Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jul 27.
Published in final edited form as: Appl Psychophysiol Biofeedback. 2016 Jun;41(2):181–189. doi: 10.1007/s10484-015-9328-3

Evaluation of Novel EMG Biofeedback for Postural Correction During Computer Use

Brecca M Gaffney 1, Katrina S Maluf 2, Bradley S Davidson 1
PMCID: PMC9328112  NIHMSID: NIHMS1824307  PMID: 26718205

Abstract

Postural correction is an effective rehabilitation technique used to treat chronic neck and shoulder pain, and is aimed toward reducing the load on the surrounding muscles by adopting a neutral posture. The objective of this investigation was to evaluate the effectiveness of real-time high-density surface EMG (HDsEMG) biofeedback for postural correction during typing. Twenty healthy participants performed a typing task with two forms of postural feedback: (1) verbal postural coaching and (2) verbal postural coaching plus HDsEMG biofeedback. The interface used activity from two HDsEMG arrays placed over the trapezius designed to shift trapezius muscle activity inferiorly. The center of gravity across both arrays was used to quantify the spatial distribution of trapezius activity. Planar angles taken from upper extremity reflective markers quantified cervicoscapular posture. During the biofeedback condition, trapezius muscle activity was located 12.74 ± 3.73 mm more inferior, the scapula was 2.58 ± 1.18° more adducted and 0.23 ± 0.24° more depressed in comparison to verbal postural coaching alone. The results demonstrate the short-term effectiveness of a real-time HDsEMG biofeedback intervention to achieve postural correction, and may be more effective at creating an inferior shift in trapezius muscle activity in comparison to verbal postural coaching alone.

Keywords: Biofeedback, Postural correction, Trapezius, Cervicoscapular posture, Electromyography

Introduction

Forty-nine percent of people experience at least one occurrence of neck pain or shoulder pain during their lifetime (Fejer et al. 2006). Chronic neck and shoulder pain commonly occur in a computer-dominated work environment because computer users adopt poor neck and shoulder postures (forward head posture, elevated and abducted scapulae) for long periods, which increases static loading on surrounding muscles such as the cervical extensors and upper trapezius (Ariëns et al. 2001; Szeto et al. 2005; Falla et al. 2007a, b). Chronic overactivation of the upper trapezius leads to trapezius myalgia, defined by tightness and palpable tender points in the trapezius muscle, and is documented in one-third of workers with chronic neck or shoulder pain (Sjøgaard et al. 2006).

Postural correction is an effective technique used by rehabilitation professionals to treat chronic neck and shoulder pain, and is aimed toward reducing mechanical loads on cervicoscapular muscles by adopting a neutral posture (Falla et al. 2007a, b; Jull et al. 2002). Postural correction shifts trapezius muscle activity more inferiorly (Enwemeka et al. 1986; McLean 2005; Wegner et al. 2010), and reduces loading of the upper trapezius which is a common site of localized pain. In previous work, scapular postural correction in patients with neck pain effectively shifted trapezius muscle activity inferiorly during computer work, and resulted in similar activation patterns to healthy controls (Wegner et al. 2010).

It may be possible to correct cervicoscapular posture by redistributing muscle activity inferiorly in the trapezius using real-time electromyography (EMG) biofeedback with high-density surface arrays (HDsEMG). EMG biofeedback has been used to retrain motor patterns and may help achieve long-lasting changes in muscle activity to prevent or treat musculoskeletal pain (Hermens and Hutten 2002). EMG biofeedback has also been used to retrain muscles of the upper extremity (Nord et al. 2001; Voerman et al. 2004; Holtermann et al. 2009, 2010), and to improve cervical and scapular postures (Samani et al. 2010). In addition, Peper et al. (2003) implemented a surface EMG biofeedback system designed to lower muscle activity in the cerviscoscapular region during computer work in healthy individuals. However, the biofeedback techniques used in these investigations were limited to the small detection zones used in bipolar EMG recordings.

One investigation demonstrated that real-time biofeedback from two sets of bipolar electrodes on the clavicular and descending portions of the upper trapezius during computer work altered the spatial organization in the trapezius measured from simultaneous HDsEMG recordings (Samani et al. 2010). The participants received feedback based on bipolar EMG recordings that signaled active (approximately 30 % maximum voluntary contraction) or passive (rest) pauses during the typing task. However, they did not report the effects of the biofeedback on cervicoscapular posture. Given the known heterogeneity of trapezius muscle activation across individual subdivisions (Falla and Farina 2008; Farina et al. 2008; Madeleine and Farina 2008; Gallina et al. 2013), feedback based on HDsEMG signals may help redistribute trapezius activity between the upper and lower subdivisions. Heterogeneity of muscle activation across the trapezius is well documented using HDsEMG, and indicates a non-uniform distribution of motor unit recruitment within individual subdivisions (Kleine et al. 2000; Holtermann et al. 2005; Farina et al. 2008; Samani et al. 2010).

