Healthcare and fitness professionals must select exercises as a plan of care. Often the clinician struggles with which exercise to prescribe that best targets the shoulder muscles during functional movement patterns. Two upper extremity functional movement patterns include scaption and the spiral-diagonal movements associated with proprioceptive neuromuscular facilitation (PNF). Scaption, originally created by Perry, is described as abduction of the arm in the plane of the scapula (20). Although rehabilitation professionals have commonly used pure frontal plane abduction to assess overall shoulder function, investigators contend that scapular plane abduction is a more efficient type of abduction movement (9,14,21,23). Abducting the shoulder in the scapular plane—approximately 35° anterior to the frontal-coronal plane—is a more natural movement than the traditional movement of shoulder abduction (1) in the pure frontal plane. Scapular plane abduction places the apex of the greater tubercle under the high point of the coracoacromial arch, which diminishes the potential for external impingement (23). Previous investigators have reported muscle activation patterns of the shoulder-arm complex for scapular plane abduction in healthy subjects when lifting dumbbell free weights against gravity with the arm in glenohumeral external rotation (“full can” or thumbs-up position) (10,27,28). The anterior deltoid and infraspinatus muscles were recruited at 71 ± 39% maximum voluntary isometric contraction (MVIC) and 60 ± 21% MVIC, respectively (27). Of the scapular upward rotators, Ekstrom et al. (10) found the serratus anterior to have the greatest activation level (96 ± 24% MVIC), followed in descending order by the upper trapezius (79 ± 19% MVIC), lower trapezius (61 ± 19% MVIC), and middle trapezius (49 ± 16% MVIC). Presently, no one has provided normalized muscle activation ratios for the trunk or back muscles during the performance of scapular plane abduction in standing against resistance provided by dumbbell free weights. This information is important for a clinician because the trunk and back muscles provide stability to the trunk during movement (3).
Proprioceptive neuromuscular facilitation is an approach to therapeutic exercise that uses functionally based multijoint, multiplanar, spiral-diagonal patterns of movement originally described during the 1940s and 1950s by Knott and Voss (16). Upper extremity PNF movement patterns are named according to the position of the shoulder when the diagonal pattern has been completed (15). Component motions of upper extremity diagonal 1 flexion (D1F) involve flexion, adduction, and external rotation at the shoulder with scapular elevation, abduction, and upward rotation. In contrast, motions of the upper extremity diagonal 2 flexion (D2F) pattern involve flexion, abduction, and external rotation of the shoulder with scapular elevation, abduction, and upward rotation. Individuals performing upper extremity D1F and D2F movement patterns are trained to follow their moving hand with their eyes and hence activate the muscles of the head, neck, and trunk. Presently there is a lack of information reporting muscle recruitment (%MVIC) of the shoulder and scapular muscles associated with DIF and D2F. Ekstrom et al. (10) studied D1F only and reported the serratus anterior to have the greatest activation (100 ± 24% MVIC), followed in descending order by the upper trapezius (66 ± 10% MVIC), lower trapezius (39 ± 15% MVIC), and middle trapezius (21 ± 9% MVIC). There is a paucity of research reporting muscle activation ratios for the trunk and back during D1F, D2F, and scaption when the movement pattern occurs in standing and is resisted by dumbbell free weights.
Because research is limited on the aforementioned topics, the purpose of this study is to quantify electromyographic (EMG) activation and recruitment levels for 8 muscles of the shoulder, arm, trunk and back during performance of (a) arm abduction in the plane of the scapula, (b) D1F, and (c) D2F in standing while lifting a dumbbell load in the dominant hand.
