Anticipatory postural adjustments to arm movement reveal complex control of paraspinal muscles in the thorax

Anatomical and empirical data suggest that deep and superficial muscles may have different functions for thoracic spine control. This study investigated thoracic paraspinal muscle activity during anticipatory postural adjustments associated with arm movement. Electromyographic (EMG) recordings were made from the right deep (multifidus/rotatores) and superficial (longissimus) muscles at T5, T8, and T11 levels using fine-wire electrodes. Ten healthy participants performed fast unilateral and bilateral flexion and extension arm movements in response to a light. EMG amplitude was measured during 25 ms epochs for 150 ms before and 400 ms after deltoid EMG onset. During arm flexion movements, multifidus and longissimus had two bursts of activity, one burst prior to deltoid and a late burst. With arm extension both muscles were active in a single burst after deltoid onset. There was differential activity with respect to direction of trunk rotation induced by arm movement. Right longissimus was most active with left arm movements and right multifidus was most active with right arm movements. All levels of the thorax responded similarly. We suggest that although thoracic multifidus and longissimus function similarly to control sagittal plane perturbations, these muscles are differentially active with rotational forces on the trunk.     

Effects of stance width on postural movement pattern and anticipatory postural control associated with unilateral arm abduction

We investigated the effects of stance width on postural movement pattern and activation timing of postural muscles during unilateral arm abduction. Thirty-two healthy subjects abducted the right arm at their own timing. Stance width was 0, 9, 18 or 27 cm. Movement angles of leg lateral inclination and trunk lateral flexion to the leg in the frontal plane were analyzed. Based on movement angles at 0 cm width, subjects were classified into three groups: contralateral whole body leaning (CWBLg); ipsilateral trunk flexion (ITFg); and contralateral trunk flexion (CTFg). A high correlation between the movement angles was obtained at 0 cm width (r = 0.82). With increasing stance width, postural movement pattern in the ITFg shifted to patterns characterized by lateral flexion of the trunk toward the side opposite to arm movement, and movement angle of leg-inclination in ITFg and CWBLg decreased. At 0 cm width, left gluteus medius and tensor fascia latae were activated significantly about 40 ms ahead of the right middle deltoid in CWBLg and CTFg, but not in ITFg. However, preceding activation became prominent (about 20 ms) in ITFg for wide stances. Moreover, bilateral activation of the tensor fascia latae was observed in CTFg for all widths.     

The use of intermittent trunk flexion to alleviate low back pain during prolonged standing

The current study examined of the effect of intermittent, short-term periods of full trunk flexion on the development of low back pain (LBP) during two hours of standing. Sixteen participants completed two 2-h standing protocols, separated by one week. On one day, participants stood statically for 2 h (control day); on the other day participants bent forward to full spine flexion (termed flexion trials) to elicit the flexion relaxation (FR) phenomenon for 5 s every 15 min (experimental day). The order of the control and experimental day was randomized. During both protocols, participants reported LBP using a 100 mm visual analogue scale every 15 min. During the flexion trials, lumbar spine posture, erector spinae and gluteus medius muscle activation was monitored. Ultimately, intermittent trunk flexion reduced LBP by 36% (10 mm) at the end of a 2-h period of standing. Further, erector spinae and gluteus medius muscle quietening during FR was observed in 91% and 65% of the flexion trials respectively, indicating that periods of rest did occurred possibly contributing to the reduction in LBP observed. Since flexion periods do not require any aids, they can be performed in most workplaces thereby increasing applicability.     