The objectives of this investigation were to develop a custom HDsEMG biofeedback interface to monitor relative activation of the upper and lower subdivisions of the trapezius muscle, and to assess the effects of this interface on cervicoscapular posture and the distribution of trapezius muscle activity during computer work in healthy individuals. We hypothesized that the HDsEMG biofeedback interface would help participants: (1) maintain a more neutral cervical and scapular posture across the duration of the task and (2) achieve a more inferior distribution of trapezius muscle activity compared to postural correction without of biofeedback.

Methods

Participants

Twenty healthy participants with no prior history of neck or shoulder pain (10 females, age, mean ± SD 36.5 ± 13.9 years, height, 171.1 ± 5.9 cm, weight, 77.5 ± 9.5 kg and 10 males, age, 30.4 ± 9.9 years, height, 183.8 ± 8.0 cm, weight, 84.1 ± 12.8 kg) participated in this investigation. The study was conducted in accord with procedures approved by the Colorado Multiple Institutional Review Board, and written informed consent was obtained from all participants prior to the experiment.

Experimental Protocol

Each participant performed a one-min baseline data collection to establish their self-selected posture in which they were given no postural instruction (Fig. 1a). Each participant was coached by a researcher on how to sit with a neutral posture at a standard computer workstation without arm rests that was otherwise adjusted according to ergonomic recommendations (Burgess-Limerick et al. 1999) with the elbows positioned in line with the trunk and flexed 90°–120°. Each participant was instructed to maintain a “neutral” posture with the trunk upright, chin tucked, and scapulae slightly adducted and depressed (Fig. 1b). After coaching and practice, each participant maintained the neutral posture during a one-min baseline data collection. Each participant performed a 15-min typing task under two different feedback conditions:

Fig. 1.

Fig. 1

Open in a new tab

Cervical posture (θCV), scapular elevation/depression (θEL), and scapular adduction/abduction (θAB) defined by planar angles in the sagittal, transverse, and frontal planes during a self-selected posture and b coached neutral posture. Posture angles expressed in a torso-fixed reference frame with origin at C7

  1. Verbal feedback from the researcher to adopt a neutral posture prior to the start of the typing task, with no subsequent verbal or EMG feedback during the task (No Biofeedback Condition).

  2. Verbal feedback from the researcher to adopt a neutral posture prior to the task, plus real-time HDsEMG corrective feedback throughout the task (Biofeedback Condition).

Each participant was given sentences to mirror (Bruce’s Unusual Typing Wizard) and was instructed to type at a self-selected speed as accurately as possible (Fig. 2). Each participant rested for one to 3 min between feedback conditions. Presentation order of the feedback condition was randomized across all participants.

Fig. 2.

Fig. 2

Open in a new tab

Screen shots of typing task with real-time corrective feedback from HDsEMG signals. Biofeedback interface displayed an activity ratio (AR) between superior and inferior arrays. Higher AR represented a more superior distribution of trapezius muscle activity. Mean AR from coached neutral posture served as the threshold for the interface. When the threshold was exceeded for one continuous second, the interface displayed an onscreen visual alert (bar graph changed from green to red). a Indicates AR below threshold (green bar) and b indicates AR above threshold (red bar)

A custom biofeedback interface in Matlab (Mathworks, Natick, MA) that displayed real-time trapezius muscle activity using two HDsEMG electrode arrays placed over the upper, middle, and lower fibers of the trapezius (Fig. 3) was developed to alter the relative distribution (i.e., shift) trapezius muscle activity inferiorly (Fig. 4). The interface displayed an activity ratio (AR), which indicated how muscle activity was distributed between the upper and middle/lower fibers of the trapezius. The AR was calculated by dividing the sum of the rectified signals from all electrodes in the superior array (upper fibers) by the summation of the rectified signals from all electrodes in the inferior array (middle/lower fibers), which provides an index of the relative magnitude of activation between different regions of the trapezius muscle. Therefore, an increase in AR indicated a more superior distribution of the geometric center (center of gravity) of trapezius muscle activation. The mean AR recorded during the one-min coached neutral posture at baseline was used to create an individual threshold for the biofeedback interface for each participant. When the real-time AR exceeded the mean AR for one consecutive second, the participant received an onscreen visual alert in which the interface turned from green to red.

