Upper Extremity Kinematics and Electromyographic Activity in Uninjured Tennis Players

Featured Application

This normative dataset provides an objective analysis of individual swing patterns, allowing for the detection of maladaptive or injury-producing swing mechanics and guiding individualized injury treatment and prevention.

Abstract

There has been an increase in ulnar-sided wrist pain among tennis players. The purpose of this study was to establish a normative dataset of kinematic and electromyography (EMG) data during the forehand and two-handed backhand groundstrokes. In total, 20 adolescent United States Tennis Association (USTA) ranked tennis players (11/20 Male, Age = 15.0 ± 1.8 years, Height = 1.7 ± 1.1 m, BMI = 21.3 ± 3.4 kg/m², 18/20 right-arm dominant) participated in this study. Kinematics (range of motion and angular velocity) and EMG data were simultaneously acquired during the forehand and two-handed backhand groundstrokes. Minimal differences were found between groupings of age, sex, and USTA ranking. The two-handed backhand groundstroke is characterized by bilaterally flexed elbows and ulnarly deviated wrists, with a flexed wrist and pronated forearm on the non-dominant side and an extended wrist and supinated forearm on the dominant side. EMG activation occurs bilaterally by peak backswing. The forehand groundstroke is characterized by a flexed elbow, pronated forearm, and ulnarly deviated and extended wrist. The wrist is at maximum ulnar deviation at ball impact. This study established an initial foundation for normative data for the forehand and two-handed backhand groundstrokes, which can be used for injury detection, rehabilitation, prevention, and ultimately performance improvement of tennis athletes.

1. Introduction

Tennis is one of the most popular participation sports, with an estimated 87 million athletes worldwide. Participation is only expected to grow, with the International Tennis Federation committed to increasing the total player population to 120 million by 2030 [1]. The repetitive nature of the sport creates specific demands on the musculoskeletal system, placing players at risk for injury. While lower extremity injuries are the most common, 20–49% of injuries occur in the upper extremity, with the majority being chronic overuse in nature [2,3]. The wrist, representing up to 31% of all upper extremity injuries [4], is vulnerable to injury due to both the complicated mechanics and repetitive utilization of backhand and forehand groundstrokes.

The wrist and hand play predominant roles in tennis strokes, functioning as the final link between kinetic energy generated by the lower body, trunk, and upper extremity at the time of ball contact [5]. In a recent review, the incidence and prevalence of wrist pain, specifically ulnar-sided, as well as wrist injury, were noted to account for a higher percentage of total injuries than ever previously described [6]. Extensor carpi ulnaris (ECU) pathology represents the most common etiologies of ulnar-sided wrist pain [7]. Generating topspin during the forehand groundstroke requires the player to grip the racket in excessive ulnar deviation, predisposing the ECU to injury [8]. Additionally, the same injury pathology occurs in the non-dominant wrist during the two-handed backhand groundstroke, specifically the wrist closest to the racket head [7]. This injury mechanism impacts both elite and less experienced players alike. Less experienced players may subject themselves to larger stress due to poor technique and increased vibration loads in the elbow and wrist during stroke production [9]. Elite tennis players experience increased volumes of play, which has been suggested as positively correlated with injury rate [10,11,12].

While many studies have examined the kinematic characteristics of the tennis serve [13,14,15,16], limited information exists regarding the mechanics of the forehand and two-handed backhand groundstrokes. Busuttil et al. recently analyzed the non-dominant wrist during the two-handed backhand [17], but no electromyographic (EMG) recordings were acquired. To allow for the development of rehabilitation, return to play, and ultimately injury prevention strategies, normative movement patterns must be characterized. To date, to our knowledge, kinematic and EMG normative data for the forehand and two-handed backhand do not exist.

To address the lack of data, this study examined kinematic and electromyographic (EMG) data during the forehand and two-handed backhand groundstrokes for the creation of a normative database. It was hypothesized that uninjured adolescent United States Tennis Association (USTA) ranked tennis players, regardless of age, sex, or USTA ranking, produce consistent stroke mechanics in the forehand and two-handed backhand groundstrokes to allow for the creation of a single normative database for both kinematic and EMG data. These data will provide the necessary foundation to determine the contributing factors for the increased incidence of ulnar-sided wrist pain and injury recently observed in tennis athletes.