The role of anticipatory postural adjustments in compensatory control of posture: 2. Biomechanical analysis

The central nervous system (CNS) utilizes anticipatory (APAs) and compensatory (CPAs) postural adjustments to maintain equilibrium while standing. It is known that these postural adjustments involve displacements of the center of mass (COM) and center of pressure (COP). The purpose of the study was to investigate the relationship between APAs and CPAs from a kinetic and kinematic perspective. Eight subjects were exposed to external predictable and unpredictable perturbations induced at the shoulder level while standing. Kinematic and kinetic data were recorded and analyzed during the time duration typical for anticipatory and compensatory postural adjustments. When the perturbations were unpredictable, the COM and COP displacements were larger compared to predictable conditions with APAs. Thus, the peak of COM displacement, after the pendulum impact, in the posterior direction reached 28 ± 9.6 mm in the unpredictable conditions with no APAs whereas it was 1.6 times smaller, reaching 17 ± 5.5 mm during predictable perturbations. Similarly, after the impact, the peak of COP displacement in the posterior direction was 60 ± 14 mm for unpredictable conditions and 28 ± 3.6 mm for predictable conditions. Finally, the times of the peak COM and COP displacements were similar in the predictable and unpredictable conditions. This outcome provides additional knowledge about how body balance is controlled in presence and in absence of information about the forthcoming perturbation. Moreover, it suggests that control of posture could be enhanced by better utilization of APAs and such an approach could be considered as a valuable modality in the rehabilitation of individuals with balance impairment.     

The recruitment order of scapular muscles depends on the characteristics of the postural task

Previous studies show that the scapular muscle recruitment order could possibly change according to the characteristics of the postural task. We aimed to compare the activation latencies of serratus anterior (SA), upper, middle, and lower trapezius (UT, MT and LT, respectively) between an unpredictable perturbation (sudden arm destabilization) and a predictable task (voluntary arm raise) and, to determine the differences in the muscle recruitment order in each task. The electromyographic signals of 23 participants were recorded while the tasks were performed. All scapular muscles showed earlier onset latency in the voluntary arm raise than in the sudden arm destabilization. No significant differences were observed in the muscle recruitment order for the sudden arm destabilization (p > 0.05). Conversely, for voluntary arm raise the MT, LT SA and anterior deltoid (AD) were activated significantly earlier than the UT (p < 0.001). Scapular muscles present a specific recruitment order during a predictable task: SA was activated prior to the AD and the UT after the AD, in a recruitment order of SA, AD, UT, MT, and LT. While in an unpredictable motor task, all muscles were activated after the destabilization without a specific recruitment order, but rather a simultaneous activation.     

Bilateral impairments of shoulder abduction in chronic hemiparesis: Electromyographic patterns and isokinetic muscle performance

ObjectiveTo analyze electromyographic (EMG) patterns and isokinetic muscle performance of shoulder abduction movement in individuals who sustained a cerebrovascular accident (CVA).DesignTwenty-two individuals who sustained a CVA and 22 healthy subjects volunteered for EMG activity and isokinetic shoulder abduction assessments. EMG onset time, root mean square (RMS) for upper trapezius and deltoid muscles, as well as the isokinetic variables of peak torque, total work, average power and acceleration time were compared between limbs and groups.ResultsThe paretic side showed a different onset activation pattern in shoulder abduction, along with a lower RMS for both muscles (21.8 ± 13.4% of the maximal voluntary isometric contraction (MVIC) for the deltoid and 25.9 ± 15.3% MVIC for the upper trapezius, about 50% lower than the control group). The non-paretic side showed a delay in both muscles activation and a lower RMS for the deltoid (32.2 ± 13.7% MVIC, about 25% lower than the control group). Both sides of the group of individuals who sustained a CVA presented a significantly lower isokinetic performance compared to the control group (paretic side ～60% lower; non-paretic side ～35% lower).ConclusionsShoulder abduction muscle performance is impaired in both paretic and non-paretic limbs of individuals who sustained a CVA.     

