low-level exertions of the neck musculature: a study of ...jwilson/files/joines06jek.pdf ·...

Journal of Electromyography and Kinesiology 16 (2006) 485–497

www.elsevier.com/locate/jelekin

Low-level exertions of the neck musculature: A study ofresearch methods

Sharon M.B. Joines a,*, Carolyn M. Sommerich a, Gary A. Mirka a,James R. Wilson a, Samuel D. Moon b

a Department of Industrial Engineering, North Carolina State University, Campus Box 7906, Raleigh, NC 27695-7906, USAb Division of Occupational & Environmental Medicine, Department of Community and Family Medicine, Duke University Medical Center, Durham, NC, USA

Received 17 March 2004; received in revised form 14 September 2005; accepted 20 September 2005

Abstract

Musculoskeletal neck discomfort is prevalent in many occupations and has been the focus of much research employing surfaceelectromy ography (sEMG). Significant differences in experimental methods among researchers make comparisons across studies dif-ficult. The goal of the current research was to use empirical methods to answer specific methodological questions concerning use ofsEMG in evaluation of the neck extensor system. This was accomplished in two studies. In Experiment 1, ultrasound technology wasused to: (a) determine accessibility of m. splenius and semispinalis capitis with surface electrodes, (b) identify appropriate electrodelocations for these muscles/muscle groups, and (c) illustrate potential benefits of using ultrasound in locating muscles/placing elec-trodes. Experiment 2 sought to assess effects of posture when normalizing sEMG data. Results from Experiment 1 showed no directaccess to semispinalis capitis for surface electrodes; their activity can only be sampled as part of a group of muscles. In most sub-jects, m. splenius was found to be accessible to surface electrodes. Electrode placement recommendations are provided. Results ofExperiment 2 showed significant differences in normalized EMG data between a posture-specific technique and a reference posturetechnique. Posture-specific normalization is recommended for accurately assessing the relative intensity of contractions of thesemuscles.� 2005 Elsevier Ltd. All rights reserved.

Keywords: Electrode location; Normalization method; Muscle access; Neck muscles

1. Introduction

Musculoskeletal discomfort in the neck has been associ-ated with a variety of occupations and types of work, fromthose categorized as physically demanding, such as farmwork [39,46] and patient care [5,23] to those that are con-sidered fairly static, such as dental work [20,34,36], workat video display terminals [4,26], and sewing machine oper-ation [1,47]. In a group of Danish CAD operators, Jensenet al. [20] found a 70% 12-month prevalence of musculo-

1050-6411/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jelekin.2005.09.007

* Corresponding author. Tel.: +1 919 513 0825; fax: +1 919 515 7330.E-mail address: [email protected] (S.M.B. Joines).

skeletal symptoms in the neck, higher than any other bodyregion, including low back (54%), shoulders (54%), andhands/wrists (52%). Many types of tissue in the cervicalregion can be sites of pain origin, including the neck mus-cles, intervertebral discs, the posterior longitudinal liga-ment, and the facet joints [6]. Whereas heavy or morephysical work has been associated with diagnoses of cervi-cal spondylosis [15] and degenerative changes [54], staticwork is more often associated with muscle-related prob-lems, including tension neck syndrome (myofascial pain)or myalgia [13–15]. Static work concentrates the workloadon fewer, smaller muscle groups [50], which may be selec-tively overloaded through prolonged activation of someof the fibers in those muscles.

mailto:[email protected]

486 S.M.B. Joines et al. / Journal of Electromyography and Kinesiology 16 (2006) 485–497

Muscle activity in occupational tasks is often investi-gated using surface electromyographic (sEMG) recordings.Electromyography can provide insight into patterns ofmuscle activation, intensity of activation, and informationabout localized muscle fatigue. Under static conditions,sEMG provides a means of assessing the relative level oftension or force developed in a muscle. Electrodes attachedto the skin detect this activity. Just as a microphone picksup any signal in its vicinity, so does a surface electrode.Therefore, placement over the muscle of interest, and awayfrom other muscles is essential to obtaining a signal that isnot contaminated by cross-talk from neighboring muscles.This requirement poses a significant challenge to research-ers interested in studying the neck musculature, becauseapproximately 20 pairs of muscles, organized in several lay-ers, support, traverse and move the neck and skull [24].Neck muscle pain is most common in the posterior regionof the neck. Therefore, the superficial muscles that are posi-tioned for head and neck extension and support and thatmay be accessible via surface electrodes were of primaryinterest to this investigation. These muscles are the semispi-nalis capitis (primary neck extensor) and splenius capitis(extension, rotation, and lateral movement).

A review was conducted of the methods of 23 studies inwhich the primary neck extensors and nine studies in whichsplenius muscle activity was sampled using sEMG. Themethods review assessed several aspects of experimentationprotocol, including electrode positioning, method of data

Table 1Summary of inconsistencies in electrode positioning

# of studies reviewed Primary neck extensors

23

Electrode locations 5 distinct locations:� cervical paraspinal – C3–C6 [12],� neck extensor – C1–C4, where specified [16,3� semispinalis capitis – C1–C2 [25,37],� trapezius par descendens in the cervical regio

[40,51,52], and� cervical erector spinae – C2–C6 [3,9,18,2

45,53]; the location range in each bullet inthe studies sited chose different electrodwithin the named range

Inter-electrode distances 65% of studies reviewed failed to reportOf those reporting [3,12,16,18,25,29,31,37,41–4� 6 distances were reported,� ranging from 1 to 6 cm

Distance to midline Seven studies reported [9,12,16,25,29,31,37,48]� 4 distances were reported, ranging from 1 to

Electrode orientation 10 studies reported:� 5 aligned with the spine [3,12,16,28,48],� 5 aligned with the direction of the mu

[25,29,38,41,53],2 studies did not report [9,52]9 indicated in figure as along direction of mus[18,31,37,40,42–45,51]

normalization, EMG signal sampling rate, and EMG sig-nal processing. Choices made in each of these areas canhave significant effects on results of an EMG study[21,22], and this paper takes a close examination of the for-mer two items. The variation in data collection and pro-cessing techniques employed in previous studies makescomparisons of results difficult and can result in inconsis-tent findings among studies. Additionally, some EMG nor-malization methods require subjects to perform maximalexertions. These methods needed to be carefully assessedand justified in their application to collection of data forthe neck muscles, especially when collecting data fromolder subjects.

1.1. Electrode locations and muscle accessibility

The research methods review revealed a substantialdegree of inconsistency in muscle terminology and electrodepositioning (location, inter-electrode distances, and dis-tances from the midline). The position descriptions of themuscles indicated that the studies were referring to the samemuscles, yet sampling at different locations and depths ofthe muscles. These inconsistencies, summarized in Table1, make the comparison of results between studies difficult.