Fig. 3.

Fig. 3

Open in a new tab

a HDsEMG arrays placed on the upper, middle, and lower fibers of the trapezius and b HDsEMG array schematic. The missing electrode of the inferior array represented the origin for array orientation. To standardize array placement, the superior array was placed with the 4th row along the C7-acromion line

Fig. 4.

Fig. 4

Open in a new tab

Topographical map (interpolation by a factor of 8) of 51 bipolar average rectified values of superior HDsEMG electrode array showing the center of gravity (white dot) in a self-selected posture and b during use of HDsEMG biofeedback designed to shift trapezius muscle activity inferior

To interpret the biofeedback, each participant was informed of the meaning of the numerical value interface (AR) (Fig. 2). Each participant was instructed to maintain their AR value as low as possible to avoid exceeding the threshold, which would cause the interface to turn red (i.e. “keep the interface green for as long as possible”). To comply with the biofeedback (lower AR), each participant was instructed to alter the position of the scapulae (slightly adduct and depress).

Instrumentation

Reflective markers placed on the acromion process, seventh cervical vertebrae (C7), and the tragus of the ear were used to form planar angles and quantify changes in cervical and scapular posture across the two feedback conditions. Motion capture signals were recorded with a sampling frequency of 100 Hz (Vicon, Centennial, CO).

Surface EMG signals were detected from the dominant upper, middle, and lower fibers of the trapezius muscle with two semi-disposable electrode arrays (ELSCH064NM2 Pin Out, OT Bioelettronica, Torino, Italy). The superior array was placed over the upper (UT) and middle trapezius (MT). The inferior array was placed over the MT and lower trapezius (LT) (Fig. 3a). Each electrode array was composed of 64 electrodes in a 13 × 5 orientation (3-mm electrode diameter, 8-mm inter-electrode distance). One electrode was missing in the bottom right corner of each array, which served as the origin of the coordinate system used for electrode orientation in post processing (Fig. 3b). EMG signals were amplified by a gain of 5000 (64 channel surface EMG-USB2 amplifier, OT Bioelettronica, Torino, Italy; bandwidth 10–500 Hz), sampled at 2048 Hz, and converted to a digital signal by a 12-bit A/D converter. Visual inspection of raw EMG signals was performed offline to identify channels with poor contact or short circuits, and these signals were replaced by linear interpolation using adjacent channels (Gallina et al. 2013).

Electrode placement was standardized across all participants according to previous work (Farina et al. 2002). The 4th row of the superior electrode array was placed along the C7-acromion line with the most medial column 10-mm from the innervation zone (Farina et al. 2008). The innervation zone was identified using a dry linear array (SA 16/5, OT Bioelettronica, Torino, Italy) composed of 16 silver bar electrodes (5-mm inter-electrode distance, 1-mm width). The inferior electrode array was placed in line with the superior electrode array. Prior to electrode placement, the skin was lightly scrubbed with gauze pads and an abrasive cream until the skin turned slightly red, indicating removal of the superficial skin layer. When necessary, excess body hair was shaved to ensure high quality EMG signals. A reference electrode was placed on the right wrist.

Data Collection

HDsEMG and motion capture signals were recorded simultaneously. To ensure synchronization between systems throughout the duration of each typing task, a 5-s signal was sent to both the HDsEMG amplifier and motion capture system every 30 s for the duration of each typing task. Dependent variables were averaged across the middle 3 s of this synchronization window and used for statistical analysis.

Data Processing

Cervical and scapular postures were quantified during each feedback condition by two measures: cranial-vertebral angle and scapular position. Marker data were filtered with a 4th-order zero-phase-lag Butterworth filter (5 Hz lowpass cutoff frequency). The cranial-vertebral angle (θCV) was defined as the angle formed between C7 and the tragus with respect to the frontal plane expressed in a torso-fixed reference frame (Fig. 1). Scapular posture was quantified using the angle formed by C7 and the acromion process in the frontal and transverse planes as defined in a torso reference frame (Fig. 1) (Gaffney et al. 2014); scapular adduction/abduction was defined as the angle between C7 and the acromion in the transverse plane (θAB) (Fig. 1), and scapular elevation/depression was defined as the angle between C7 and the acromion in the frontal plane (θEL).

