FOREARM POSITION MATTERS DURING ECCENTRIC SHOULDER EXERCISES: AN EMG RECRUITMENT STUDY WITH IMPLICATIONS FOR REHABILITATION

Abstract

Background:

Eccentric exercise has demonstrated great utility in the rehabilitation of various shoulder pathologies. Research on the electromyographic (EMG) activity of the shoulder musculature during these activities is limited, however. Furthermore, no studies have observed how forearm positioning during exercise affects EMG output.

Purpose/Hypothesis:

The purpose was to examine the degree of specific muscle recruitment among commonly used eccentric exercises in rehabilitation of the upper extremity and shoulder. Secondarily, the authors hypothesized that different hand/forearm positions would alter EMG activity within the targeted musculature during a given exercise.

Study Design:

Prospective cross-sectional observation of EMG analysis

Methods:

This study analyzed surface EMG data obtained from 10 healthy individuals during five eccentric exercises of the dominant extremity, performed in a randomized order: side-lying eccentric horizontal abduction (SL ER), half-kneeling weighted ball decelerations (BALL DC), seated eccentric external rotation in scaption (STD ER), standing eccentric external rotation at 0deg (STND ER), supine eccentric external rotation at 90deg (SUP ER). Each exercise was performed with two to three forearm position variants commonly used in clinical environments: neutral, pronation, and/or supination. EMG data were collected from the upper trapezius, infraspinatus, teres minor, latissimus dorsi, and anterior/middle/posterior deltoid. Data were analyzed for each individual exercise and within each muscle using a mixed-model ANOVA repeated across forearm position. Significant interactions were followed by a Bonferroni post-hoc test for pairwise comparisons. Effect size was calculated for all significant pairwise comparisons using a Cohen's d statistic.

Results:

Significant differences in EMG activity for the selected musculature exist between forearm positions for four of the five exercises and Cohen's d effect sizes 0.178 – 1.159.

Conclusion:

Specific eccentric shoulder exercises activate muscles of the shoulder complex differently based on forearm positioning.

Level of Evidence:

Level 2

INTRODUCTION

In rehabilitation of the shoulder, the rotator cuff and posterior shoulder girdle musculature are regions of high interest.^1,2 Strengthening these structures can contribute to greater balance among the muscles of the shoulder, including the rotator cuff, lending to improvement of upper extremity function after injury or surgical procedures.^2,3,4,5 Specific strengthening exercises are employed to accomplish this, often inclusive of eccentric training, in order to maximize shoulder performance and reduce pain.^2,6,7,8,9 While a sizeable amount of literature exists examining muscle recruitment patterns of certain isometric and concentric exercises,^{7,10,11,12,13} there is relatively little data regarding shoulder muscle activation during eccentric exercises (along with any potential variations) utilized for the posterior rotator cuff. Therapeutic treatment efficacy often depends on sufficient recruitment of targeted muscles, so understanding this relationship may help to improve the effectiveness of exercise prescription.

Eccentric exercises consist of a muscular contraction wherein the contractile unit (muscle and tendon) lengthen under external load.^14,15 There are unique molecular and neural characteristics which distinguish eccentric activities from isometric and concentric contractions.^14,15 These characteristics can positively affect the morphology of the target muscle as well as neuromuscular control – both of which are primary goals in rehabilitation and injury prevention.¹⁵ Eccentric exercises have shown promising outcomes in the rehabilitation of many different conditions in both the upper and lower extremities.^{3,4,16,17,18,19,20} A majority of research has indicated that tendon-related injuries respond especially well to eccentrics, likely due to histological changes about the muscle-tendon interface.^{4,14,17,18,19} Furthermore, the shoulder complex has been a topic of more recent research as eccentric exercises seem to provide greater strength and pain improvements versus general exercise protocols in the short and long-term.^16,17 However, information on eccentric exercise selection based on targeted structures is scarce. Different variations in upper extremity positioning (hand, elbow, forearm) have also not been studied with these activities.