The primary neck extensor muscle described as ‘cervicalparaspinal muscle’, ‘neck extensor’, ‘semispinalis capitis’,‘trapezius par descendens in the cervical region’, or ‘cervi-cal erector spinae’ were examined in 23 studies that

Splenius muscle

9

3 distinct locations:

1,48,55,56],

n – C1–C6

5,29,38,41–dicates thate locations

� C2 [2],� C2–C3 [10,11,43–45],� C4 [25,37]; ranged from C2 to C4 levels

60% of studies reviewed failed to report3,48,51,53]: Of those reporting:

� 3 distances were reported [2,10,11,25,30,37],� ranging from 1 to 2.1 cm

:2.5 cm � 3.5–4.0 cm lateral to C2 spinous process [2],

� 6–8 cm lateral to midline [25,37],� no other studies provided this information

4 studies reported:

scle fibers� 3 aligned with the spine [2,10,11],� 1 aligned with the direction of the muscle fibers [25],

1 study did not report [30]cle fiber 4 indicated in figure as along direction of muscle fiber

[37,43�45]

S.M.B. Joines et al. / Journal of Electromyography and Kinesiology 16 (2006) 485–497 487

employed sEMG [3,9,12,16,18,25,28,29,31,37,38,40–45,48,51–53,55,56]. In four of the studies, no specific electrodelocations were provided [9,38]. The remaining studiesdocumented five distinct electrode locations: three studiesidentified the C1/C2 level, one at the C2 level, nine at theC2–C3 level, three between C2 and C6, and one at theC5/C6 level. The splenius muscles were examined in ninestudies [2,10,11,25,30,37,43–45]. One study [30] provideda general location, while the remaining studies identifiedthree distinct electrode locations.

Of primary significance was the assertion, by Keshneret al. [25], that the semispinalis capitis m. is accessible usingsurface electrodes. This claim was consistent with prelimin-ary cadaver work conducted as part of the current study, butwas inconsistent with references sampling similar locations,yet referring only to the less specific neck extensor musclegroup [16] and cervical erector spinae muscles [3,43–45].

An obvious question that stems from this diversity ofelectrode locations is whether or not there is an ideal loca-tion, and that would seem to depend, in part, on the extentto which the muscles of interest were accessible to surfaceelectrodes. This is one of the two primary points of focusof the original research study described later in this report.

1.2. Normalization methods

Among the studies relating to sEMG activity in the neck,there were 5 that performed no normalization [2,37,38,51], 4that normalized to a reference task or submaximal exertion[19,31,53,57], and 19 that normalized to maximum exertions(five studies reported two methods of normalization)[2,10,16,18,25,28,29,37,41–43,45,48,51,53,55–57]. In addi-tion to the method of normalization, posture during maxi-mal exertions has been shown to influence normalized data[35]. Posture-specific maximal or reference exertions affordnormalization of sEMG data to data collected from a simi-lar portion of the muscle. This is important to consider whenpostures change throughout a task. When this occurs, theportion of the muscle being sampled may change due to skinand electrode movement relative to the muscle of interest.Even so, many of the studies utilized postures that generatedmaximum EMG activity from each neck muscle [44], ratherthan postures specifically relevant to the task of interest.

EMG data are collected while muscles are activated toperform a task. To quantify the amount of muscle activitythe task requires relative to the muscle’s capability necessi-tates a benchmark for comparison. Typical benchmarks forcomparison include a muscle’s activity while resting, whileperforming a standardized task, while performing a refer-ence contraction, or while performing a maximum volun-tary exertion. The EMG data is then represented as apercentage of the reference value. This process is referredto as normalization of EMG data. The majority (19 of23) of the research studies that were reviewed, which usedsEMG data collected in the neck region, documented thenormalization technique as a percentage of a maximumvoluntary exertion (MVE) [2,10,16,18,25,28,29,37,41–

43,45,48,51,53,55–57]. In order to obtain reliable data ofthe muscle’s capability, accurate MVEs must be obtained.There are several factors that may affect the data collectedduring an MVE, including posture, fatigue, and subjectmotivation. The effect of posture on the normalization ofsEMG data collected from the neck extensor muscles wasthe second primary point of focus of the original researchstudy described later in this report.

Posture can affect the muscle’s ability to generate forcedue to changes in the length–tension relationship and mus-cle’s moment arm. Schuldt and Harms-Ringdahl [43] foundthat the peak activity levels for the cervical erector spinaeoccurred in a slightly flexed posture. This posture has beenadopted by several researchers when defining the postureduring their MVEs. Changes in joint posture may changethe portion of the muscle being sampled and may resultin additional muscles being sampled. The argument forminimizing error due to normalization technique wasaddressed in a study of eight trunk muscles under staticand dynamic conditions, conducted by Mirka [35]. Mirkafound that posture significantly affected the normalizedEMG data collected in the low back region. Althoughthe merits of Mirka’s multiple reference normalizationmethod are clear for moderate to high force exertions per-formed over a wide range of trunk angles, the benefits ofsuch a normalization approach have not resulted in suchnormalization of EMG data collected from the neck whennormalizing low force exertions. Potential reasons for thelack of transference of the multiple reference normalizationmethod presented by Mirka to EMG data collected fromthe neck include the reduced range of motion for typicalneck related tasks, and the concern for subject discomfortand risk of injury that could be associated with multiplemaximum exertions performed with the head and neck.The objective of this experiment was to determine if theposture in which the normalization or maximum exertionis performed has a significant impact on the results ofsEMG data collected from the neck extensor muscles.

It is important to note that there were several othersources of variability identified between the studies. Theyincluded differences in signal collection methods, includingsampling rate, bandwidth, smoothing, and filtering. Eightdifferent sampling rates (from 5.7 to 2000 Hz) were docu-mented by 12 of the studies [2,3,10,11,16,29,30,48,51,53,55,56]. The other studies [3,9,12,16,18,25,28,29,31,37,38,40–45,48,51–53,55,56] did not report this aspect of theirmethods. Nine studies documenting bandwidth chose dif-ferent pairs of cut-off frequencies ranging from 2 to39 Hz for the low frequency cutoff and from 100 to2000 Hz for the high frequency cutoff [2,3,10,11,30,51,53,55,56]. Five different sizes (ranging from 0.1 to0.5 s) were documented for the width of the smoothingwindows used in 10 studies [2,17,27,29,40,42,45,48,53,56].Only one study explicitly reported filtering data to remove(50 Hz due to European power characteristics) noise [11].These issues are not discussed further here, but were exam-ined more closely by Joines [21]. The focus of this report is

Fig. 1. Subject’s neck marked using structured neck marking procedurefor m. splenius electrode locations.


restricted to the important questions regarding appropriatesurface electrode location and normalization procedures.

The two part study described in the remaining sections ofthis report was designed to address the following questions:

(1) Are these primary neck extensor muscles (spleniuscapitis and semispinalis capitis) accessible using sur-face electromyography? Where should surface elec-trodes be located to provide the best access to thesemuscles? Is palpation sufficient to identify the bestlocation, or would use of modern imaging techniquesidentify better locations? (Experiment 1.)

(2) Does the posture of the neck that is assumed during amaximum exertion is performed have a significanteffect on the normalized EMG magnitude? (Experi-ment 2.)

2. Methods

This paper provides descriptions of two experimentsspecifically conducted to examine electrode location anddata normalization methods, in the context of collectingsEMG data from the neck extensor musculature under sta-tic/semi-static loading conditions. The methods used forcollecting data in both experiments are presented sequen-tially in this section.

2.1. Experiment: muscle access for surface

electromyography

Experiment 1 consisted of identifying electrode locationsfor the left and right SPL and left and right SEMI using astructured neck marking procedure that included tradi-tional palpation methods and then evaluating the qualityof those electrode placement locations using ultrasoundtechnology.