EMG signals were filtered using a 4th-order, zero-phaselag, Butterworth filter (10–500 Hz bandpass cutoff frequencies). Fifty-one bipolar signals were calculated from each electrode array. The average rectified value (ARV) was computed from each bipolar recording from adjacent, non-overlapping signal epochs of 0.5 s. To characterize the spatial distribution of trapezius muscle activity, the center of gravity (COG) in the medial–lateral direction (XCOG) and the inferior-superior direction (YCOG) (Fig. 3b) was calculated from the 51 bipolar ARV recordings (Farina et al. 2008).

All variables were referenced to the coached neutral posture recorded at baseline (Gaffney et al. 2014); therefore, small values indicated similarities to the coached neutral posture. Changes in cervical posture were defined as ΔθCV. Positive ΔθCV angles indicated movement in the direction of cervical flexion, and negative ΔθCV angles indicated movement in the direction of cervical extension. Changes in scapular posture were defined as ΔθAB and ΔθEL. Positive ΔθAB angles indicated movement in the direction of scapular abduction, and negative ΔθAB angles indicated movement in the direction of scapular adduction. Positive ΔθEL angles indicated movement in the direction of scapular elevation, and negative ΔθEL angles indicated movement in the direction of scapular depression. Changes (shifts) in trapezius muscle activity were defined as ΔXCOG and ΔYCOG, where positive ΔXCOG and ΔYCOG values indicated a more medial and superior distribution of trapezius muscle activity with respect to the coached neutral posture, respectively.

Statistical Analysis

To determine whether muscle activity or posture changed as a function of time during each 15-min typing task, the average slope for each dependent variable was tested against zero with two-tailed one-sample t tests. Slopes of each dependent variable for each participant were calculated and then averaged across all participants for inferential tests. Differences in each dependent variable between the two feedback conditions were examined using two-tailed paired t-tests. All statistical calculations were performed with JMP Pro 10 (Cary, NC). Level of significance was set at 0.05 for all statistical tests.

Results

The neutral posture resulted in a slight greater cervical extension, scapular adduction, and scapular elevation; and the distribution of trapezius muscle activity was slightly more superior and medial in comparison to the self-selected posture (Table 1).

Table 1.

Dependent variables during the self-selected posture and the coached neutral posture baseline data collections

Self-selected Neutral
θ CV 33.2 ± 5.9° 38.2 ± 4.1°
θ AB 17.3 ± 4.5° 14.6 ± 5.0°
θ EL −19.4 ± 4.2° −18.0 ± 3.4°
X COG 13.9 ± 1.8 mm 13.5 ± 3.5 mm
Y COG 107.4 ± 27.4 mm 102.9 ± 31.6 mm
Open in a new tab

Effects of Feedback Condition on Cervicoscapular Posture and Muscle Activity

θAB was 2.6 ± 1.2° smaller when typing with HDsEMG biofeedback in comparison to no biofeedback (P < 0.001) (Fig. 5a). θEL was 0.2 ± 0.2° smaller when typing with HDsEMG biofeedback in comparison to no biofeedback (P < 0.001) (Fig. 5b). There were no differences in θCV across feedback conditions (P = 0.30) (Fig. 5c). XCOG was 0.8 ± 0.2 mm more lateral when typing with HDsEMG biofeedback in comparison to no biofeedback (P < 0.001) (Fig. 5d). YCOG was 12.7 ± 3.7 mm more inferior when typing with HDsEMG biofeedback in comparison to typing with no biofeedback (P < 0.001) (Fig. 5e).

Fig. 5.

Fig. 5

Open in a new tab

Changes in a scapular abduction/adduction (ΔθAB), b scapular elevation/depression (ΔθEL), c cervical posture (ΔθCV); distribution of trapezius muscle activity in the d medial–lateral direction (ΔXCOG), and e superior-inferior direction (ΔYCOG) with respect to coached neutral posture between feedback conditions (No Biofeedback and Biofeedback). Asterisk indicates statistically significant difference from neutral posture (P < 0.05). Filled diamond indicates statistically significant difference between feedback conditions (P < 0.05)

Time-Dependent Changes in Cervicoscapular Posture and Trapezius Muscle Activity

Positive slopes for cervical and scapular posture variables during the 15-min typing tasks indicate a migration into poorer posture over the duration of the task. The average slope of ΔθCV was less than zero (P = 0.02) (Fig. 6a) when typing with HDsEMG biofeedback, and indicates increasing cervical extension throughout the typing task. The average slope of ΔYCOG was positive (P = 0.02) (Fig. 6b) when typing with HDsEMG biofeedback, and indicates a superior shift of trapezius muscle activity throughout the duration of the typing task (Table 2). No statistically significant differences were found for all other variables. Similarly, no statistically significant differences were found when comparing the slopes of variables between feedback conditions.