Patients with chronic,but not episodic,migraine display altered activity of their neck extensor muscles

The current study aimed to investigate differences in activity of neck flexor and extensor muscles in women with migraine considering the chronicity of their condition. Thirty-one subjects with episodic migraine, 21 with chronic migraine and 31 healthy controls participated. Surface electromyography signals were recorded bilaterally from the sternocleidomastoid, anterior scalene, splenius capitis and upper trapezius muscles as subjects performed 5 stages of cranio-cervical flexion (CCF), representing a progressive increase in range of CCF motion. Comparison of normalized root-mean-square among groups was conducted with 3 × 5 ANCOVA with task level as the within-subject variable, group as the between-subject variable, and the presence of neck pain and disability as co-variates. The group with chronic migraine exhibited increased activity of their extensor muscles compared to the control and episodic migraine groups (splenius capitis: F = 3.149, P = 0.045; upper trapezius: F = 3.369, P = 0.041). No significant between-group differences were found for the superficial neck flexors (sternocleidomastoid: F = 1.161, P = 0.320; anterior scalene: F = 0.135, P = 0.874). In conclusion, women with chronic migraine exhibit increased activity of their superficial neck extensor muscles when acting as antagonists during low-load isometric CCF contractions in comparison to non-headache subjects.     

Quantifying ‘normal’ shoulder muscle activity during abduction

The purpose of this experiment was to obtain electromyographic (EMG) activity from a sample of healthy shoulders to allow a reference database to be developed and used for comparison with pathological shoulders. Temporal and intensity shoulder muscle activation characteristics during a coronal plane abduction/adduction movement were evaluated in the dominant healthy shoulder of 24 subjects. Surface and intramuscular fine wire electrodes recorded EMG activity from 15 shoulder muscles (deltoid × 3, trapezius × 3, subscapularis × 2, latissimus dorsi, pectoralis major, pectoralis minor, supraspinatus, infraspinatus, serratus anterior and rhomboids) at 2000 Hz for 10 s whilst each subject performed 10 dynamic coronal plane abduction/adduction movements from 0° to 166° to 0° with a light dumbbell. Results revealed that supraspinatus (?.102 s before movement onset) initiated the movement with middle trapezius (?.019 s) and middle deltoid (?.014 s) also activated before the movement onset. Similar patterns were also found in the time of peak amplitude and %MVC with a pattern emerging where the prime movers (supraspinatus and middle deltoid) were among the first to reach peak amplitude or display the highest %MVC values. In conclusion, the most reproducible patterns of activation arose from the more prime mover muscle sites in all EMG variables analysed and although variability was present, there emerged ‘invariant characteristics’ that were considered ‘normal’ for this group of non pathological shoulders. The authors believe that the methodology and certain parts of the analysis in this study can be duplicated and used by future researchers who require a reference database of muscle activity for use as a control group in comparisons to their respective pathological shoulder group.     

Mcconnell’s patellar taping does not alter knee and hip muscle activation differences during proprioceptive exercises: A randomized placebo-controlled trial in women with patellofemoral pain syndrome

The purpose of this study was to assess the effect of patellar taping on muscle activation of the knee and hip muscles in women with Patellofemoral Pain Syndrome during five proprioceptive exercises. Forty sedentary women with syndrome were randomly allocated in two groups: Patellar Taping (based in McConnell) and Placebo (vertical taping on patella without any stretching of lateral structures of the knee). Volunteers performed five proprioceptive exercises randomly: Swing apparatus, Mini-trampoline, Bosu balance ball, Anteroposterior sway on a rectangular board and Mediolateral sway on a rectangular board. All exercises were performed in one-leg stance position with injured knee at flexion of 30° during 15 s. Muscle activation was measured by surface electromyography across Vastus Medialis, Vastus Lateralis and Gluteus medius muscles. Maximal voluntary contraction was performed for both hip and knee muscles in order to normalize electromyography signal relative to maximum effort during the exercises. ANOVA results reported no significant interaction (P > 0.05) and no significant differences (P > 0.05) between groups and intervention effects in all exercise conditions. Significant differences (P < 0.01) were only reported between muscles, where hip presented higher activity than knee muscles. Patellar taping is not better than placebo for changes in the muscular activity of both hip and knee muscles during proprioceptive exercises. Trial registration number: ClinicalTrials.gov NCT02322515.     