2.1.1. Subjects

Twenty-one subjects from North Carolina State Univer-sity and the surrounding area were recruited for participa-tion in this study. Subjects ranged in age from 21 to 75years of age and consisted of 8 females and 13 males.

2.1.2. Apparatus

Ultrasound imaging was used in this study to examinethe arrangement of the subjects’ muscles. Use of ultra-sound poses minimal risks to subjects, and scans are rela-tively easy to interpret with a moderate amount oftraining. The use of ultrasound equipment in this researchwas provided by Duke University Medical Center(DUMC). Although use of ultrasound equipment is fairlyintuitive due to the dynamic, real-time nature of the tech-nology, it was necessary to obtain hands-on experience inorder to develop the dexterity and coordination requiredto identify the neck muscles. This experience was gainedduring intensive pilot work conducted at DUMC.

The ATL HDI 5000 ultrasound system with a CL 10-5transducer (frequency 12 MHz) and the musculoskeletalprogram option was chosen based on pilot work. Ultra-sound technology generates gray scale images from theresponse of soundwaves reflecting off tissues. Muscle tissuesappear dark with light outlines formed by fascia. The darkmuscle area and light outline change shape (typically appearslightly more cylindrical) when the muscle is contracted.Myotendonous junctions gradually brighten in the centerof the fascia outline as the tissue transitions from muscleto tendon. Tendons appear bright. Adipose (fat) tissueappears speckled gray without defined boundaries. In addi-tion to providing high quality images that are easily inter-preted, use of ultrasound allowed subjects to assume aseated posture, typical of a working environment.

2.1.3. Protocol

After the experiment was described, subjects were askedto read and sign an informed consent form. Anthropomet-ric measurements were made. Posterior regions of the sub-ject’s neck were shaved where needed. The sites for theelectrode pairs for the SEMI and SPL were located usingtraditional muscle location methods (palpation during acti-vation) augmented by a structured neck marking procedurefor both the left and right side.

The steps for identifying the location of the electrodelocations for the m. splenius were as follows:

1. With the subject’s head in a fully flexed posture, a linewas drawn from the top of the ear to C7 (line 1 inFig. 1).

2. With the subject’s head in an upright posture, the pos-terior edge of the sternocleidomastoid was palpated dur-ing a resisted flexion exertion.


3. After the subject relaxed, the posterior edge of the ster-nocleidomastoid was marked (line 2 in Fig. 1).

4. With the subject’s head in an upright posture, the bellyof the splenius was palpated. The palpation was per-formed while providing resistance superior to the sub-ject’s temple while subject rotated his/her head andneck.

5. After the subject relaxed, the line of action of the sple-nius was marked at the belly of the muscle.

6. The electrode location (or center of the electrode pair)was chosen at the intersection of the C7-Ear line andthe splenius line of action muscle.

7. If the electrode location was estimated to overlap theposterior edge of the sternocleidomastoid, a locationslightly inferior and posterior was chosen until it nolonger overlapped with the sternocleidomastoid.

The sites for the electrode pair for the m. semispinaliscapitis were located using the following procedure:

1. While the subject sat in an upright posture, the experi-menter supported the weight of the subject’s head withboth hands and palpated the occiput.

2. After the subject resumed support of his or her head, thelocation of the occiput was marked (line 1 in Fig. 2).

3. This was repeated for locating and marking the spinousprocess of the C2 (line 2 in Fig. 2).

4. The belly of each neck extensor muscle was identifiedduring several resisted neck extensions (line 3 in Fig. 2).

5. The electrode location (or center of the electrode pair)was marked at the intersection of the line marking thebelly of the muscle and the C2 level.

Fig. 2. Subject’s neck marked using structured neck marking procedurefor m. semispinalis capitis electrode locations.

The muscle edges, lines of action and electrode locationswere marked on the subject using a temporary, non-toxicmarker (no electrodes were actually placed on the subjectsfor Experiment 1, because they would have blocked accessto the skin that was required for the ultrasound scanning).The electrode locations described above are referred to aselectrode level (E-level), meaning the most likely spot toselect for electrode placement, based on palpation.

While the subject’s head was in a relaxed, neutral pos-ture, a minimum of three scans were taken for each of thefour muscles (left and right for SEMI and SPL). For eachmuscle, the first scan was taken at the electrode placementsite (E-level). The second scan was taken 1 cm superior toE-level; the third was taken 1 cm inferior to E-level. Thesubject was also scanned at E-level with the neck in arelaxed, flexed posture as if reading a book placed on his/her lap. The ultrasound scanning session lasted approxi-mately 1.5 h.

2.1.4. Data analysis

The ultrasound scans were printed using a high resolu-tion laser printer, and the muscle accessibility at the variouselectrode locations was examined. Each scan represented a2.4 cm deep section of the tissue under the scan head. Thescans were examined to determine (a) if the muscle wasaccessible to surface electrodes, which meant that othermuscle(s) did not come between the muscle of interestand the skin at the location of interest, and (b) if the musclewas accessible, then the quality of the proposed electrodelocation was rated. The sternocleidomastoid and trapeziusm. are superficial to SEMI and SPL. This can be seen inFig. 3. Figs. 3(a)–(c) illustrate three representations of thesame muscle arrangement. Fig. 3(a) provides a transverseview highlighting the location of the surface electrode onthe skin surface over the left SPL. Fig. 3(c) presents anultrasound scan taken at the location depicted in Fig. 3a.Fig. 3(b) is a graphic identifying the muscles in the ultra-sound scan. Fig. 3(b) shows the SPL to be accessible, butonly to an electrode located at a particular location. Thatlocation is referred to here as the ‘gap of accessibility’,and is the gap between the edges of the sternocleidomastoid(SCM) and the trapezius (TRAP). The gap is also shownon the scan in Fig. 3(c). If the proposed electrode place-ment was found to be over the gap of accessibility, it wasrated as either A, B, or C, according to how centrallylocated the proposed electrode placement was relative tothe edges of the gap of accessibility (see Fig. 4 for exam-ples). If there was no gap at the electrode location, thethicknesses of the muscles were measured. The thicknessof the overlapping muscle (TRAP or SCM) was measured,as was the thickness of the muscle of interest (SPL orSEMI). These measures were used to calculate the percent-age of muscle mass underneath the ultrasound scan head(i.e., the location of the electrode) that was overlappingmuscle. Additionally, the data were examined as a func-tion of sex and subject anthropometry, in an effort toidentify a means of predicting presence and/or quality of

Fig. 3. Muscle accessibility determination utilizing ultrasound images: (a) cross-sectional view of neck; (b) drawing identifying muscles in scan below;(c) ultrasound scan of left splenius (SPL), E-level in the upright posture.


accessibility. Scan measurements of muscle thicknesses anddepths are documented in more detail in [22].