Experimental Approach to the Problem
Electromyography is used to record motor unit activity within a muscle or group of muscles. Researchers in health and wellness professions use EMG to observe differences in muscle activity between exercise conditions (17,18). According to Distefano et al. (8), investigators have discovered that higher EMG signals correspond with larger strength gains. Scaption is performed approximately 35° anterior to the frontal plane and up to approximately 120° with the wrist and forearm in neutral (wrist not flexed or extended and the forearm not supinated or pronated) (Figure 1). D1F is started with the dominant hand near the ipsilateral hip in a pronated position. It is then followed by supination, diagonal flexion, adduction, and humeral external rotation until the end position is reached (Figure 2). Diagonal 2 flexion is started with the dominant hand near the contralateral hip in a pronated position. It is followed by supination, diagonal flexion, and abduction with humeral external rotation (Figure 3). Both diagonal patterns are accompanied by scapular elevation, abduction, and upward rotation while subjects tracked their hand movement with their eyes. We examined 8 muscles during each of these 3 exercise conditions including the following: (a) upper trapezius, (b) anterior deltoid, (c) serratus anterior, (d) infraspinatus, (e) middle trapezius, (f) lower trapezius, (g) erector spinae, and (h) external oblique. The independent variables were the 3 exercise conditions (scaption, D1F, and D2F). The dependent variables were the normalized EMG values of the 8 muscles. Electromyographic data were analyzed with Delsys software by processing the root mean square algorithm and normalizing values to the MVIC. Peak EMG activity recorded during flexion or the concentric phase during the second repetition of each test condition was normalized to the muscle's respective peak activity level in the MVIC trial and, therefore, expressed as a percentage of MVIC. The question we sought to answer was as follows: “Is there a difference in activation during the upper extremity PNF patterns compared to scaption, which would suggest that PNF patterns provide a greater strengthening effect?”
Twelve men (26.1 ± 4.4 years) and 13 women (24.5 ± 1.9 years) participated in this study. One man who volunteered to participate was excluded because of a previous shoulder dislocation. The average height, mass, and body mass index of the men were 180.6 ± 5.3 cm, 83.3 ± 9.7 kg, and 25.6 ± 3.5 kg·m-2, respectively, and 168.9 ± 6.4 cm, 68 ± 9.8 kg, and 23.8 ± 3.1 kg·m-2 for the women, respectively. All subjects demonstrated upper extremity (shoulder, elbow, wrist, and hand) range of motion within normal limits before participating. A sample size of 22 subjects was required to detect a mean difference in EMG recruitment of 10% MVIC (effect size = 0.20) between conditions with a statistical power (1 - β) equal to 0.80 at α = 0.05 (12). Before participating in the study, 100% of men (n = 12) and 53.8% of women (n = 7) self-reported that they were engaged in a regular upper extremity weight training program. Subjects self-reporting a history of the following upper extremity conditions were excluded from the study: (a) previous shoulder subluxation, dislocation, or fracture; (b) a history of joint instability, tendonitis, bursitis, impingement, adhesive capsulitis, neurovascular complications, or any condition that limited physical activity for greater than 2 days over the past 6 months; and (c) current complaints of neuromuscular pain, numbness, or tingling in the upper extremity, neck, or back during the 10 repetition maximum (10RM) testing. The study procedures were approved by the Mayo Institutional Review Board, Mayo Clinic, Rochester, MN, USA. Before enrollment in the study, all subjects provided written informed consent.
Electromyographic signals were collected with DE-3.1 double-differential surface electrodes at a sampling frequency of 1,000 Hz. The sensor contacts were made from 99.9% pure silver bars 10 mm in length and spaced 10 mm apart and encased within preamplifier assemblies measuring 41 × 20 × 5 mm. The preamplifiers had a gain of 10 V/V. The combined preamplifier and main amplifier permitted a gain from 100 to 10,000. The common mode rejection ratio was 92 dB at 60 Hz, and input impedance was >1,015 Ω at 100 Hz. Raw EMG signals were processed with EMG works Data Acquisition and Analysis software (Delsys Inc., Boston, MA, USA). The Iron Master 75 lb Quick Lock Adjustable Dumbbell System (Iron Master, Woodinville, WA, USA) was used as the load during each condition.