Kinematic chain reactions on trunk and dynamic postural steadiness in subjects with recurrent low back pain

Although subjects with recurrent low back pain (LBP) demonstrate altered trunk control, the kinematic and kinetic responses of the trunk have not been carefully investigated. This study was conducted to compare the standing time, spine range of motion (ROM), and dynamic postural steadiness index (DPSI) based on visual condition between subjects with and without recurrent LBP during upright one leg standing. Sixty-three individuals participated in the study, including 34 control subjects and 29 subjects with recurrent LBP. The DPSI was a composite of the medio-lateral (MLSI), anterior-posterior (APSI), and vertical steadiness indices (VSI) on a force platform. The control group demonstrated longer standing time (s) during the eyes-open condition than the LBP group (26.82 ± 6.03 vs. 19.87 ± 9.36; t = 2.96, p = 0.01). Regarding spine ROM, visual condition was significantly different between groups (F = 7.09, p = 0.01) and demonstrated interactions with spine region and group (F = 5.53, p = 0.02). For the kinetic measures, there was a significant interaction between visual conditions and indices (F = 25.30, p = 0.001). In the LBP group, the DPSI was significantly correlated with the MLSI (r = 0.59, p = 0.002), APSI (r = 0.44, p = 0.03), and VSI (r = 0.98, p = 0.01) in the eyes-closed condition. Overall, the results of this study indicated that the LBP group decreased thorax and lumbar spine rotations during the eyes-closed condition. The LBP group also demonstrated positive correlations with the kinetic indices, enhancing dynamic postural steadiness in the eyes-closed condition in order to possibly avoid pain or further injury. This dynamic postural steadiness strategy is necessary to improve kinetic and kinematic chain reactions in the LBP group. This compensatory pattern supports the development of optimal postural correction strategies to prevent LBP recurrence and might represent a chain reaction to protect trunk control without visual input.     

Internal consistency and test–retest reliability of an instrumented functional reaching task using wireless electromyographic sensors

The purpose of this study was to establish the internal consistency and test–retest reliability of the electromyographic and accelerometric data sampled from the prime movers of the dominant arm during an antigravity, within-arm’s length stand-reaching task without trunk restraint. Ten healthy young adults participated in two experimental sessions, approximately 7–10 days apart. During each session, subjects performed 15 trials of both a flexion- and an abduction-reaching task. Surface EMG and acceleration using wireless sensors were sampled from the anterior and middle deltoid. Reliability was established using Cronbach’s alpha, intraclass correlation coefficients (ICC 2, k) and standard error of measurements (SEM) for electromyographic reaction time, burst duration and normalized amplitude along with peak acceleration. Results indicated high degrees of inter-trial and test–retest reliability for flexion (Cronbach’s α range = 0.92–0.99; ICC range = 0.82–0.92) as well as abduction (Cronbach’s α range = 0.94–0.99; ICC range = 0.81–0.94) reaching. The SEM associated with response variables for flexion and abduction ranged from 1.55–3.26% and 3.33–3.95% of means, respectively. Findings from this study revealed that electromyographic and accelerometric data collected from prime movers of the arm during the relatively functional stand-reaching task were highly reproducible. Given its high reliability and portability, the proposed test could have applications in clinical and laboratory settings to quantify upper limb function.     

The architecture and contraction time of intrinsic foot muscles

Although critical for effective human locomotion and posture, little data exists regarding the segmentation, architecture and contraction time of the human intrinsic foot muscles. To address this issue, the Abductor Hallucis (AH), Abductor Digiti Minimi (ADM), Flexor Digitorum Brevis (FDB) and Extensor Digitorum Brevis (EDB) were investigated utilizing a cadaveric dissection and a non-invasive whole muscle mechanomyographic (wMMG) technique. The segmental structure and architecture of formaldehyde-fixed foot specimens were determined in nine cadavers aged 60–80 years. The wMMG technique was used to determine the contraction time (Tc) of individual muscle segments, within each intrinsic foot muscle, in 12 volunteers of both genders aged between 19 and 24 years.While the pattern of segmentation and segmental –architecture (e.g. fibre length) and –Tc of individual muscle segments within the same muscle were similar, they varied between muscles. Also, the average whole muscle Tc of FDB was significantly (p < 0.05) shorter (faster) (Tc = 58 ms) than in all other foot muscles investigated (ADM Tc = 72 ms, EDB Tc = 72 ms and ABH Tc = 69 ms). The results suggest that the architecture and contraction time of the FDB reflect its unique direct contribution, through toe flexion, to postural stability and the rapid development of ground reaction forces during forceful activities such as running and jumping.     