2.2. Experiment 2: utility of posture specific normalization

The primary factor of interest in this experiment was theeffect the data from a maximum exertion performed in aparticular posture has on EMG task data to which the taskdata are normalized. As such, in this study maximum exer-tions were performed in two postures: neck straight (headupright, as would be adopted when looking across a room)and neck flexed about 25� (as if reading a book or viewinga computer on one’s lap). Two static tasks were performed:head/neck extension at 5% MVE and at 10% MVE. Thesewere performed in the same two postures as the maximumexertions. The EMG data from the static tasks were nor-malized two ways: relative to the MVE in the same postureas the task and relative to the MVE recorded in the otherposture. The dependent variables were the normalized

EMG values from the electrode locations identified inExperiment 1 (bilaterally, over m. splenius and over m.semispinalis capitis; the latter is hereafter referred to inExperiment 2 as the neck extensor (NE) group, for reasonsthat will be made clear later in the report).

2.2.1. Subjects

Ten subjects (seven males and three females) from theuniversity population and surrounding areas were recruitedto participate in the experiment. Each provided informedconsent prior to participation. Subjects ranged in age from25 to 54 years and participated in one practice session andone test session. During the practice session, each subjectwas screened by questionnaire regarding health history.

2.2.2. Apparatus

This experiment was performed using a Kin-Com 125EIsokinetic Dynamometer (Chattecx Co., Chattanooga,TN) and an EMG data collection and processing system

Fig. 4. Examples of employment of rating method for proposed electrode locations. The rating grades the proposed location of the electrodes with respectto the centerline of the muscle’s gap of accessibility.


to collect muscle activity data. While the subject was per-forming MVEs, the subject was seated in a chair that waslocated in one of three positions such that the cushionedpad on the Kin-Com arm was located either to the left,right, or back of the subject’s head. The subject’s torsoand upper arms were restrained to the chair with a velcrostrap to isolate the exertion to just the subject’s neck.The subject was also asked to rest his/her heels on blocksto minimize any ability or tendency to push with the legs.The adjustable arm and head on the Kin-Com were usedto position the cushioned pad attached to the load cell rel-ative to the subject while the subject was in the neutral,upright seated position (see Fig. 5).

Subjects were videotaped from the side during the exper-iment in order to document and ensure correct head/neckposture throughout the experiment. The center of the cam-era lens was set to the subject’s seated eye height. Reflectivemarkers were placed on the subject’s canthus, tragus, and C7locations to identify the bony landmarks. Digital pictures ofthe subject were also taken during maximum and submaxi-mum exertions in the upright and flexed neck postures.

The EMG data collection system consisted of Ag-Ag/Clbipolar surface electrodes (Model E22x, InVivo Metric,CA), preamplifiers (1000·), main amplifier (52.2·), hard-ware filters (60 Hz notch filter, 1000 Hz low pass filter),and an A/D system connected to a Pentium 90 MHz datacollection computer. Electrodes with an inner diameter of4 mm were placed with a center-to-center distance of1.5 cm. The electrodes were placed over the splenius and

the neck extensor muscles along each muscle’s line-of-action, using procedures described in Experiment I. Thecenter of the electrode pair for the SPL was located at theintersection of the C7-Ear line and the line of action forthe SPL posterior to the sternocleidomastoid. The centerof the electrode pair for the NE was located over the bellyof the neck extensor muscles at the C2 level. Prior to apply-ing the electrodes, the subject’s neck was shaved, includingthe region above the hair line on the neck, if necessary, inorder to expose the skin at the level of C2. The skin was thencleaned and abraded, and then the electrodes were applied.

The four unprocessed EMG signals (bilateral collectionfrom the two muscle locations) were collected along withthe Kin-Com force data. All data were collected at1024 Hz. The analogue signals were converted to digitalsignals and recorded using the data collection programGlobal Lab (Data Translation, Inc. V 3.00, Marlboro,MA). The signal was subsequently digitally filtered (highpass: 10 Hz; low pass: 512 Hz; notch: 60 and 180 Hz), rec-tified, and smoothed (0.1 s moving window).

2.2.3. TasksThe tasks performed during this experiment consisted of

a set of low-level neck extension exertions and a set of max-imum exertions. The low-level exertions were performed at5% and 10% of MVE in both upright and neck flexed pos-tures. Each was performed twice. Subjects also performeda series of MVEs, which provided the reference values fornormalization of the low-level, sub-maximal exertions.

Fig. 5. Subject performing task in flexed neck and upright postures in Experiment 2.


The method for performing MVEs and collecting MVE datais described below. The subjects were provided a practicesession during which they learned to perform the tasks. Thiswas followed, at least 24 h later, by the actual test session.

2.2.4. Practice session

During the practice session, each subject was screenedfor the following health conditions: a history of or exis-tence of current neck-related diagnoses or symptomsincluding injury or surgery; a history of dizziness, fainting,or symptomatic irregular heartbeat; diagnosed or currentsymptoms of coronary artery disease; very high risk ofcarotid or coronary artery disease; poorly or marginallycontrolled high blood pressure; severe limitations in pul-monary function capacity; or a heart condition. If thesubject’s health history was free from these conditions,the experiment was described in detail and then the subjectwas asked to read and sign a consent form.

After a series of anthropometric measurements weretaken, the subject was guided through a set of moderatestretching exercises. The subject was then taught and prac-ticed the maximum voluntary exertions. The subject per-formed the isometric MVEs with an upright head andwith the neck flexed forward about 25�, by pushing his/her head against a cushioned pad in three directions (exten-sion, right lateral extension, left lateral extension) in sepa-rate trials. Each maximum exertion lasted for 6 s, duringwhich time subjects received verbal encouragement. Thesubject was instructed to build up his/her effort for the firsttwo seconds, then exert as forcefully as possible for four sec-onds. Two minutes rest was provided between exertions.Two repetitions of each maximum exertion were performed,

with a third performed if the first two were not within 10%of each other. The subjects then practiced the low-level exer-tions, during which they were required to exert 5% or 10%of their maximum extension exertion for 5 s, in both theupright and flexed neck postures. Visual feedback for thelow-level exertions was provided on a computer screen thatdisplayed a target line bracketed by two lines indicating thetolerance range, and the subject’s extension force registeredby the Kin-Com. The practice session lasted about 1 h.

2.2.5. Test session

The test session was conducted in a similar manner tothe practice session, except that electrodes and reflectivemarkers were placed on the subject after the stretchingand prior to performing the maximum exertions. The testsession lasted approximately 1.5 h.

2.2.6. EMG processing and normalizationFollowing initial processing (filtering, rectification, and

smoothing), the EMG data were reduced. For each muscleand each posture, a maximum smoothed MVE value wasidentified to be used in the normalization process. Forthe low-level exertions, an average EMG level was calcu-lated from each exertion, with the first and last secondsof data excluded from each file. Then average EMG valuesfrom each of the four submaximal tasks were normalizedusing two different techniques.

The first method of normalization, referred to as pos-

ture-specific normalization, is defined in Eq. (1). Using thismethod, the task data were normalized to the maximumcollected in the same posture as the task. A priori, thiswas considered to be the gold standard, based on the workby Mirka [35]:


NormalizedEMGi;j;k ¼EMGði; j; kÞ

MaxEMGði; j; kÞ ; ð1Þ

where EMG(i, j,k) is the task average EMG signal at neckangle i, from muscle j, for subject k; MaxEMG(i, j,k) is themaximum EMG value for neck angle i, from muscle j, forsubject k; i = upright or flexed; j = SPL or NE; and k = 1to 10.

