlström

. Ergonomics, musculoskeletal disorders and computer work. Occup Med (Lond). 2005;55:168–76. doi:10.1093/occmed/kqi083.

30.

Mathiassen

. Diversity and variation in biomechanical exposure: What is it, and why would we like to know? Appl Ergon. 2006;37:419–27. doi:10.x1016/j.apergo.2006.04.006.

31.

Richter

, Mathiassen

, Slijper

, Over

E a B

, Frens

M a

. Differences in muscle load between computer and non-computer work among office workers. Ergonomics. 2009;52:1540–55. doi:10.1080/00140130903199905.

32.

Robertson

, Ciriello

, Garabet

. Office ergonomics training and a sit-stand workstation: Effects on musculoskeletal and visual symptoms and performance of office workers. Appl Ergon. 2013;44:73–85. doi:10.1016/j.apergo.2012.05.001.

33.

Barbieri

, Srinivasan

, Mathiassen

, Oliveira

. Comparison of Sedentary Behaviors in Office Workers Using Sit-Stand Tables With and Without Semiautomated Position Changes. Hum Factors J Hum Factors Ergon Soc. 2017;59:782–95. doi:10.1177/0018720817695771.

34.

Galinsky

, Swanson

, Sauter

, Dunkin

, Hurrell

, Schleifer

. Supplementary breaks and stretching exercises for data entry operators: A follow up field study. Am J Ind Med. 2007;50:519–27. doi:10.1002/ajim.20472.

35.

Samani

, Holtermann

, Søgaard

, Madeleine

. Active pauses induce more variable electromyographic pattern of the trapezius muscle activity during computer work. J Electromyogr Kinesiol. 2009;19:e430–7. doi:10.1016/j.jelekin.2008.11.011.

36.

Januario

, Moreira R de

, Cid

, Samani

, Madeleine

, Oliveira

. Effects of active pause pattern of surface electromyographic activity among subjects performing monotonous tasks: A systematic review. J Electromyogr Kinesiol. 2016;30:196–208. doi:10.1016/j.jelekin.2016.07.009.

37.

Andersen

, Kjaer

, Søgaard

, Hansen

, Kryger

, Sjøgaard

. Effect of two contrasting types of physical exercise on chronic neck muscle pain. Arthritis Rheum. 2008;59:84–91. doi:10.1002/art.23256.

38.

Straker

, Mathiassen

, Holtermann

. The “Goldilocks Principle”: Designing physical activity at work to be “just right” for promoting health. Br J Sports Med. 2017: bjsports–2017–097765. doi:10.1136/bjsports-2017-097765.

39.

Johansen

, Samani

, Antle

, Côté

, Madeleine

. Gender effects on the coordination of subdivisions of the trapezius muscle during a repetitive box-folding task. Eur J Appl Physiol. 2013;113:175–82. doi:10.1007/s00421-012-2425-6.

40.

Szucs

, Molnar

. Differences in the activation and co-activation ratios of the four subdivisions of trapezius between genders following a computer typing task. Hum Mov Sci. 2017;52:181–90. doi:10.1016/j.humov.2017.02.001.