Influence of the knee flexion on muscle activation and transmissibility during whole body vibration

The influence of the knee flexion on muscle activation and transmissibility during whole body vibration is controversially discussed in the literature. In this study, 34 individuals had electromyography activity (EMG) of the vastus lateralis and the acceleration assessed while squatting with 60° and 90° of knee flexion either with or without whole-body vibration (WBV). The conditions were maintained for 10 s with 1 min of rest between each condition. The main findings were (1) the larger the angle of knee flexion (90° vs. 60°), the greater the EMG (p < 0.001), with no difference on acceleration transmissibility; (2) for both angles of knee flexion, the addition of WBV produced no significant difference in EMG and higher acceleration compared to without WBV (p < 0.001). These results suggest that the larger the knee flexion angle (60° vs. 90°), the greater the muscle activation without acceleration modification. However, the addition of WBV increases the transmissibility of acceleration in the lower limbs without modification in EMG of vastus lateralis.     

Kinematics and muscle activities of the lumbar spine during and after working in stooped postures

Existing biomechanical evidence suggests mechanisms of low back injuries and disorders associated with prolonged stooping. However, no research has tested realistic and more natural stooped work conditions with human subjects in the investigation of the biomechanical responses of the low back in prolonged stooping. The current study was aimed to explore various biomechanical responses of the low back in more realistic and work-related loading and posture conditions of prolonged stooping. Twenty two subjects performed stooped work tasks for 7 min with periodic micro-breaks in upright standing, and various measures for assessing biomechanical responses of the low back were obtained before, during and immediately after the stooped work period. Study results found significant increases (p < 0.05) in the range of lumbar flexion and myoelectric activation of the low back muscles after the stooped work period. During stooped work, the low back extensor muscles did not show flexion–relaxation. It could be concluded that the natural and unrestricted stooped work conditions produced similar viscoelastic responses of the low back to what more severe stooping conditions with posture restrictions did in previous research, but could be more fatigue-prone due to low but consistent activation of the low back extensor muscles during stooped work activities.     

Improved compensatory postural adjustments of the deep abdominals following exercise in people with chronic low back pain

The purpose of this study was to determine if 8 weeks of exercise affects motor control in people with chronic low back pain (CLBP), measured by anticipatory (APAs) and compensatory postural adjustments (CPAs). APAs and CPAs were measured prior to and following 8 weeks in two groups of people with CLBP: an exercise group (n = 12) who attended three exercise sessions per week for 8 weeks; and a non-exercise control group (n = 12) who were advised to continue their usual activities for the duration of the study. APAs and CPAs were recorded during unilateral arm flexion, bilaterally from rectus abdominis (RA), transverse abdominis/internal oblique (TA/IO), and erector spinae (ES) via surface electromyography. Analysis of muscle onsets and APA amplitudes suggests APAs did not change for either group. Ipsi-lateral TA/IO CPAs increased for the exercise group and ipsi-lateral TA/IO CPAs decreased for the control group. Only exercise promoted a pattern of TA/IO activity during CPAs similar to healthy individuals, suggesting improved control of rotational torques. These results show motor control improvement following exercise in people with CLBP, highlighted by improved side specific control of TA/IO.     