2. Materials and Methods

2.1. Subjects

The study was reviewed and approved by the Mayo Clinic Institutional Review Board. Prior to participation in the study, the subjects signed a written consent form. To be eligible for the study, the subjects had to be between ages 12 and 21, actively participate in USTA-sanctioned events, have pain-free range of motion (ROM) of the bilateral upper and lower limbs, and not have any current musculoskeletal upper extremity injuries. In total, 20 subjects participated in this study (11/20 Male, Age = 15.0 ± 1.8 years, Height = 1.7 ± 1.1 m, BMI = 21.3 ± 3.4 kg/m², 18/20 right-arm dominant). This sample size is consistent with ranges established by previous sport kinematic normative database studies, in which the sample sizes ranged from 7 to 29 subjects (mean = 21) [14,18,19,20,21,22].

Male subjects were asked to perform testing shirtless while females wore a tight-fitting sports bra or tank top. All subjects wore tight-fitting shorts. Subjects used their own racket to keep swing patterns consistent with match play. Retroreflective markers were placed on anatomic landmarks and were used to define joint centers and joint axes, measure segment length, and track motion. All markers were placed directly onto the skin, if possible. The markers were secured with double-sided tape to allow for unrestricted movement. Four additional markers were attached to the subject’s racket to assist with event detection.

2.2. Data Collection

Three-dimensional marker trajectories were collected at 400 samples/second with a 10-camera motion capture system (Raptor 12HS, Motion Analysis Corporation, Rohnert Park, CA, USA). Simultaneously, surface EMG electrodes (MA300, Motion Lab Systems, Inc., Baton Rouge, LA, USA) collected data at 2400 Hz from the bilateral extensor carpi radialis longus/brevis (ECR), extensor carpi ulnaris (ECU), flexor carpi radialis (FCR), and flexor carpi ulnaris (FCU) muscles throughout the swing cycle. EMG data were collected from each muscle before beginning data collection to establish the quiescent activity level. A ball machine (Elite Liberty, Lobster Sports, Inc., North Hollywood, CA, USA) located approximately 18 m (60 feet) away delivered the ball at a constant speed of 18 m/s (40 mph) without spin.

The subjects completed their typical warm-up prior to test initiation and then took 20–30 practice forehand and two-handed backhand groundstrokes to become familiarized with the testing environment, as well as the speed and trajectory of the tennis ball. Before data collection, the participants stood in the center of the motion-capture volume, facing the ball machine in an anatomic position to define their neutral position. Once the tennis ball was projected by the machine, the subject performed their typical two-handed backhand groundstroke with instructions to return straight back to the ball machine as naturally as possible, without excessive topspin. Subjects were instructed to perform groundstrokes with maximum intensity and technique, consistent with match play. A trial was considered satisfactory if the ball reached the ball machine. Hits that did not follow the intended trajectory were not analyzed. A minimum of 5 acceptable backhand strokes were collected. The data collection process was repeated for the forehand stroke, with time to warm up for the forehand stroke if needed, and a minimum of 5 acceptable forehand strokes were collected. Subjects were videotaped in the frontal and sagittal plane using multiple cameras.

2.3. Data Analysis

Joint kinematics and EMG signals were analyzed from three trials using a commercial software program (Visual3D, C-Motion, Inc., Germantown, MD, USA). Both fatigue and physiological variability were assumed to be negligible; subjects were given a consistent 10 s rest between groundstrokes, which is more than the timing between subsequent hits in competitive match play. Position data were filtered using a GCVSPL filter [23]. Segment coordinate systems were defined in accordance with International Society of Biomechanics (ISB) recommendations [24]. The upper-body kinematic chain was constrained to have six degrees of freedom at the pelvis, trunk, and upper arm, and two DOF at the forearm and hand. Each subject’s stroke mechanics were divided into three phases: (1) racket preparation, (2) acceleration, and (3) follow-through, which were defined by four time points: (1) initiation of backswing, (2) termination of backswing, (3) ball impact, and (4) end of follow-through. Ball impact was determined based on the acceleration of the four racket markers. Wrist and elbow kinematics were calculated per ISB standards [24].