Data were collected in a research laboratory at the Mayo Clinic in Rochester. Before formal data collection, random order of conditions was determined for each subject by having 1 investigator select note cards and write down the order for each subject. Additionally, each subject completed a separate 10RM test for D1F, D2F, and scaption (15). Subjects were allowed at least 1 day of rest between each 10RM test. Subjects were supervised as they performed 10 repetitions with proper form and speed. The highest load that subjects could lift 10 times while fatiguing on the 10th repetition was considered their 10RM. Average 10RM values for scaption, D1F, and D2F of the men were 10.1 ± 2.5, 11.3 ± 3.9, and 9.9 ± 2 kg, respectively, and 9.7 ± 0.8, 5.1 ± 1.3, and 4.8 ± 0.8 kg for the women.
All men removed their shirt, and women wore a sports bra or sports tank top to permit appropriate placement of EMG electrodes. To allow optimal EMG reading, the skin was shaved if necessary and each subject's skin abraded with an alcohol wipe until erythema was observed. Electrodes were placed superficially and parallel to the direction of the muscle fibers on the subject's hand-dominant side according to established locations for each muscle of interest (5). A ground electrode was secured ipsilaterally on the medial malleolus. The gain on the Delsys EMG instrumentation was set individually for each muscle by having each subject first perform an informal manual muscle break test for 2–3 seconds. Maximum voluntary isometric contraction for each muscle were collected by having the subject perform a formal manual muscle break test for 5 seconds as described by Hislop and Montgomery (13). Next, each subject performed D1F, D2F, and scaption (see Figure 1) in the predetermined random order as previously described. One examiner demonstrated the specific exercise to the subjects, and subjects were allowed to practice with their nondominant upper extremity before formal data collection. Subjects followed a 40-b·min−1 metronome during the 4 repetitions of each exercise condition. We chose 4 repetitions per condition to avoid muscle fatigue. Subjects were permitted 2 minutes between each exercise condition to rest. Time to complete the entire testing sequence was approximately 20 minutes.
An intraclass correlation coefficient (ICC3,1) was used to estimate the test-retest reliability of the EMG recordings. Descriptive and interferential statistics were calculated with SPSS 15.0 for Windows statistical software (SPSS, Inc., Chicago, IL, USA). Because our primary interest was in comparing muscle recruitment across conditions within each muscle, we conducted 1-way analysis of variances for each of the 8 muscles tested (α = 0.05) and used post hoc Bonferroni corrections to determine which condition (scaption, D1F, or D2F) differed from one another. Greenhouse-Geisser corrections were applied when assumptions of sphericity were violated.
To establish intratester reliability, peak %MVIC values from all 4 repetitions during scaption from 8 subjects were analyzed. The ICCs (ICC3,1) for the 8 muscles were as follows: (a) upper trapezius, 0.78 (95% confidence interval [CI], 0.50–0.94; SEM = 7.5% MVIC); (b) serratus anterior, 0.78 (95% CI, 0.51–0.94; SEM = 15.5% MVIC); (c) anterior deltoid, 0.83 (95% CI, 0.60–0.96; SEM = 7.7% MVIC); (d) middle trapezius, 0.95 (95% CI, 0.85–0.99; SEM = 5.3% MVIC); (e) lower trapezius, 0.89 (95% CI, 0.72–0.97; SEM = 4.9% MVIC); (f) infraspinatus, 0.89 (95% CI, 0.71–0.97; SEM = 9.4% MVIC); (g) erector spinae, 0.78 (95% CI, 0.51–0.94; SEM = 6% MVIC); and (h) external oblique, 0.83 (95% CI, 0.60–0.96; SEM = 4.9% MVIC). These values indicate that intrarater reliability was good to excellent (22).
Electromyographic Exercise Data
We found statistically significant differences in EMG recruitment (%MVIC) between exercise conditions for the anterior deltoid, middle trapezius, lower trapezius, and erector spinae muscles. Mean differences between statistically significant conditions are displayed in Table 1. No significant differences in EMG recruitment were detected between exercise conditions for the upper trapezius, serratus anterior, infraspinatus, and external oblique muscles.