The second method of normalization, referred to as ref-

erence-posture normalization, is defined in Eq. (2). Usingthis method, the task data were normalized to the maxi-mum collected in a standard reference posture. Both theupright and flexed postures were investigated as standardreference postures:

NormalizedEMGi;j;k ¼EMGði; j; kÞ

MaxEMGðr; j; kÞ ; ð2Þ

where EMG(i, j,k) is the task average EMG signal at neckangle i, from muscle j, for subject k; i = upright or flexed;j = SPL or NE; k = 1 to 10; MaxEMG(r, j,k) is the maxi-mum EMG value with the head and neck in a standard ref-erence posture r, from muscle j, for subject k; r = flexed ifi = upright; r = upright if i = flexed.

2.2.7. Data and statistical analyses

The analysis of the data consisted of calculating the dif-ferences between the results from processing the data usingthe two different normalization techniques. The differencedata was first evaluated using the Shapiro-Wilk test fornormality. Data sets that failed the Shapiro-Wilk test fornormality were evaluated using a Wilcoxon signed-ranktest. A paired t-test was used on the remaining data todetermine if the mean difference between the data setswas different from zero. No statistical difference from zerowould indicate that there was no difference between thenormalization methods.

3. Results and discussion

3.1. Experiment 1

Unilaterally or bilaterally, SEMI was accessible in onlytwo subjects, while SPL was accessible in 20 of the 21 sub-jects. For the electrode locations identified for the left andright SPL, 52% and 71%, respectively, were positioned overa gap of accessibility at E-level. Logistic regression wasemployed to test the predictability of accessibility as a func-tion of anthropometry. Results indicated that none of theanthropometric measurements or combinations of the 14measurements collected would predict the presence of gapsover SPL.

3.1.1. Accessibility of m. semispinalis capitisThe measurements from the ultrasound scans indicated

that 19 of 21 subjects were without a gap of accessibilityfor SEMI (either unilaterally or bilaterally). For the leftand right SEMIs, 0% and 9% of the electrode locations

at E-level were positioned over gaps, respectively. Thesefindings stand in contrast to the assertions of Keshneret al. [25], who reported that SEMI was accessible at C1–C2, 2 cm to the midline. Those authors stated, ‘Anatomicalexamination reveals that most of SEMI lies close to midlineand can be palpated by resisting head extension. The mus-cle is covered by either SPL or TRAP except for a smallarea extending from its insertion to about the level of thesecond cervical vertebrae. . .’ In the current study, approx-imately 30% of the muscle thickness under the electrodelocation was muscle that overlapped SEMI. That percent-age decreased to 19–25% for the superior electrode location(E-level +1) and increased to about 39% for the inferiorelectrode location (E-level �1). When comparing the flexedto the upright posture at E-level, the percent of muscleoverlap was similar, as was the extent of inaccessibilityacross subjects. Given these findings, in concurrence withthe recommendation from Sommerich et al. [49], the loca-tion identified for the SEMI in this study should be consid-ered as a location-specific site, rather than a muscle-specificsite. Muscle activity from this site should be designated asemanating from the neck extensor (NE) muscle group,rather than a specific muscle.

3.1.2. Accessibility of m. splenius

In this investigation, 15 of 21 subjects had gaps of acces-sibility at E-level for the right SPL (Gap Rating A: 7 sub-jects; B: 4; and C: 4) and 11 of 21 had gaps for the left SPL(Gap Rating A: 7; B: 2; and C: 2). For the right and leftSPL, the number of gaps was maximized at E-level(n = 15) and E-level +1 (n = 14), respectively. When com-paring the flexed to the upright posture, the total numberof gaps decreased for the right SPL (from 15 to 14) andincreased for the left (from 11 to 13). The average thicknessand depth of the SPL decreased in the flexed posture forboth males and females.

The assessment of Keshner et al. [25], regarding SPLaccessibility, is consistent with Mayoux-Benhamou et al.[32] who reported finding a small superficial area of theSPL using CT scans. However, Mayoux-Benhamou et al.concluded from their study that the surface electrodes theypositioned over m. splenius picked up activity from thesplenius as well as the SCM during certain exertions, andso they recommended using intramuscular electrodes tostudy m. splenius in certain circumstances: dynamic testsand neurophysiological studies of motor control of thehead-neck system. They suggested that for ergonomic stud-ies in which the dorsal neck musculature as a whole was ofinterest, surface electrodes could be appropriate. Theirmaximum isometric tests showed SPL activity from thesurface and intramuscular electrodes were similar for thetests that were relevant to the current study (isometricextension and, less directly, ipsilateral rotation). Together,these findings suggest that for assessment during isometricextension, as in the current study, cross-talk from the SCMis not likely to be a concern for the SPL electrode locationidentified in the current study. Though not utilized in the


current study, a double-differential electrode configurationmay reduce cross-talk in other conditions [7,8].

3.2. Experiment 2

Individual differences among subjects, between normal-ization methods, ranged from 0% to 6% MVE for the SPLand from 0% to 24.7% MVE for the NE site. Results of sta-tistical testing of the effects of normalization method areprovided in Table 2. For the data collected in the flexedposture during submaximal exertions, the differencesbetween the data sets when normalized to data collectedin the upright and flexed posture were significantly different(at both 5% and 10%) for the right SPL, and left and rightNE. No significant difference was found between data col-lected from the left SPL regardless of exertion level, for thedata collected in the flexed posture. For the data collectedin the upright posture during submaximal exertions, thedifferences between the data sets when normalized to datacollected in the upright and flexed posture were significantfor the right NE (at both 5% and 10%) and for the left NE(at 10% and nearly significant at 5%). There were no statis-tically significant differences between the data sets for theleft and right SPL based on the posture during the maxi-mum exertion for either exertion level for the data collectedin the upright posture.

To summarize, for data collected in a flexed neck pos-ture during submaximal exertions, except for the leftSPL, there was a statistically significant difference betweenthe neck sEMG data normalized using the posture specificnormalization method and the reference posture normali-zation method. For data collected in an upright postureduring submaximal exertions, there was a statistically sig-

Table 2The effect of neck posture during MVE on normalization of EMG activity du

Submaximal exertion: 5% of MVE

MVE neck posture Diff. between upright and

Upright Flexed

Submax – flexed neck postureLeft extensor 0.19b 0.15c P = 0.006*d

Right extensor 0.17 0.14 P = 0.024*d

Left splenius 0.08 0.07 P = 0.856Right splenius 0.07 0.06 P = 0.007*

Submax – upright neck postureLeft extensor 0.15e 0.13f P = 0.056Right extensor 0.13 0.12 P = 0.0001*Left splenius 0.07 0.07 P = 0.75Right splenius 0.06 0.06 P = 0.11

Note: Values in bold font indicate the posture-specific normalized values. Pair-wactivity normalized to the upright MVE and normalized to the flexed MVEmethods produced difference normalized EMG values. All NIEMG data presen(n = 2).

a Differences between upright and flexed groups were evaluated using the Wb NIEMG = (EMG during subMax exertion at 5% of MVE in flexed posturc NIEMG = (EMG during subMax exertion at 5% of MVE in flexed posturd t Test used to evaluate difference between groups.e NIEMG = (EMG during subMax exertion at 5% of MVE in upright postuf NIEMG = (EMG during subMax exertion at 5% of MVE in upright postu

nificant difference between the neck extensor group sEMGdata normalized using the two normalization methods, andno difference for the splenius. Prudence would suggest thatthe posture specific method of normalization for data col-lected in the cervical region be employed to afford compar-isons of the current research with future efforts.