Fig. 6.

Fig. 6

Open in a new tab

Change in a cervical posture (ΔθCV) and b trapezius muscle activity in the inferior-superior direction (ΔYCOG) for Biofeedback (gray) and No Biofeedback (black) conditions for a representative participant

Table 2.

Mean ± SD of slopes of dependent variables across both feedback conditions

HDsEMG biofeedback Verbal postural feedback
θ CV −0.05 ± 0.09 * −0.02 ± 0.13
θ AB 0.01 ± 0.11 0.03 ± 0.14
θ EL −0.02 ± 0.08 −0.003 ± 0.08
X COG 0.00 ± 0.00 0.00 ± 0.00
Y COG 0.29 ± 0.50 * 0.01 ± 0.68
Open in a new tab

Bold indicates statistically significant difference

*

Slope was significantly different from zero

Discussion

We investigated the ability of a custom real-time biofeedback intervention using HDsEMG to alter cervicoscapular posture and the spatial distribution of trapezius muscle activity during computer work. When compared to verbal postural correction alone, the biofeedback interface successfully shifted trapezius muscle activity inferiorly (i.e. toward the middle and lower trapezius). HDsEMG biofeedback improved the ability to maintain a neutral scapular posture compared to verbal cueing alone, as demonstrated by greater scapular adduction and depression in the biofeedback condition. These findings indicate that postural correction guided by real-time HDsEMG biofeedback results in short-term improvements in scapular posture, achieved through an inferior shift in trapezius muscle activity.

To our knowledge, this is the first investigation that uses real-time biofeedback guided by muscle activity recorded from HDsEMG electrodes for correction of cervicoscapular posture. Previous investigations with real-time postural biofeedback have implemented postural rather than EMG signals within the feedback loop to maintain ideal posture (McLean, 2005; Wegner et al. 2010), or have used bipolar surface electrodes, which do not reflect the overall spatial distribution of trapezius muscle activity (Nord et al. 2001; Peper et al. 2003; Holtermann et al. 2009; Voerman et al. 2004). HDsEMG biofeedback helped participants achieve a more adducted and depressed scapular posture that corresponded to an inferior shift in trapezius muscle activity during the typing task. In contrast, participants assumed a less ideal posture in the absence of real-time postural biofeedback. These observations demonstrate the ability to accomplish postural correction using HDsEMG in the absence of ongoing verbal or other types of postural feedback cues. We previously demonstrated that an inferior shift of trapezius muscle activity could be achieved by altering the position of the scapulae (Gaffney et al. 2014); however, until now it was unclear if the opposite relationship held true (shifting trapezius muscle activity inferior causing a change in position of the scapulae).

Both forms of postural feedback resulted in greater cervical extension (decreased forward head posture) compared to the coached neutral posture, with no difference observed between the two feedback conditions. Research consistently supports that computer users adopt a more forward head posture (greater cervical flexion) over the duration of a task (Ankrum and Nemeth 2000; Szeto et al. 2002; McLean 2005; Falla et al. 2007a, b; Yoo and Kim 2010). However, our results showed the opposite trend, with a gradual increase in cervical extension across time observed only for the biofeedback condition. Although verbal instruction and biofeedback both appear to prevent the migration into a forward head posture typically seen during computer work, the addition of biofeedback may augment verbal instruction by encouraging active adjustments of cervical posture across time.

HDsEMG biofeedback resulted in an inferior shift in trapezius muscle activity compared to the coached neutral posture, whereas verbal postural feedback alone resulted in a superior shift in trapezius muscle activity. Differences in the distribution of muscle activity between feedback conditions may be associated with the continuity of feedback administered throughout the duration of the typing task. HDsEMG biofeedback provided continuous real-time postural correction, whereas verbal postural feedback was given only before the typing task began. Therefore, participants may have neglected the verbal postural instructions when confronted with additional cognitive demands during the typing task. Postural correction through verbal cueing is commonly used in clinical settings; however, implementing postural correction using HDsEMG biofeedback may be more effective to help individuals maintain an inferior location in trapezius muscle activity while completing daily tasks in the workplace. Real-time HDsEMG biofeedback is currently impractical to implement in most clinical and work settings. However, this approach could be selectively implemented for individuals at highest risk for cervical overuse injuries or who are resistant to other types of interventions for trapezius myalgia. The efficacy of lower cost options for EMG biofeedback using an AR interface derived from bipolar electrode pairs placed over the upper and lower subdivisions of the trapezius should also be explored.