To our knowledge, no previous research has normalized EMG activity of the erector spinae or external oblique for scaption, D1F, or D2F. Additionally, a limited amount of research exists reporting the normalized EMG values of the shoulder and scapular muscles for D1F and specifically D2F. Therefore, the primary purpose of this study was to compare the activation levels of the 8 muscles of the shoulder, arm, trunk, and back during the 3 exercise conditions to determine if a difference exists between the 3 conditions and which exercise may provide the greatest strengthening effect. To classify and compare the relative EMG activation of each muscle, we used the system proposed by DiGiovine et al. (7), which classified muscular activation into 4 categories: low (<20% MVIC), moderate (20–40% MVIC), high (41–60% MVIC), and very high (>60% MVIC) (Table 1). The proposed scheme (7) agreed with subsequent researchers who proposed that the threshold value for muscle strength gains during exercise requires EMG activation greater than 50–60% (2,4,18). Therefore, the normalized EMG values were chosen to discuss the relative muscle activation and strengthening effects of scaption, D1F, and D2F. Electromyographic analysis does not genuinely provide a direct measure of muscle strength; instead, it merely measures the neuromuscular activity underneath the surface electrode (Table 2). However, normalized muscle recruitment can be used to determine which exercise pattern produces the highest external demand upon a muscle.
We are only aware of 1 other study that examined shoulder muscle activity during D2F (25). However, the authors did not normalize the data to %MVIC; therefore, we could not compare between the 2 studies. Ekstrom et al. (10) performed 3 repetitions of several upper extremity exercises, including scaption and D1F, using hand-held loads at the subjects' predetermined 5 repetition maximum (5RM) while recording EMG activation of the trapezius and serratus anterior. An experiment by Decker et al. (6) recorded EMG activity of the deltoid, trapezius, and serratus anterior muscles while performing standing humeral elevation in the plane of the scapula. Subjects performed 10 repetitions up to shoulder height while holding a 4.28-kg (9.4 pounds) dumbbell as measured by a force transducer.
The function of the anterior deltoid is to flex, adduct, and internally rotate the arm at the glenohumeral joint (19). Previous EMG studies recorded anterior deltoid activity during scaption, but not during PNF patterns. Decker et al. (6) reported activation of 183 ± 92% MVIC, whereas Townsend et al. (27) reported anterior deltoid recruitment of 71 ± 39% MVIC. During the current study, activation of the anterior deltoid during scaption was 92 ± 26% MVIC, which falls within the range of values from previous studies on scaption. Scaption when performed with a 10RM load in healthy persons is beneficial for strengthening the anterior deltoid muscle.
Actions of the upper trapezius muscle at the scapulothoracic joint (20) include elevation and upward rotation. Other investigators (6,10) have recorded higher values for the upper trapezius activation during scaption than the value reported in the present study (77 ± 19% MVIC). Methodology of the current study was most similar to that of Ekstrom et al. (10), who used a 5RM load and also requested each subject perform scaption to above 120°. Ekstrom et al. (10) reported an upper trapezius EMG signal amplitude of 79 ± 19% MVIC while performing scaption in the standing position. In contrast to the present study and the study by Ekstrom et al. (10), Decker et al. (6) found a higher peak value for upper trapezius recruitment during the performance of scaption (97 ± 41% MVIC). Ekstrom et al. (10) found EMG activation of the upper trapezius to be 66 ± 10% MVIC for D1F, whereas the present study found EMG activation to be 74 ± 17% MVIC. Although, the MVIC values were slightly different across the 3 studies, all studies presented values appropriate for strengthening the upper trapezius during scaption and D1F.
The middle trapezius retracts the scapulothoracic joint (19). The middle trapezius produced very high EMG activation during scaption (68 ± 21% MVIC). This value was inconsistent when compared with previous studies. Ekstrom et al. (10) and Decker et al. (6) recorded EMG activity of 49 ± 16% MVIC and 91 ± 35% MVIC, respectively. Therefore, according to the current study and Decker et al. (6), scaption is adequate for strengthening the middle trapezius. High EMG values of the middle trapezius (46 ± 20% MVIC) were also recorded during D1F in the present study. In contrast, Ekstrom et al. (10) reported EMG activity at 21 ± 9% MVIC for D1F. Values from the present study and Ekstrom et al. (10) indicate that D1F may not be appropriate for strengthening the middle trapezius because muscle recruitment was below 50% MVIC.