The efficacy of most rehabilitation protocols is often measured utilizing outcome measures and subjective reporting. These protocols are guided by studies that indicate the feasibility and effectiveness of the desired exercises or activities. Electromyographic (EMG) data is one indicator of efficacy as clinicians can better understand the degree of muscle activation (or lack thereof) within targeted muscle groups for given exercises and exercise variations.^21,22 The glenohumeral (GH) and scapulothoracic (ST) joints have been studied extensively utilizing EMG data to examine muscle activity with certain shoulder injuries,^23,24,25 during functional activities (such as overhead throwing),^26,27,28 and with particular therapeutic and exercise interventions.^{10,12,13,29,30,31,32} Other researchers have gone further, isolating precise hand positions and placements during similar exercises to identify the most appropriate techniques.^7,33,34 This descriptive data serves as a foundational base for physical therapists and other clinicians to better understand how the shoulder complex operates under different stimuli, offering a more selective and effective means of treatment approach.

The aim of this study was two-fold. Firstly, to examine the degree of specific muscle recruitment among commonly used eccentric exercises in rehabilitation of the upper extremity and shoulder. Secondly, hand/forearm positions were varied throughout the exercises to determine if differences exist in amplitude of EMG activity within the targeted musculature. The authors hypothesized that EMG activity would predominantly differ between exercises due to the intended nature of each activity and the dissimilar planes of motion. The authors also hypothesized that hand/forearm positioning would significantly alter the EMG patterns within a given exercise.

METHODS

Approval from the Houston Methodist Institutional Review Board (IRB) was first obtained for this ­prospective observational study. Healthy, young participants were enrolled from an outpatient clinical setting with the following exclusion criteria: previous history of shoulder injury occurring in dominant extremity, current painful dysfunction resulting in exercise limitation, any health-related exercise limitation as ordered by physician, ages outside of 18-40, and inability to access clinic and equipment. Subjects were also not considered if they reported actively training for competition (body-building, power-lifting, professional sports). All potential participants were informed (verbally and written) of the intended procedures and risks of the study. Consenting participants signed the written informed consent.

A total of 10 subjects (8 male, 2 female; age: 28.6 ± 3.69 years; height: 179.3 ± 7.55 cm; weight: 82.6 ± 12.80 kg; body mass index (BMI): 25.5 ± 2.93 kg/cm² [Table 1]) were consented for this study. Sample size was selected based on previous literature reporting on similar EMG measures to those used in this investigation.^7,12,35,36 Subjects were initially provided visual demonstration and verbal instruction on the eccentric exercises. Each subject was allowed to perform the exercise without resistance to ensure appropriate technique was achieved.

Table 1.

Participant Demographics.

Subject	Sex	Age	Height (cm)	Weight (kg)	BMI
1	Female	29	162.5	52.6	19.9
2	Male	28	180.3	79.4	24.4
3	Male	35	175.3	86.2	28.1
4	Male	32	189	90.7	25.4
5	Male	25	189.5	104.3	29
6	Male	24	172.7	90.7	30.4
7	Male	25	182.9	85.3	25.5
8	Male	28	177.8	80.3	25.4
9	Female	26	180.3	72.6	22.3
10	Male	34	182.9	83.5	25
Average		28.6	179.32	82.56	25.54
Standard Deviation		3.69	7.55	12.8	2.93

Once the participant was adequately instructed on the exercises, they were prepared for electrode placement. The skin surface of the dominant upper extremity was cleaned using an alcohol swab to optimize electrode contact and reduce skin impedance. Self-adhesive snap surface bipolar electrodes (Delsys, Inc. Natick, Massachusetts) were placed with an interelectrode distance of at least 2cm to the posterior rotator cuff and shoulder musculature (anterior deltoid, middle deltoid, ­posterior deltoid, infraspinatus, teres minor, upper trapezius, and latissimus dorsi). Electrode placement was performed by an experienced research scientist (>15 years) with palpation assistance from an experienced physical therapist (>10 years) with specialization in sports and orthopedics. Each muscle was palpated individually against resistance to attain optimal positioning parallel to muscle fiber orientation. After securing the electrodes, adequate connectivity and signal validity was assured. Signal was calibrated using EMGworks® data acquisition and analysis software (Delsys) for each individual lead. The same software was used once leads were placed to confirm adequate signal strength. Maintenance of proper lead contact was monitored during and after each trial. The investigators responsible for electrode placement remained ­consistent throughout the study.