In addition to the inferior shift, a more lateral distribution of trapezius muscle activity was present with HDsEMG feedback. This observation is consistent with previous investigations showing a lateral shift in the center of gravity (Kleine et al. 2000; Madeleine et al. 2006; Farina et al. 2008; Samani et al. 2010) when the overall spatial distribution of relative trapezius muscle activation is altered. Although these findings are statistically significant, the amount of change (<1 mm) may not be clinically relevant.

Although the biofeedback interface was designed to constrain the activity ratio (AR) to that of the neutral posture, a subgroup of four participants demonstrated higher AR when typing with HDsEMG biofeedback in comparison to verbal postural coaching. There are at least three plausible explanations regarding the inability of these individuals to comply with the biofeedback instructions. First, the higher AR during HDsEMG biofeedback may represent an inability to redistribute muscle activity toward the middle and lower trapezius. These participants may have lacked sufficient muscle strength, endurance, or coordination to successfully maintain a low AR during the biofeedback condition. Second, this subset may have required additional time to learn the motor task. These participants demonstrated high variability in scapular movements, and exceeded the AR threshold more times than the other participants during the biofeedback condition (subgroup exceeded AR: 301 ± 228 times, others exceeded AR: 183 ± 67 times). Third, the distribution of trapezius muscle activity in the self-selected sitting posture was slightly superior for this subgroup in comparison to the rest of the population (subgroup neutral posture YCOG: 112.6 ± 7.6 mm, remaining self-selected posture YCOG: 105.9 ± 30.9 mm). This may indicate that individuals with more extreme deviations from ideal patterns of muscle activity in their preferred sitting posture may respond less favorably to HDsEMG biofeedback. Future research is needed to identify relevant characteristics that predict which individuals are likely to benefit most from EMG-based biofeedback interventions.

The overall superior shift in trapezius muscle activity (positive slope) and increased cervical extension (negative slope) during the HDsEMG feedback condition may be related to the design of the biofeedback interface. The interface was designed to prevent a superior shift in trapezius activity above that recorded during the coached neutral posture. Therefore, the gradual superior shift in trapezius activity may represent a redistribution of activity more similar to habitual patterns of muscle activity in the preferred sitting posture. Although the slope of trapezius muscle activity was positive throughout the duration of the HDsEMG feedback condition, the average location of activity was inferior to the coached neutral posture, which supports the effectiveness of the interface. Because the interface was not designed for cervical posture correction, the gradual increase of cervical extension throughout the HDsEMG feedback condition may indicate a migration toward their habitual cervical posture, which was slightly more extended (Table 1).

Several limitations should be considered in this investigation. First, the participants in this investigation had no history of chronic neck or shoulder pain. Although we cannot immediately extend these results to a clinical population, the effective short-term results of HDsEMG biofeedback on muscle activity and scapular posture in healthy participants indicates that the proposed interface may hold promise for future clinical applications. Previous studies indicate that reducing muscle activity and taking rest breaks during computer work may help reduce neck and shoulder pain (Enwemeka et al. 1986; Hermens and Hutten, 2002; Vollenbroek-Hutten et al. 2006; Wegner et al. 2010; Yip et al. 2008); however, it is not known whether changing the distribution of muscle activity between different subdivisions of the trapezius muscle is sufficient to relieve pain. Second, the participants were given only one to 2 min to familiarize themselves with the HDsEMG biofeedback interface, which may have been an inadequate amount of time to effectively learn the task. This investigation examined short-term changes in posture and muscle activation during a single experimental collection; therefore, long-term retention of training effects with more prolonged or repeated training sessions remains to be determined.

Conclusion

This investigation is the first to demonstrate the ability to modify scapular posture using real-time HDsEMG biofeedback to encourage an inferior shift in trapezius muscle activity. Healthy participants were able to maintain a neutral posture while typing, which coincided with a more inferior distribution of trapezius muscle activity. Future studies should examine the long-term efficacy of EMG biofeedback interventions to improve posture and prevent or treat neck/shoulder pain during computer work.

Acknowledgments

We thank Brett Donnermeyer and Oscar Reyes for their contributions during data collections.