4. Conclusions and recommendations

Mirka [35] outlined several goals researchers have forusing electromyography in which they ‘(1) use EMG signalsto find the on/off points for a muscle, (2) use EMG to quan-tify the absolute electrical activity of a muscle, (3) use EMGto quantify the relative activity of the muscle, and (4) use theEMG signal to quantify the actual force the muscles areexerting across a joint’. In addition to these goals, research-ers may also wish to compare muscle activity across studies.Comparisons between experiments (meta-analysis and vali-dation experimentation) are facilitated when similar meth-ods are used, and hampered or precluded when they arenot. Both of the experiments in this paper were conductedin order to draw attention to the significant differences inmethods utilized by experimentalists who study the neckmusculature, as well as some ways in which those method-ological differences may impair the ability to compareresults from one study to another, or build upon prior work.

4.1. Experiment 1

The semispinalis capitis was found to be, essentially,inaccessible to surface electrodes. The splenius was foundto be easily locatable when traditional palpation methodswere augmented by a structured neck marking procedure.

ring low force exertions

Submaximal exertion: 10% of MVE

flexeda MVE neck posture Diff. between upright and flexeda

Upright Flexed

0.21 0.18 P = 0.004*0.20 0.17 P = 0.006*d

0.10 0.10 P = 0.990.09 0.07 P = 0.007*

0.18 0.15 P = 0.01*d

0.15 0.14 P = 0.002*d

0.08 0.08 P = 0.8790.07 0.06 P = 0.09

ise statistical testing was conducted, such that the difference between EMGwas compared to the value 0. P values <0.05 indicate the normalizationted in this table have been averaged across subjects (n = 10) and repetitions

ilcoxon signed rank test unless other wise noted.e)/(EMG during MVE in an upright posture).e)/(EMG during MVE in an flexed posture).

re)/(EMG during MVE in an upright posture).re)/(EMG during MVE in an flexed posture).

Tab

le3

Rec

om

men

dat

ion

sfo

rd

ata

coll

ecti

on

par

amet

ers

of

SE

MG

inth

en

eck

mu

scu

latu

red

uri

ng

low

-lev

el,

stat

icex

erti

on

s

Par

amet

erR

eco

mm

end

atio

n

Bas

edo

nth

efi

nd

ing

pre

sen

ted

her

e

Ele

ctro

de

loca

tio

n:

nec

kex

ten

sor

(in

clu

des

sem

isp

inal

isca

pit

is.

man

dtr

apez

ius

m.)

Th

esi

tes

for

the

elec

tro

de

pai

rsfo

rth

eS

EM

Ian

dS

PL

wer

elo

cate

du

sin

gtr

adit

ion

alm

usc

lelo

cati

on

met

ho

ds

(pal

pat

ion

du

rin

gac

tiva

tio

n)

augm

ente

db

ya

stru

ctu

red

nec

km

ark

ing

pro

ced

ure

for

bo

thth

ele

ftan

dri

ght

sid

e(s

eeS

ecti

on

2.1.

3)

Th

eel

ectr

od

ep

airs

sho

uld

be

cen

tere

dar

ou

nd

C2

ove

rth

eb

elly

of

the

mu

scle

(sem

isp

inal

isca

pit

is)

Ele

ctro

de

loca

tio

n:

Sp

len

ius

m.

Th

ece

nte

ro

fth

eel

ectr

od

ep

air

sho

uld

be

loca

ted

ove

rth

esp

len

ius

m.

wh

ere

the

lin

eb

etw

een

C7

and

the

top

of

the

ear

cro

sses

the

sple

niu

sm

.(p

ost

erio

rto

the

ster

no

clei

do

mas

toid

)N

orm

aliz

atio

nP

ost

ure

spec

ific

no

rmal

izat

ion

isre

com

men

ded

Bas

edo

nJo

ines

[21]

fin

din

gsS

amp

lin

gra

te10

24H

z(f

or

sign

alp

roce

ssin

gu

sin

gF

FT

s)

Ban

dw

idth

10–5

12H

zA

dd

itio

nal

filt

ers

60H

zan

dal

iase

s(d

ata

coll

ecte

din

the

US

)S

amp

lin

gd

ura

tio

nF

req

uen

cyd

om

ain

asse

ssm

ents

:�

1–2

sfo

rgl

ob

alp

aram

eter

s(e

.g.,

mea

nan

dm

edia

nfr

equ

ency

)�

5s

for

no

n-g

lob

alp

aram

eter

s(e

.g.,

po

wer

asso

ciat

edw

ith

smal

lra

nge

of

freq

uen

cies

)T

ime

do

mai

nas

sess

men

ts:

�le

ngt

ho

fw

ork

cycl

e(e

.g.,

gap

anal

ysis

)�

rep

rese

nta

tive

sam

ple

per

iod

s(e

.g.,

NIE

MG

or

avg,

amp

litu

de)


Splenius was accessible in most subjects, but access was notpredictable, based on anthropometric measures collected inthis study. Changes in neck posture did not result in signif-icant differences in accessibility of the semispinalis capitisor splenius. Although no strict rules can be outlined forEMG surface electrode placement based on the subjects’anthropometric dimensions or sex, surface electrodes maybe placed by relying on muscle palpation, bony landmarks,and a structured neck marking procedure.

The center of the electrode pair for the splenius should belocated at the intersection of the C7-Ear line and the line ofaction for the splenius muscle posterior to the sternocleido-mastoid. It is recommended that electrode pairs for semispi-nalis capitis be centered around C2 over the belly of themuscle, which is similar to the locations documented byHamilton [16], Mathiassen and Winkel [31], Keshner et al.[25], and Queisser et al. [37], and that the muscle mass sam-pled be referred to as the neck extensor muscle group, ratherthan a single muscle (semispinalis capitis). This location ispreferred, because the ratio of overlapping muscle to SEMIincreases substantially at the inferior location that was stud-ied (E-level �1). The use of ultrasound technology was use-ful in evaluating the underlying muscle arrangement of livesubjects in positions typical of working postures. It allowedfor verification of access to splenius and for evaluation ofthe electrode locations. Use of ultrasound for identifyingelectrode locations could be helpful in improving spleniusplacement, from ‘B’ and ‘C’ rated locations to ‘A’ positions.

4.2. Experiment 2

For the researcher, normalizing data to a single uprightposture makes conducting an experiment less complicated.For the subjects in an experiment with multiple neck pos-tures, normalizing data to a single upright posture mayresult in fewer MVEs, thereby making the experiment lesstaxing. However, based on the results of this study, if sEMGdata is normalized to data collected in a posture that is fairlydifferent from the posture in which the task was performed,accuracy may be sacrificed on the behalf of convenience.The EMG levels for the neck muscles associated with typicalworking tasks involving static loading of the neck are rela-tively low. The changes in accuracy associated with collect-ing MVCs in the posture of interest would appear worth theadditional inconvenience for the accuracy gained; therefore,posture specific normalization is recommended.