The following exercises were performed on the self-reported dominant side in a randomized order: 1) side-lying eccentric horizontal abduction [SL ER], 2) half-kneeling weighted ball decelerations [BALL DC], 3) seated eccentric external rotation in scaption [STD ER], 4) standing eccentric external rotation at 0deg [STND ER], 5) supine eccentric external rotation at 90deg [SUP ER]). Figure 1 provides a visual example of the exercise techniques. Each exercise utilized a similar resistance across all participants. SL ER, 5-pound dumbbell; BALL DC, 2kg ball; STD ER, 5-pound dumbbell; STND ER/SUP ER, 2-meter resistance standing 3 meters away from anchor (9.1 lbs of resistance (blue band) based on 150% elongation – Theraband, Performance Health, Akron, Ohio). Ordering was performed utilizing a spreadsheet randomization algorithm for each exercise and then within each exercise for forearm positioning.

Figure 1.

Visual representation of eccentric shoulder exercises. A. Sidelying Eccentric Horizontal Abduction (SL ER), B. Half-kneeling Weighted Ball Decelerations (BALL DC), C. Seated Eccentric ER in Scaption (STD ER), D. Standing Eccentric ER at 0deg (STND ER), E. Supine Eccentric ER at 90deg (SUP ER).

All participants then performed the exercises with hand/forearm position variants (pronation, neutral, supination). Exercises were completed with one initial repetition without resistance to verify EMG signal strength and software connectivity, followed by 10 repetitions under resistance with a trained physical therapist ensuring appropriate movement quality and cadence throughout the exercise. The eccentric component of the exercise was performed with a full five-second loading phase.³⁷ All subjects were provided a rest period of at least two minutes in between exercises.³⁸

STATISTICAL METHODS

EMG signal for each lead was analyzed following root-mean-square transformation for each exercise using EMG analysis software (EMG Works, Delsys ®). Within each individual muscle, data were analyzed using a Mixed Model analysis of variance repeated across grip. Significant interactions were followed by a Bonferroni post-hoc test for pairwise comparisons. Type I error was set at α = 0.05 for all analyses. Effect size was calculated for all significant pairwise comparisons using a Cohen's d statistic.⁴⁰ Effect sizes were interpreted as follows: 0-0.1, Negligible (N); 0.1-0.3, Small; 0.3-0.5, Moderate (M); 0.5-0.7, Large (L); > 0.7, very large (VL).

RESULTS

Table 2 provides descriptive data for each muscle analyzed by mean EMG contraction amplitude (mV) across all exercises and forearm positions. The SL ER exercise produced significant differences in muscle activation of the infraspinatus, teres minor, middle deltoid, and posterior deltoid based on forearm position. The supinated position elicited higher recruitment of the infraspinatus (5.247 ± 0.410 mV*10²; p<0.05) versus neutral (4.633 ± 0.332 mV*10²) and pronated (4.592 ± 0.338 mV*10²) positions. For the teres minor, the supinated position (6.261 ± 0.580 mV*10²) was significantly different (greater recruitment) from the pronated position (4.449 ± 0.391 mV*10²; p<0.05), but neither differed from the ­neutral position. The middle deltoid produced the highest level of recruitment (p<0.05) in the pronated position (10.037 ± 0.769 mV*10²) versus either neutral (8.738 ± 0.687 mV*10²) or supinated (7.752 ± 0.670 mV*10²). This remained consistent for the posterior deltoid as well (pronated 7.321 ± 0.959 mV*10², neutral 6.433 ± 0.812 mV*10²; p<0.05). BALL DC was performed in the neutral and pronated forearm positions, producing significant differences (p<0.05) in the middle deltoid (neutral 10.393 ± 1.709 mV*10², pronated 14.639 ± 2.118 mV*10²), posterior deltoid (neutral 4.028 ± 0.488 mV*10², pronated 5.901 ± 0.848 mV*10², latissimus dorsi (neutral 2.645 ± 0.546 mV*10², pronated 4.172 ± 0.963 mV*10²), and upper trapezius (neutral 3.502 ± 0.764 mV*10², pronated 4.349 ± 1.527 mV*10²) – all greater in the pronated position.