Biomechanical variables of interest included elbow (flexion–extension) and wrist (pronation–supination, flexion–extension, radial–ulnar deviation) range of motion, and angular velocities. To create a normative database, the mean and ±1 standard deviation were created for each variable. Data were additionally grouped based on age (12–13 and 15–17), sex, and USTA ranking (<500 and >500, taken at the time of testing). Statistical parametric mapping (SPM) was used to investigate if there were any significant differences in the kinematic data of these groupings [25]. This analysis was performed using open-source spm1d code implemented in MATLAB (MathWorks, Natick, MA, USA). Joint range of motion and angular velocities were compared between groupings for both swing types using an SPM two-tailed t-test (α = 0.05). A scalar output statistic, SPM1, was calculated for each time point. When the SPM1 value exceeds the critical threshold value of α = 0.05, there are significant differences between groups at this time point. Time points of significant difference were indicated by solid bars under each kinematic graph.

Raw EMG signals were bandpass-filtered at 10 Hz and 500 Hz. EMG data were then filtered with a fourth-order low-pass Butterworth filter using a 10 Hz cutoff, and the root-mean-square values were calculated. The threshold was determined from each subject’s quiescent trial. For each groundstroke trial, EMG activation was taken to occur when the amplitude was 3 standard deviations above the threshold value for a 50 ms window, and EMG termination occurred when the amplitude dropped 3 standard deviations below the threshold value [26].

3. Results

Swing times and kinematics at peak backswing and ball impact are summarized in

Table 1. Swing time metrics and joint kinematics at peak backswing and ball impact (Mean ± SD) for the forehand and two-handed backhand groundstrokes. Negative angle values indicate ulnar deviation, extension, and pronation.

Swing Metrics	Two-Handed Backhand		Forehand
TOTAL SWING TIME	1.6 ± 0.4 s		1.8 ± 0.2 s
RACKET PREPARATION	1.0 ± 0.3 s		1.1 ± 0.2 s
ACCELERATION	0.2 ± 0.1 s		0.3 ± 0.1 s
FOLLOW-THROUGH	0.4 ± 0.1 s		0.4 ± 0.1 s
KINEMATICS AT PEAK BACKSWING (°)	Non-Dominant	Dominant	Dominant
WRIST RADIAL–ULNAR DEVIATION	−13 ± 13°	−11 ± 12°	−16 ± 11°
WRIST FLEXION–EXTENSION	−43 ± 17°	10 ± 20°	−35 ± 19°
FOREARM SUPINATION–PRONATION	−18 ± 25°	11 ± 20°	−8 ± 17°
ELBOW FLEXION–EXTENSION	77 ± 26°	48 ± 16°	69 ± 22°
KINEMATICS AT BALL IMPACT (°)	Non-Dominant	Dominant	Dominant
WRIST RADIAL–ULNAR DEVIATION	−18 ± 11°	−20 ± 13°	−23 ± 16°
WRIST FLEXION–EXTENSION	−42 ± 15°	19 ± 14°	−41 ± 20°
FOREARM SUPINATION–PRONATION	−27 ± 24°	7 ± 12°	−11 ± 15°
ELBOW FLEXION–EXTENSION	69 ± 24°	69 ± 19°	73 ± 22°

. The kinematics and EMG data of the dominant and non-dominant elbow and wrist for both the two-handed backhand (Figure 1 and Figure 2) and forehand (Figure 3 and Figure 4) strokes were used to create a normative database. To reduce the variation due to each subject’s individual windup, each graph represents 500 ms before ball impact to 300 ms after ball impact, where ball impact represents t = 0 s. Joint angles are expressed in degrees and angular velocity in degrees/second.

There was essentially no difference between age, sex, and USTA ranking, with a few exceptions. During the two-handed backhand groundstroke, five areas of significant differences existed: non-dominant forearm supination–pronation between age (p = 0.02) (Figure 5a), dominant (Figure 5b) and non-dominant (Figure 5c) elbow flexion–extension and dominant wrist extension angular velocity (Figure 5d) between sex (p = 0.04, 0.01, and <0.01, respectively); and dominant wrist flexion–extension angular velocity between USTA rankings (p = 0.01) (Figure 5e). Only one area of significant difference existed during the forearm groundstroke in the dominant forearm supination–pronation angular velocity between sex (p < 0.01) (Figure 5f).