4.3. Summary

A review by the authors of the literature related to the useof surface electromyography on the neck musculaturerevealed a wide variation in methodologies including elec-trode location and normalization techniques. The experi-ments described in this paper were conducted to shed lighton these methodological issues. The results of these empiri-cal studies have illustrated the importance of posture-specific normalization and have documented the challenges


in isolating the semispinalis capitis and recommend a moregeneral description of neck extensor group. It is believedthat these recommendations can be used in conjunction withthose identified in previous investigations [21] (see Table 3)and those provided by the SENIAM project funded by theEuropean commission [33] to improve the standardizationof data collection methods which will facilitate meta-analy-sis across studies.

Acknowledgments

The assistance of Mr. Kevin Fiest and Dr. Mark Kleiverof Duke University Medical Center is greatly appreciated.

References

[1] Andersen JH, Gaardboe O. Prevalence of persistent neck and upperlimb pain in a historical cohort of sewing machine operators. Am JInd Med 1993;24(6):677–87.

[2] Bansevicius D, Westgaard R, Jensen C. Mental stress of longduration: emg activity, perceived tension, fatigue, and pain develop-ment in pain-free subjects. Headache 1997;37(8):499–510.

[3] Bauer W, Wittig T. Influence of screen and copy holder positions onhead posture, muscle activity and user judgment. Appl Ergon1998;29(3):185–92.

[4] Bernard B, Sauter S, Fine L, Petersen M, Hales T. Job task andpsychosocial risk factors for work-related musculoskeletal disordersamong newspaper employees. Scand J Work Environ Health1994;20(6):417–26.

[5] Bork BE, Cook TM, Rosecrance JC, Engelhardt KA, Thomason ME,Wauford IJ, et al. Work-related musculoskeletal disorders amongphysical therapists. Phys Ther 1996;76(8):827–35.

[6] Cailliet R. Neck and arm pain. 3rd ed. Philadelphia: FA DavisCompany; 1991.

[7] De Luca CJ. The use of surface electromyography in biomechanics. JAppl Biomech 1997;13:135–63.

[8] Farina D, Merletti R, Indino B, Nazzaro M, Pozzo M. Surface emgcrosstalk between knee extensor muscles: experimental and modelresults. Muscle Nerve 2002;26(5):681–95.

[9] Fernstrom E, Ericson M, Malker H. Electromyographic activityduring typewriter and keyboard use. Ergonomics 1994;37(3):477–84.

[10] Finsen L. Biomechanical aspects of occupational neck posturesduring dental work. Int J Ind Ergon 1999;23(5–6):397–406.

[11] Finsen L, Christensen H, Bakke M. Musculoskeletal disorders amongdentists and variation in dental work. Appl Ergon 1997;29(2):119–25.

[12] Gogia P, Sabbahi M. Median frequency of the myoelectric signal incervical paraspinal muscles. Arch Phys Med Rehabil 1990;71(6):408–14.

[13] Grieco A, Molteni G, De Vito G, Sias N. Epidemiology ofmusculoskeletal disorders due to biomechanical overload. Ergonom-ics 1998;41(9):1253–60.

[14] Hagberg M, Silverstein B, Wells R, Smith M, Hendrick H, CarayonP, et al. Work related musculoskeletal disorders (wmsds): a referencebook for prevention. London: Taylor & Francis; 1995.

[15] Hagberg M, Wegman D. Prevalence rates and odds ratios ofshoulder-neck diseases in different occupational groups. Br J IndMed 1987;44(9):602–10.

[16] Hamilton N. Source document position as it affects head position andneck muscle tension. J Appl Physiol 1996;39(4):593–610.

[17] Harms-Ringdahl K. On assessment of shoulder exercise and load-elicited pain in the cervical spine. Biomechanical analysis of load–EMG–methodological studies of pain provoked by extreme position.Scand J Rehabil Med Suppl 1986;14:1–40.

[18] Harms-Ringdahl K, Ekholm J, Schuldt K, Nemeth G, Arborelius U.Load moments and myoelectric activity when the cervical spine isheld in full flexion and extension. Ergonomics 1986;29(12):1539–52.

[19] Hautekiet M, Hermans V, Spaepen A, Snijders K. The influences ofthe height of the vdu screen, and the position of the back rest on theneck muscle load and the position of the head. In: Proceedings ofIEA, Tampere, Finland; 1997.

[20] Jensen C, Borg V, Finsen L, Hansen K, Juul-Kristensen B, Chris-tensen H. Job demands, muscle activity and musculoskeletal symp-toms in relation to work with the computer mouse. Scand J WorkEnviron Health 1998;24(5):418–24.

[21] Joines S. Using surface electromyography to study cervical extensormuscle activity: an investigation of methodological considerations andthe effects of age on fatigue development and recovery. Dissertation,Industrial Engineering. NorthCarolina State University,Raleigh; 2002.

[22] Joines S, Sommerich, CM. Low-level exertions of the neck musculature:a study of research methods, fatigue development, and effects of age.ReportNumber:02-02,NorthCarolinaState University,Raleigh;2002.

[23] Josephson M, Lagerstrom M, Hagberg M, Wigaeus Hjelm E.Musculoskeletal symptoms and job strain among nursing personnel:a study over a three year period. Occup Environ Med 1997;54(9):681–5.

[24] Kamibayashi LK, Richmond FJ. Morphometry of human neckmuscles. Spine 1998;23(12):1314–23.

[25] Keshner E, Campbell D, Katz R, Peterson B. Neck muscle activationpatterns in humans during isometric stabilization. Exp Brain Res1989;75:335–44.

[26] KnaveBG,WibomRI,VossM,HedstromLD,BergqvistUO.Workwithvideo display terminals among office employees. I. Subjective symp-toms and discomfort. Scand J Work Environ Health 1985;11(6):457–66.

[27] KumarS.Acomputerdeskforbifocal lenswearers,withspecialemphasison selected telecommunication tasks. Ergonomics 1994;37(10):1669–78.

[28] Kumar S, Scaife W. A precision task, posture, and strain. J Safety Res1979;11(1):28–36.

[29] Lannersten L, Harms-Ringdahl K. Neck and shoulder muscle activityduring work with different cash register systems. Ergonomics1990;33(1):49–65.

[30] Lie I, Watten R. Oculomotor factors in the aetiology of occupationalcervicobrachial diseases (ocd). Eur J Appl Physiol 1987;56(2):151–6.

[31] Mathiassen S, Winkel J. Electromyographic activity in the shoulder-neck region according to arm position and glenohumeral torque. EurJ Appl Physiol 1990;61(5-6):370–9.

[32] Mayoux-Benhamou MA, Revel M, Vallee C. Surface electrodes arenot appropriate to record selective myoelectric activity of spleniuscapitis muscle in humans. Exp Brain Res 1995;105(3):432–8.

[33] Merletti R, Farina D, Hermans H, Freriks B, Harlaar J. Europeanrecommendations for signal processing methods for surface electromy-ography. In: Hermans et al., editors. European recommendations forsurface electromyography: results of the seniam project. Enschede, TheNetherlands: Roessingh Research and Development; 1999. p. 57–68.

[34] Milerad E, Ekenvall L. Symptoms of the neck and upper extremitiesin dentists. Scand J Work Environ Health 1990;16(2):129–34.

[35] Mirka G. The quantification of emg normalization error. Ergonomics1991;34(3):343–52.