Table 2.

Comparison ranking of each muscle per exercise and forearm position. Data are presented as means.

Exercise	RC Muscle	EMG (mV*102)			Sig.	Pairwise Comparisons p-value | Effect Size, Cohen's d
		Neutral	Pronated	Supinated		Neutral vs Pronated	Neutral vs Supinated	Pronated vs Supinated
Sidelying Eccentric Horizontal Abduction (SL ER)	Infraspinatus	4.633 ± 0.332	4.592 ± 0.338	5.247 ± 0.410	p < 0.05	-	p = 0.018 | 0.525 (L)	p = 0.035 | 0.554 (L)
	Teres Minor	4.939 ± 0.396	4.449 ± 0.391	6.261 ± 0.580	p < 0.05	-	-	p = 0.005 | 1.159 (VL)
	Anterior Deltoid	2.211 ± 0.125	2.092 ± 0.096	2.473 ± 0.230	NS	-	-	-
	Middle Deltoid	8.738 ± 0.687	10.037 ± 0.769	7.752 ± 0.670	p < 0.05	p = 0.007 | 0.564 (L)	-	p = 0.016 | 1.002 (VL)
	Posterior Deltoid	6.433 ± 0.812	7.321 ± 0.959	6.355 ± 0.773	p < 0.05	p = 0.041 | 0.316 (M)	-	p = 0.016 | 0.351 (M)
	Latissimus Dorsi	1.885 ± 0.205	1.993 ± 0.259	1.801 ± 0.186	NS	-	-	-
	Upper Trapezius	2.017 ± 0.496	2.005 ± 0.395	1.983 ± 0.585	NS	-	-	-
Half-kneeling Weighted Ball Decelerations (BALL DC)	Infraspinatus	3.403 ± 0.174	3.678 ± 0.204	-	NS	-	-	-
	Teres Minor	3.502 ± 0.764	4.349 ± 1.527	-	NS	-	-	-
	Anterior Deltoid	5.214 ± 0.345	5.284 ± 0.345	-	NS	-	-	-
	Middle Deltoid	10.393 ± 1.709	14.639 ± 2.118	-	p < 0.05	p = 0.002 | 0.698 (L)	-	-
	Posterior Deltoid	4.028 ± 0.488	5.901 ± 0.848	-	p < 0.05	p = 0.002 | 0.856 (VL)	-	-
	Latissimus Dorsi	2.645 ± 0.546	4.172 ± 0.963	-	p < 0.05	p = 0.044 | 0.617 (L)	-	-
	Upper Trapezius	3.502 ± 0.764	4.349 ± 1.527	-	p < 0.05	p = 0.042 | 0.223 (S)	-	-
Seated Eccentric External Rotation in Scaption (STD ER)	Infraspinatus	3.306 ± 0.358	-	2.861 ± 0.314	p < 0.01	-	p<0.001 | 0.418 (M)	-
	Teres Minor	4.412 ± 0.613	-	3.308 ± 0.538	p < 0.01	-	p = 0.002 | 0.606 (L)	-
	Anterior Deltoid	1.809 ± 0.069	-	1.915 ± 0.138	NS	-	-	-
	Middle Deltoid	1.828 ± 0.018	-	1.818 ± 0.016	NS	-	-	-
	Posterior Deltoid	1.539 ± 0.166	-	1.399 ± 0.144	NS	-	-	-
	Latissimus Dorsi	0.982 ± 0.055	-	0.917 ± 0.042	NS	-	-	-
	Upper Trapezius	-	-	-	-	-	-	-
Standing Eccentric External Rotation at 0 deg (STND ER)	Infraspinatus	4.004 ± 0.477(ab)	4.224 ± 0.549(a)	3.931 ± 0.492(b)	p < 0.05	-	-	p = 0.004 | 0.178 (S)
	Teres Minor	2.646 ± 0.390	3.043 ± 0.522	2.699 ± 0.393	NS	-	-	-
	Anterior Deltoid	1.886 ± 0.078	1.871 ± 0.093	1.893 ± 0.067	NS	-	-	-
	Middle Deltoid	1.928 ± 0.092	1.957 ± 0.069	1.899 ± 0.076	NS	-	-	-
	Posterior Deltoid	3.037 ± 0.998	3.118 ± 0.466	2.402 ± 0.481	p < 0.05	-	-	p = 0.021 | 0.478 (M)
	Latissimus Dorsi	1.262 ± 0.110	1.589 ± 0.277	1.154 ± 0.087	NS	-	-	-
	Upper Trapezius	-	-	-	-	-	-	-
Supine Eccentric External Rotation at 90deg (SUP ER)	Infraspinatus	2.437 ± 0.091	-	2.396 ± 0.109	NS	-	-	-
	Teres Minor	1.819 ± 0.211	-	1.955 ± 0.270	NS	-	-	-
	Anterior Deltoid	2.635 ± 0.259	-	2.533 ± 0.188	NS	-	-	-
	Middle Deltoid	3.723 ± 0.423	-	4.056 ± 0.418	NS	-	-	-
	Posterior Deltoid	1.372 ± 0.210	-	1.432 ± 0.204	NS	-	-	-
	Latissimus Dorsi	1.597 ± 0.466	-	2.012 ± 0.695	NS	-	-	-
	Upper Trapezius	-	-	-	-	-	-	-