Strengthening of the lower trapezius is important because it depresses and retracts the scapulothoracic joint in addition to upwardly rotating the scapula along with the serratus anterior and upper trapezius (20). The current study reported similar EMG values to previous research for the lower trapezius in both scaption and D1F. Lower trapezius muscle activation during scaption (55 ± 20% MVIC) was found to be similar to the value previously reported (61 ± 19% MVIC) by Doody et al. (9). In the current study during D1F, lower trapezius activation was 40 ± 16% MVIC when compared with the previously reported value of 39 ± 15% MVIC (10). Based on these data, scaption would be more advantageous for strengthening the lower trapezius than D1F.
The external oblique and erector spinae were of particular interest to the investigators of the present study because each muscle provides a stabilization or fixation force to the trunk during movement (3) of the upper extremities. As previously stated, previous studies have not examined these trunk muscles during standing scaption or upper extremity PNF patterns. During the concentric phase of lifting, the erector spinae were more active for D1F (42 ± 21% MVIC) and D2F (34 ± 12% MVIC) than during scaption (15 ± 12% MVIC). However, EMG activity of the external obliques during the scaption condition (23 ± 28% MVIC) was not significantly different than either D1F (36 ± 20% MVIC) or D2F (28 ± 16% MVIC). Both the erector spinae and external oblique demonstrated moderate (20–40% MVIC) to high (41–60% MVIC) muscle activity during PNF motions, but not at levels that are not considered to be sufficient for strength training (7).
The serratus anterior is a primary upward rotator and protractor of the scapulothoracic joint (19). Muscle activation for the serratus anterior (74 ± 36% MVIC) during scaption was lower than the previous studies of both Ekstrom et al. (10) and Decker et al. (6), which were 96 ± 24% MVIC and 92 ± 28% MVIC, respectively. During D1F, mean EMG activation of the serratus anterior (81 ± 40% MVIC) was again less than the value reported by Ekstrom et al. (10) (100 ± 24% MVIC). Although the MVIC of the current study is lower than previously presented, D1F and scaption still elicited sufficient muscle activity for strengthening of the serratus anterior.
The role of the infraspinatus is to externally rotate the humerus at the glenohumeral joint. Furthermore, during shoulder abduction, it exerts a depression force on the humeral head to prevent impingement (19). In the current study, EMG activation of the infraspinatus was 71 ± 27% MVIC compared with a value of 60 ± 21% MVIC reported by Townsend et al. (27). Therefore, both the current study and Townsend et al. (27) concur that scaption is sufficient for strengthening the infraspinatus muscle.
As mentioned previously, scaption and D2F elicited sufficient muscle activity (>50% MVIC) to provide strengthening effects of all the shoulder muscles tested. In contrast, during D1F, only the serratus anterior, anterior deltoid, infraspinatus, and upper trapezius elicited high enough EMG activity to be appropriate for strengthening.
Because PNF upper extremity movements use multiplanar, spiral-diagonal patterns of movement (16), rehabilitation professionals have recommended their use during treatment of patients after glenohumeral dislocation to improve joint strength, range of motion, and proprioception and kinesthesia (26). Using a motor-driven shoulder wheel apparatus, Smith and Brunolli (24) examined 8 subjects after unilateral frank anterior glenohumeral dislocations. The authors discovered significant differences in shoulder joint kinesthesia between the previously dislocated and unaffected shoulder. Hence, recurrent anterior glenohumeral dislocations may occur as a result of diminished joint and muscle receptor feedback with subsequent loss of neuromuscular coordination. Muscle activation data from the present study support using PNF upper extremity diagonals as a method for strengthening the shoulder joint. Additional clinical research is needed to provide evidence for the role of PNF in restoring joint proprioception and kinesthesia after shoulder injury.