References

  1. Ankrum DR, & Nemeth KJ (2000). Head and neck posture at computer workstations—what’s neutral? In Proceedings of the 14th Triennial Congress of the International Ergonomics Association, 5, 565–568. [Google Scholar]
  2. Ariëns GA, Bongers PM, Douwes M, Miedema MC, Hoogendoorn WE, van der Wal G, & van Mechelen W (2001). Are neck flexion, neck rotation, and sitting at work risk factors for neck pain? Results of a prospective cohort study. Occupational and Environmental Medicine, 58(3), 200–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Burgess-Limerick R, Plooy A, Fraser K, & Ankrum D (1999). The influence of computer monitor height on head and neck posture. International Journal of Industrial Ergonomics, 23(3), 171–179. [Google Scholar]
  4. Enwemeka CS, Bonet IM, Ingle JA, Prudhithumrong S, Ogbahon FE, & Gbenedio NA (1986). Postural correction in persons with neck pain (II. Integrated electromyography of the upper trapezius in three simulated neck positions). The Journal of Orthopaedic and Sports Physical Therapy, 8(5), 240–242. [DOI] [PubMed] [Google Scholar]
  5. Falla D, & Farina D (2008). Non-uniform adaptation of motor unit discharge rates during sustained static contraction of the upper trapezius muscle. Experimental Brain Research. Experimentelle Hirnforschung. Expérimentation Cérébrale, 191(3), 363–370. doi: 10.1007/s00221-008-1530-6. [DOI] [PubMed] [Google Scholar]
  6. Falla D, Jull G, Russell T, Vicenzino B, & Hodges P (2007a). Effect of neck exercise on sitting posture in patients with chronic neck pain. Physical Therapy, 87(4), 408–417. [DOI] [PubMed] [Google Scholar]
  7. Falla D, O’Leary S, Fagan A, & Jull G (2007b). Recruitment of the deep cervical flexor muscles during a postural-correction exercise performed in sitting. Manual Therapy, 12(2), 139–143. [DOI] [PubMed] [Google Scholar]
  8. Farina D, Leclerc F, Arendt-Nielsen L, Buttelli O, & Madeleine P (2008). The change in spatial distribution of upper trapezius muscle activity is correlated to contraction duration. Journal of Electromyography and Kinesiology, 18(1), 16–25. [DOI] [PubMed] [Google Scholar]
  9. Farina D, Madeleine P, Graven-Nielsen T, Merletti R, & Arendt-Nielsen L (2002). Standardising surface electromyogram recordings for assessment of activity and fatigue in the human upper trapezius muscle. European Journal of Applied Physiology, 86(6), 469–478. [DOI] [PubMed] [Google Scholar]
  10. Fejer R, Kyvik KO, & Hartvigsen J (2006). The prevalence of neck pain in the world population: a systematic critical review of the literature. European Spine Journal, 15(6), 834–848. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gaffney BM, Maluf KS, Curran-Everett D, & Davidson BS (2014). Associations between cervical and scapular posture and the spatial distribution of trapezius muscle activity. Journal of Electromyography and Kinesiology, 24(4), 542–549. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Gallina A, Merletti R, & Gazzoni M (2013). Uneven spatial distribution of surface EMG: What does it mean? European Journal of Applied Physiology, 113(4), 887–894. [DOI] [PubMed] [Google Scholar]
  13. Hermens HJ, & Hutten MM (2002). Muscle activation in chronic pain: its treatment using a new approach of myofeedback. International Journal of Industrial Ergonomics, 30(4–5), 325–336. [Google Scholar]
  14. Holtermann A, Mork PJ, Andersen LL, Olsen HB, & Søgaard K (2010). The use of EMG biofeedback for learning of selective activation of intra-muscular parts within the serratus anterior muscle: A novel approach for rehabilitation of scapular muscle imbalance. Journal of Electromyography and Kinesiology, 20(2), 359–365. [DOI] [PubMed] [Google Scholar]
  15. Holtermann A, Roeleveld K, & Karlsson JS (2005). Inhomogeneities in muscle activation reveal motor unit recruitment. Journal of Electromyography and Kinesiology, 15(2), 131–137. [DOI] [PubMed] [Google Scholar]
  16. Holtermann A, Roeleveld K, Mork PJ, Grönlund C, Karlsson JS, Andersen LL, & Søgaard K (2009). Selective activation of neuromuscular compartments within the human trapezius muscle. Journal of Electromyography and Kinesiology, 19(5), 896–902. [DOI] [PubMed] [Google Scholar]