STD ER was performed in the neutral and supinated positions, producing significant differences (p<0.01) for both the infraspinatus (neutral 3.306 ± 0.358 mV*10², supinated 2.861 ± 0.314 mV*10²) and teres minor (neutral 4.412 ± 0.613 mV*10², supinated 3.308 ± 0.538 mV*10²) with higher EMG activity in the neutral position for both.

STND ER elicited differences (p<0.05) about the infraspinatus and posterior deltoid. The infraspinatus produced greater EMG activity in pronation (4.224 ± 0.549 mV*10²) versus supination (3.931 ± 0.492 mV*10²), but neither differed significantly from neutral (4.004 ± 0.477 mV*10²). The posterior deltoid also demonstrated higher EMG activity in pronation (3.118 ± 0.277 mV*10²) compared to supination (2.402 ± 0.481 mV*10²), but neither were significantly different from the neutral forearm (3.037 ± 0.998 mV*10²).

Figure 2 provides a comparison ranking, by muscle, of each exercise and forearm position. The anterior deltoid produced the highest EMG signaling with the BALL DC activity in pronation and neutral positions. The middle deltoid elicited the highest EMG activity with SL ER in supination followed by SUP ER in supination and BALL DC in the neutral then pronated positions. For the posterior deltoid, the SUP ER exercise in supination elicited the greatest muscle activity followed by SL ER in supination and BALL DC in neutral. The infraspinatus and teres minor produced the highest activation during SL ER in supination followed by SL ER in neutral and then pronation. The latissimus dorsi and upper trapezius both produced maximal recruitment during BALL DC in pronation followed by the neutral position. The upper trapezius was not able to be ranked for all exercises due to negligible muscle activation relative to background EMG signal, indicating no significant contractile activation.

Figure 2.

Muscle activity rankings per exercise. Abbreviations: EMG, Electromyographic activity; mV, millivolts; BALL DC, half-kneeling weighted ball decelerations; SUP ER, supine eccentric external rotation at 90 degrees; SL ER, sidelying eccentric horizontal abduction; STND, standing eccentric external rotation at 0 degrees; STD, seated eccentric external rotation in scaption; SUP, supinated; PRO, pronated; NEUT, neutral.