4. Discussion

This study provides kinematic and EMG normative bands for the forehand and two-handed backhand groundstrokes. When comparing kinematics between age, sex, and USTA ranking, six instances of significance were found. The most differences were found between males and females, specifically in wrist angular velocities. Fleisig et al. also observed differences in Olympic caliber athletes during the tennis serve, where male players generated more shoulder internal rotation velocity than females. The differences were attributed to increased internal rotation strength in males over females, and no suggestion was made to teach mechanics differently to males and females [14]. The lack of differences observed in this study supports our hypothesis and improves confidence in the establishment of a single normative database.

4.1. Two-Handed Backhand

The normative data for the two-handed backhand groundstroke have applications in injury prevention. At peak backswing, the elbow is flexed bilaterally, and the wrist ulnarly deviates bilaterally, with the non-dominant wrist positioned in a slightly increased position of ulnar deviation, compared with the dominant wrist. The non-dominant side is positioned with a flexed wrist and pronated forearm, while the dominant side presents an extended wrist and supinated forearm. Moving from peak backswing to ball impact, the elbow maintains flexion, and increases in wrist ulnar deviation are seen bilaterally. Forearm and wrist sagittal plane positions are maintained bilaterally. Peak ulnar deviation does not occur at ball impact bilaterally, but at 0.10 ms and 0.12 ms after ball impact, on dominant and non-dominant sides, respectively. The kinematics described in this study agree with previous reports of the two-handed backhand [17]. EMG activation occurred by peak backswing for all muscles and was maintained throughout the swing. Non-dominant FCU onset occurred the closest to peak backswing, and the ECR experienced the most activation among subjects throughout the entirety of the swing bilaterally. The EMG activation onset and offset times reported are like those observed in previous studies [27]. With the dominant wrist moving towards extension during the racket acceleration phase, it allows for a concentric coactivation of extensor muscles during ball impact, which has been shown to be advantageous [28]. Current knowledge of EMG studies suggests that novice players impact the ball with a dominant flexing wrist, causing repetitive eccentric contraction and lengthening of the wrist extensor muscles [29], which may cause injury due to the extensor activation during eccentric contraction where higher forces can be generated due to the muscle force–velocity relationship.

4.2. Forehand

Similarly, knowledge of the forehand groundstroke is significant for injury prevention. Peak backswing is characterized by a flexed elbow, pronated forearm, and ulnarly deviated and extended wrist. Moving from peak backswing to ball impact, the elbow maintains this position. Unlike the two-handed backhands stroke, the wrist is positioned at maximum ulnar deviation at ball impact. Like the two-handed backhand, EMG activation occurred by peak backswing for all muscles. Muscle onset occurred 500 ms prior to ball impact at all muscles for most subjects and maintained activation throughout the swing pattern. The forehand is the most powerful and commonly used groundstroke in tennis, and the positioning of a maximal ulnar deviation at ball impact subjects the ECU tendon to increased injury risk. The traditional tennis forehand stroke has changed dramatically in the last 10 years, moving towards the modern topspin forehand [30]. Ulnar-sided injuries have been shown to increase in the forehand stroke with the utilization of Western and semi-Western grip types, which are most effective in generating topspin [31]. The increased incidence of non-professional tennis players moving towards techniques mimicking professional athletes may not be supported by the players’ individual strength and technical development [32].

The present study establishes a solid initial foundation of normative data that enable an in-depth objective analysis of stroke patterns. This study also has some limitations. The study was limited to uninjured adolescent, sub-elite level tennis players. While subjects were uninjured at the time of testing, they may have employed stroke mechanics that might predispose them to injury. Grip-type and racket specifications were not standardized, although players were instructed to return the ball as naturally as possible, without excessive spin. These factors may have resulted in increased variability in the normative bands. In addition, different muscle activation patterns have been observed due to increased racket mass and weight distribution, and spin types [26,33]. Future studies should investigate the kinematics and EMG of different grip types and topspin, enhancing the understanding of stroke mechanics that predispose players to ulnar-sided wrist injuries [34].

5. Conclusions

This study established an initial foundation of normative data for the forehand and two-handed backhand groundstroke biomechanics of adolescent USTA tennis athletes. The database contains the upper extremity kinematics and EMG for the elbow, wrist, and forearm. This information may benefit players and coaches by providing an objective analysis of individual swing patterns, allowing for the detection of maladaptive or injury-producing swing mechanics, and guiding individualized injury treatment and prevention. This normative dataset can be used for injury detection, rehabilitation, prevention, and ultimately performance improvement of tennis athletes.