  17. Jull G, Trott P, Potter H, Zito G, Niere K, Shirley D, & Richardson C (2002). A randomized controlled trial of exercise and manipulative therapy for cervicogenic headache. Spine, 27(17), 1835–1843. [DOI] [PubMed] [Google Scholar]
  18. Kleine BU, Schumann NP, Stegeman DF, & Scholle HC (2000). Surface EMG mapping of the human trapezius muscle: The topography of monopolar and bipolar surface EMG amplitude and spectrum parameters at varied forces and in fatigue. Clinical Neurophysiology, 111(4), 686–693. [DOI] [PubMed] [Google Scholar]
  19. Madeleine P, & Farina D (2008). Time to task failure in shoulder elevation is associated to increase in amplitude and to spatial heterogeneity of upper trapezius mechanomyographic signals. European Journal of Applied Physiology, 102(3), 325–333. doi: 10.1007/s00421-007-0589-2. [DOI] [PubMed] [Google Scholar]
  20. Madeleine P, Leclerc F, Arendt-Nielsen L, Ravier P, & Farina D (2006). Experimental muscle pain changes the spatial distribution of upper trapezius muscle activity during sustained contraction. Clinical Neurophysiology, 117(11), 2436–2445. [DOI] [PubMed] [Google Scholar]
  21. McLean L (2005). The effect of postural correction on muscle activation amplitudes recorded from the cervicobrachial region. Journal of Electromyography and Kinesiology, 15(6), 527–535. [DOI] [PubMed] [Google Scholar]
  22. Nord S, Ettare D, Drew D, & Hodge S (2001). Muscle learning therapy—efficacy of a biofeedback based protocol in treating work-related upper extremity disorders. Journal of Occupational Rehabilitation, 11(1), 23–31. [DOI] [PubMed] [Google Scholar]
  23. Peper E, Wilson VS, Gibney KH, Huber K, Harvey R, & Shumay DM (2003). The integration of electromyography (SEMG) at the workstation: assessment, treatment, and prevention of repetitive strain injury (RSI). Applied Psychophysiology Biofeedback, 28(2), 167–182. [DOI] [PubMed] [Google Scholar]
  24. Samani A, Holtermann A, Søgaard K, & Madeleine P (2010). Active biofeedback changes the spatial distribution of upper trapezius muscle activity during computer work. European Journal of Applied Physiology, 110(2), 415–423. [DOI] [PubMed] [Google Scholar]
  25. Sjøgaard G, Søgaard K, Hermens HJ, Sandsjö L, Läubli T, Thorn S, & Merletti R (2006). Neuromuscular assessment in elderly workers with and without work related shoulder/neck trouble: The NEW-study design and physiological findings. European Journal of Applied Physiology, 96(2), 110–121. [DOI] [PubMed] [Google Scholar]
  26. Szeto GPY, Straker LM, & O’Sullivan PB (2005). A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—2: Neck and shoulder kinematics. Manual Therapy, 10(4), 281–291. [DOI] [PubMed] [Google Scholar]
  27. Szeto GPY, Straker L, & Raine S (2002). A field comparison of neck and shoulder postures in symptomatic and asymptomatic office workers. Applied Ergonomics, 33(1), 75–84. [DOI] [PubMed] [Google Scholar]
  28. Voerman GE, Sandsjö L, Vollenbroek-Hutten MMR, Groothuis-Oudshoorn CGM, & Hermens HJ (2004). The influence of different intermittent myofeedback training schedules on learning relaxation of the trapezius muscle while performing a gross-motor task. Euro, 93(1–2), 57–64. [DOI] [PubMed] [Google Scholar]
  29. Vollenbroek-Hutten M, Hermens H, Voerman G, Sandsjö L, & Kadefors R (2006). Are changes in pain induced by myofeedback training related to changes in muscle activation patterns in patients with work-related myalgia? European Journal of Applied Physiology, 96(2), 209–215. doi: 10.1007/s00421-004-1212-4. [DOI] [PubMed] [Google Scholar]
  30. Wegner S, Jull G, O’Leary S, & Johnston V (2010). The effect of a scapular postural correction strategy on trapezius activity in patients with neck pain. Manual Therapy, 15(6), 562–566. [DOI] [PubMed] [Google Scholar]
  31. Yip CHT, Chiu TTW, & Poon ATK (2008). The relationship between head posture and severity and disability of patients with neck pain. Manual Therapy, 13(2), 148–154. [DOI] [PubMed] [Google Scholar]
  32. Yoo W-G, & Kim M-H (2010). Effect of different seat support characteristics on the neck and trunk muscles and forward head posture of visual display terminal workers. Work, 36(1), 3–8. [DOI] [PubMed] [Google Scholar]

RESOURCES