[36] Oberg T, Oberg U. Musculoskeletal complaints in dental hygiene: asurvey study from a swedish county. J Dental Hyg 1993;67(5):257–61.

[37] Queisser F, Bluthner R, Brauer D, Seidel H. The relationship betweenthe electromyogram-amplitude and isometric extension torques ofneck muscles at different positions of the cervical spine. Eur J ApplPhysiol 1994;68:92–101.

[38] Saito S, Miyao M, Kondo T, Sakakibara H, Toyoshima H. Ergonomicevaluationof working posture ofvdt operationusingpersonalcomputerwith flat panel display. Ind Health 1997;35(2):264–70.

[39] Sakakibara H, Miyao M, Kondo T, Yamada S. Overhead work andshoulder-neck pain in orchard farmers harvesting pears and apples.Ergonomics 1995;38(4):700–6.

[40] Schuldt K, Ekholm J, Harms-Ringdahl K, Arborelius UP, NemethG. Influence of sitting postures on neck and shoulder e.M.G. Duringarm-hand work movements. Clin Biomech 1987;2:126–39.

[41] Schuldt K, Ekholm J, Harms-Ringdahl K, Nemeth G, Arborelius U.Effects of changes in sitting work posture on static neck and shouldermuscle activity. Ergonomics 1986;29(12):1525–37.


[42] Schuldt K, Ekholm J, Harms-Ringdahl K, Nemeth G, Arborelius UP.Effects of arm support or suspension on neck and shoulder muscleactivity during sedentary work. Scand J Rehabil Med 1987;19(2):77–84.

[43] Schuldt K, Harms-Ringdahl K. Activity levels during isometric testcontractions of neck and shoulder muscles. Scand J Rehabil Med1988;20:117–27.

[44] Schuldt K, Harms-Ringdahl K. Cervical spine position versus e.M.Gactivity in neck muscles during maximum isometric neck extension.Clin Biomech 1988;3:129–36.

[45] Schuldt K, Harms-Ringdahl K. E.M.G./moment relationships inneck muscles during isometric cervical spine extension. Clin Biomech1988;3:58–65.

[46] Scutter S, Turker KS, Hall R. Headaches and neck pain in farmers.Aust J Rural Health 1997;5(1):2–5.

[47] Serratos-Perez JN, Mendiola-Anda C. Musculoskeletal disordersamong male sewing machine operators in shoemaking. Ergonomics1993;36(7):793–800.

[48] Sommerich C, Joines S, Psihogios J. Effects of computer monitorviewing angle and related factors on strain, performance, andpreference outcomes. Human Factors 2001;43(1):39–55.

[49] Sommerich CM, Joines SMB, Hermans V, Moon SD. Use of surfaceelectromyography to estimate neck muscle activity. J ElectromyogrKines 2000;10:377–98.

[50] Stock SR. Workplace ergonomic factors and the development ofmusculoskeletal disorders of the neck and upper limbs: a meta-analysis. Am J Ind Med 1991;19(1):87–107.

[51] Sundelin G, Hagberg M. The effects of different pause types on neck andshoulder emg activity during vdu work. Ergonomics 1989;32(5):527–37.

[52] Sundelin G, Hagberg M. Electromyographic signs of shoulder musclefatigue in repetitive arm work paced by the methods-time measure-ment system. Scand J Work Environ Health 1992;18(4):262–8.

[53] Turville KL, Psihogios JP, Ulmer TR, Mirka GA. The effects of videodisplay terminal height on the operator: a comparison of the 15 degreeand 40 degree recommendations. Appl Ergon 1998;29(4):239–46.

[54] Viikari-Juntura ER. The scientific basis for making guidelines andstandards to prevent work-related musculoskeletal disorders. Ergo-nomics 1997;40(10):1097–117.

[55] Villanueva MB, Jonai H, Saito S. Ergonomic aspects of portablepersonal computers with flat panel displays (pc-fpds): evaluation ofposture, muscle activities, discomfort and performance. Ind Health1998;36(3):282–9.

[56] Villanueva MB, Jonai H, Sotoyama M, Hisanaga N, Takeuchi Y,Saito S. Sitting posture and neck and shoulder muscle activities atdifferent screen height settings of the visual display terminal. IndHealth 1997;35(3):330–6.

[57] Wells R, Norman R, Shannon H, Cole D, Woo H, Bao S. Exposuremeasurement in the upper limbs of vdt operators using electromyog-raphy: Reliability within and between days. In: Proceedings ofPREMUS; 1998.

Sharon Bennett Joines, Ph.D., is Director ofResearch at the Center for Universal Designand an Assistant Professor of Industrial Designat North Carolina State University. Shereceived her B.S., M.S. and Ph.D. degrees inIndustrial Engineering from North CarolinaState University. Her current research initia-tives include ergonomics and occupationalbiomechanics, with special interests in theaging workforce, and product design andevaluations.

Carolyn M. Sommerich, Ph.D., is an AssociateProfessor in the Department of IndustrialWelding & Systems Engineering at The OhioState University and holds an adjunctappointment in the Department of IndustrialEngineering at North Carolina State Univer-sity. Her research focus is ergonomics andoccupational biomechanics, with special inter-ests in the upper body, upper extremities, officeergonomics, and student-related ergonomicissues.

Gary Allen Mirka is an Associate Professor ofIndustrial Engineering at North Carolina StateUniversity where he has been on the facultysince 1992. He received his B.S., M.S. andPh.D. degrees in Industrial and Systems Engi-neering from The Ohio State University. Hiscurrent research interests involve spine biome-chanics, stochastic modeling of biomechanicalsystems, and ergonomic intervention effective-ness research. Dr. Mirka was a recipient of the1993 Volvo Award on Low Back PainResearch.

James R. Wilson is Professor and Head in the
Department of Industrial Engineering at NorthCarolina State University, having served inthat position since 1999. He received a B.A.degree in Mathematics from Rice University in1970, and he received M.S. and Ph.D. degreesin Industrial Engineering from Purdue Uni-versity in 1977 and 1979, respectively. He hasserved on the faculties of The University ofTexas at Austin (1979–1984); Purdue Univer-sity (1985–1991); and North Carolina StateUniversity (1991–present). Dr. Wilson’s cur-
rent research interests are focused on probabilistic and statistical issues inthe design and analysis of simulation experiments. His complete cur-
riculum vitae is available via <www.ie.ncsu.edu/jwilson/vita.pdf> on theweb.
Dr. Moon was educated at Wofford College,Medical College of Virginia, Emory Univer-sity, UNC Chapel Hill and Duke University.He serves as Chief of the Division of Occu-pational and Environmental Medicine in theDepartment of Community and FamilyMedicine (CFM) at Duke. He is also on theIndustrial Engineering Faculty at North Car-olina State University in the ergonomicsdivision and on the UNC Chapel Hill Facultyin the School of Public Health. He serves inthe new Duke Center for Integrative Medicine

as Co-director of Education, Operating Team member, and Director ofClinical and Corporate activities. His research and publications focus on
musculoskeletal disorders, psychosocial factors in occupational pain,ergonomics, work capacity, health promotion, and other occupationalhealth issues.
http://www.ie.ncsu.edu/jwilson/vita.pdf

low-level exertions of the neck musculature: a study of ...jwilson/files/joines06jek.pdf ·...

Documents