There are a variety of limitations in this study. As with all surface EMG electrode systems, there is the potential for skin artifact and cross talk; however, this was minimized by cleansing the skin with alcohol and shaving excess hair if needed before placing the electrodes to prevent impedance. Electrodes were placed according to guidelines recommended by Cram and Kassman (5). To prevent the electrode from falling off the subject and chosen motor point, tape was used as reinforcement. Additionally, the external validity of our study is limited because the results only apply to young, healthy individuals and not necessarily to the general population or others with shoulder pathology. Most participants in this study also reported a history of upper-body weight training.
Normalizing raw EMG signals to obtain a %MVIC was a labor-intensive process. When normalizing data to the manual muscle testing positions described by Hislop and Montgomery (13), numerous muscle channels showed extremely high EMG activation levels during the exercise conditions. In particular, the upper and middle trapezius and serratus anterior were often greater than 200% MVIC in many subjects. However, other researchers demonstrated that the level of EMG activity during many upper extremity manual muscle tests varies widely and no single manual muscle test results in an MVIC for every individual. Because of this, Ekstrom et al. (11) advocated that kinesiologic electromyographers consider performing multiple muscle tests and normalize to each individual. Indeed, when reviewing the EMG activity of the muscles during testing, it was noticed, for example, that multiple individuals had a greater EMG signal of the middle trapezius when lying prone with their arm raised above the head in line with the lower trapezius muscle, which is the traditional lower trapezius muscle test position. Because of the variability among our own subjects, whichever muscle test produced the highest EMG activity was considered the MVIC for that particular muscle.
Healthcare and fitness professionals prescribe various strengthening exercises to their clients, including scaption and PNF patters for rehabilitation and strengthening of the shoulder and scapular muscles. Therefore, knowing which functional exercise pattern provides the greatest strengthening effect is beneficial when targeting specific muscles. Based on the results of the current study, scaption and D2F both displayed very high muscle activation at a level that would be appropriate for strength training in the serratus anterior, anterior deltoid, infraspinatus, upper, middle, and lower trapezius. In contrast, D1F showed strengthening effects only in the serratus anterior, anterior deltoid, infraspinatus, and upper trapezius. However, D2F, in addition to providing similar muscle activation and strengthening effects as compared with scaption, also exhibited a statistically greater activation of erector spinae. Although the erector spinae activation was not high enough to provide strengthening, it may be more beneficial to perform D2F for shoulder strengthening because of the additional erector spinae activation for endurance training.
Funding for this project came from the Mayo Program in Physical Therapy, College of Medicine, Mayo Clinic, Rochester, MN, USA. The authors disclose they have no professional relationship with any of the materials or equipment used during this study. The results of the present study do not constitute endorsement of the devices by the authors or the National Strength & Conditioning Association.
1. An KN, Browne AO, Korinek S, Tanaka S, Morrey BF. Three-dimensional kinematics of glenohumeral elevation. J Orthop Res 9: 143–149, 1991.
2. Andersen LL, Magnusson SP, Nielsen M, Haleen J, Poulsen K, Aagaard P. Neuromuscular activation in conventional therapeutic exercises and heavy resistance exercises: implications for rehabilitation. Phys Ther 86: 683–697, 2006.
3. Axler CT, McGill SM. Low back loads over a variety of abdominal exercises: Searching for the safest abdominal challenge. Med Sci Sports Exerc 29: 804–811, 1997.
4. Ayotte NW, Stetts DM, Keenan G, Greenway EH. Electromyographical analysis of selected lower extremity muscles during 5 unilateral weight-bearing exercises. J Orthop Sports Phys Ther 37: 48–55, 2007.
5. Criswell E. Cram's Introduction to Surface Electromyography
(2nd ed.). Sudbury, MA: Jones and Bartlett Publishers, 2011.
6. Decker MJ, Hintermeister RA, Faber KJ, Hawkins RJ. Serratus anterior muscle activity during selected rehabilitation exercises. Am J Sports Med 27: 784–791, 1999.