DISCUSSION

To the authors’ knowledge, this is the first study to investigate differences in EMG muscle activity of the shoulder with specific eccentric exercises and forearm variations commonly utilized in the rehabilitation of the shoulder and upper extremity. In accordance with the original hypotheses, the results in the given investigation support the authors’ hypotheses of dissimilar EMG activity across the exercises as well between forearm positions. The clinical implications may benefit rehabilitation professionals as the data allows for greater perspective in how the shoulder functions when placed under targeted eccentric loads.

Similar to a study by Schoenfeld et al,³⁴ this study provides insight into the general differences in shoulder muscle activation obtained when altering the forearm position. While humeral rotation was maintained throughout each forearm position variant, it is reasonable to assume that subtle differences may exist at the glenohumeral joint, even when adjusting distally at the forearm. These subtleties could preferentially position certain muscles in advantageous or disadvantageous positions for higher levels of contractility. These ideas warrant further investigation.

Other studies have examined muscle activation patterns of the shoulder and rotator cuff with other rehabilitation exercises, demonstrating high levels of utility when considering the rehabilitation of specific structures.^{7,10,12,13,33} In rehabilitation of the pathologic or post-surgical shoulder, this data serves as a crucial check for physical therapists to ensure adequate loads are being applied to the correct tissues while harmful stressors are minimized. The data presented in this investigation may be useful when attempting to target the posterior rotator cuff muscles (infraspinatus, teres minor) versus the deltoids or upper trapezius, for instance. The ­co-contractive mechanism of the rotator cuff is a vital component to healthy shoulder movement and function,^5,9,13,39 and the isolation of those muscles may be preferential to recruiting others to accomplish certain tasks. For example, the SL ER exercise studied in this investigation provides that the supinated forearm position may be best to maximally activate the infraspinatus (5.247 ± 0.410 mV*10²) and teres minor (6.261 ± 0.580 mV*10²), while minimally activating the middle (7.752 ± 0.670 mV*10²) and posterior deltoid (6.355 ± 0.773 mV*10²). However, the STND ER exercise elicits the highest level of activation of the infraspinatus (4.224 ± 0.549 mV*10²) as well as the posterior deltoid (3.118 ± 0.466 mV*10²) and latissimus dorsi (1.589 ± 0.277 mV*10²), all in the pronated position. Thus, in light of the current data, optimal forearm position to maximize targeting of the rotator cuff or other musculature varies from exercise to exercise and should be considered when designing a shoulder strengthening or rehabilitation program.

Some limitations are present in this study and should be considered prior to clinical application. Importantly, the sample size studied was small. However, Cohen's D statistic is commonly used in both large and small sample sizes or pairwise comparisons to determine the degree of the “effects” observed. This, in part, can provide insight as to whether or not pairwise differences observed are relevant or possibly due to an inflated or small sample size. ­Clinical significance may be difficult to determine from this study as the topic of EMG analysis for eccentric exercise is relatively understudied. The participants were also all young, healthy individuals and right hand dominant. Individuals with shoulder pathology in the dominant or non-dominant hand may exhibit altered neuromuscular patterns of recruitment and may not reflect the results of healthy individuals. Furthermore, the resistance within the given exercises was consistent among all individuals, regardless of individualistic variables such as height or BMI. Further study in the selection of optimal resistances may be warranted to elicit maximal training effects. Lastly, surface electrodes may not provide the same EMG detection sensitivity compared to fine-wire electrodes when ascertaining muscle activity. However, recent studies have indicated that surface EMG readings for the selected musculature produce moderate to good agreeability versus fine-wire electrodes.^{7,10,12,23,34,35} Future research may also be required to determine if the degrees of muscle activation observed in this study are consistent across other specialized populations such as older adults or those who perform a high degree of sport-specific shoulder training for athletic performance.

CONCLUSION

The results of this study demonstrate that specific eccentric shoulder exercises may activate muscles of the shoulder complex differently based on forearm position. In instances where isolation of the rotator cuff is preferential, it seems that the SL ER exercises can help to accomplish this goal while minimizing utilization of the other musculature of the shoulder. This data provides physical therapists and other rehabilitation professionals with information to more selectively and efficaciously prescribe exercises for the shoulder.