7. DiGiovine NM, Jobe FW, Pink M, Perry J. An electromyographic analysis of the upper extremity in pitching. J Shoulder Elbow Surg 1: 15–25, 1992.
8. Distefano LJ, Blackburn JT, Marshall SW, Padua DA. Gluteal muscle activation
during common therapeutic exercises. J Orthop Sports Phys Ther 39: 532–540, 2009.
9. Doody SG, Freedman L, Waterland JC. Shoulder movements during abduction in the scapular plane. Arch Phys Med Rehabil 51: 595–604, 1970.
10. Ekstrom RA, Donatelli RA, Soderberg GL. Surface electromyographic analysis of exercises for the trapezius and serratus anterior muscles. J Orthop Sports Phys Ther 33: 247–258, 2003.
11. Ekstrom RA, Soderberg GL, Donatelli RA. Normalization procedures using maximum voluntary isometric contractions for the serratus anterior and trapezius muscles during surface EMG analysis. J Electromyogr Kinesiol 15: 418–428, 2005.
12. Faul F, Erdfelder E, Lang AG, Buchner A. G* Power 3: a flexible statistical power analysis program for the social, behavioral and biomedical sciences. Behav Res Methods 39: 175–191, 2007.
13. Hislop HJ, Montgomery J. Muscle Testing: Techniques of Manual Examination. St. Louis: Saunders, 2007.
14. Johnston TB. The movements of the shoulder-joint. A plea for the use of the “plane of the scapula” as the plane of reference for movements occurring at the humero-scapular joint. Br J Surg 25: 252–260, 1937.
15. Kisner C, Colby LA. Therapeutic Exercise (5th ed.). Philadelphia: F.A. Davis, 2007.
16. Knott M, Voss DE. Proprioceptive Neuromuscular Facilitation, Patterns and Techniques (2nd ed.). Philadelphia: Harper & Row, 1968.
17. Kraemer WJ, Adams K, Cafarelli E. American College of Sports Medicine position stand: Progression models in resistance training for healthy adults. Med Sci Sports Exerc 34: 364–380, 2002.
18. Marshall SJ, Darkin JA, Sallis JF, McKenzie TL. Tracking of health-related fitness components in youth ages 9 to 12. Med Sci Sports Exerc 30: 910–916, 1998.
19. Neumann DA. Kinesiology of the Musculoskeletal System: Foundations for Rehabilitation (2nd ed.). St. Louis: Mosby, 2010.
20. Pink M, Jobe FW. Shoulder injuries in athletes. Clin Management 11: 39–47, 1991.
21. Poppen NK, Walker PS. Normal and abnormal motion of the shoulder. J Bone Joint Surg Am 58A: 195–201, 1976.
22. Portney LG, Watkins MP. Foundations of Clinical Research Applications to Practice (3rd ed.). Upper Saddle River, NJ: Pearson Education, 2009.
23. Saha AK. Mechanism of shoulder movements and a plea for the recognition of “zero position” of glenohumeral joint. Ind J Surg 12: 153–165, 1950.
24. Smith RL, Brunolli J. Shoulder kinesthesia after anterior glenohumeral joint dislocation. Phys Ther 69: 106–112, 1989.
25. Sullivan PE, Portney LG. Electromyographic activity of shoulder muscles during unilateral upper extremity proprioceptive neuromuscular facilitation patterns. Phys Ther 60: 283–288, 1980.
26. Surburg PR, Schrader JW. Proprioceptive neuromuscular facilitation techniques in sports medicine: A reassessment. J Athl Train 32: 34–39, 1997.
27. Townsend H, Jobe FW, Pink M, Perry J. Electromyographic analysis of the glenohumeral muscles during a baseball rehabilitation program. Am J Sports Med 19: 264–272, 1991.
28. Yanai T, Fuss FK, Fukunaaga T. In vivo measurements of subacromial impingement; substantial compression develops in abduction with large internal rotation. Clin Biomech (Bristol, Avon) 21: 692–700, 2006.