Effect of 14 days of bed rest in older adults on parameters of the body sway and on the local ankle function

This study explored the effects of a 14-day horizontal bed rest (BR) without countermeasures on postural sway, maximal voluntary torque and precision of voluntary torque matching. Sixteen subjects were tested before, immediately after and two weeks after BR. The increase in frequency and amplitude after BR was comparable for both sway subcomponents (rambling and trembling) in medial-lateral direction. But in anterior–posterior direction, rambling increased more in frequency (?7% vs. +31%, p < 0.05) while trembling increased more in amplitude (+35% vs. +84%, p < 0.05). The drop in maximal voluntary torque after BR was present for plantar flexion (p < 0.05) but not for dorsal flexion. After the BR, the subjects were less precise in the dorsal flexion torque matching task (p < 0.05). All the observed parameters, except the dorsal flexion torque matching error, returned back to the pre-BR values after the two weeks of re-conditioning. Results of this study indicate that body sway subcomponents responded differently to BR. Based on these findings, it was not possible to draw clear assumptions on the effects of neural and structural changes on body sway.     

Flexion–relaxation ratio in computer workers with and without chronic neck pain

This study evaluated the flexion–relaxation phenomenon (FRP) and flexion–relaxation ratios (FR-ratios) using surface electromyography (sEMG) of the cervical extensor muscles of computer workers with and without chronic neck pain, as well as of healthy subjects who were not computer users. This study comprised 60 subjects 20–45 years of age, of which 20 were computer workers with chronic neck pain (CPG), 20 were computer workers without neck pain (NPG), and 20 were control individuals who do not use computers for work and use them less than 4 h/day for other purposes (CG). FRP and FR-ratios were analyzed using sEMG of the cervical extensors. Analysis of FR-ratios showed smaller values in the semispinalis capitis muscles of the two groups of workers compared to the control group. The reference FR-ratio (flexion relaxation ratio [FRR], defined as the maximum activity in 1 s of the re-extension/full flexion sEMG activity) was significantly higher in the computer workers with neck pain compared to the CG (CPG: 3.10, 95% confidence interval [CI95%] 2.50–3.70; NPG: 2.33, CI95% 1.93–2.74; CG: 1.99, CI95% 1.81–2.17; p < 0.001). The FR-ratios and FRR of sEMG in this study suggested that computer use could increase recruitment of the semispinalis capitis during neck extension (concentric and eccentric phases), which could explain our results. These results also suggest that the FR-ratios of the semispinalis may be a potential functional predictive neuromuscular marker of asymptomatic neck musculoskeletal disorders since even asymptomatic computer workers showed altered values. On the other hand, the FRR values of the semispinalis capitis demonstrated a good discriminative ability to detect neck pain, and such results suggested that each FR-ratio could have a different application.     

The influence of adipose tissue location on postural control

A growing body of evidence suggests, that excessive body weight is inseparably connected with postural instability. In none of previous studies, body weight distribution has been considered as a factor, which may affect results of a static posturography. The purpose of the present study is to quantify some center of foot pressure (COP) characteristics in 40 obese women with android type of obesity (waist-to-hip ratio - WHR ≥ 0.85, BMI: 37.5 ± 5.4) and 40 obese women with gynoid type of obesity (WHR < 0.85, BMI: 36.9 ± 5.1). Variables of postural sway were acquired while subjects were standing quietly on a force plate with eyes open and closed. Both in the sagittal and frontal plane sway range, average velocity, and maximal velocity of COP were calculated. Moreover, the total average velocity and total maximal velocity of the COP displacement were computed.Women with abdominal obesity showed a larger sway range in the anterior-posterior plane with eyes open (p < 0.0282) and eyes closed conditions (p < 0.0115) and a greater maximal COP velocity to compare with subjects with gynoidal obese type (p < 0.0112) with eyes closed condition.The postural stability in obese women from the biomechanical point of view is strongly dependent on body distribution. Women with the abdominal obesity type may be exposed to a greater risk of postural instability as compare to women with gynoid fat distribution.     

Voluntary activation of trapezius measured with twitch interpolation

This study investigated the feasibility of measuring voluntary activation of the trapezius muscle with twitch interpolation. Subjects (n = 8) lifted the right shoulder or both shoulders against fixed force transducers. Stimulation of the accessory nerve in the neck was used to evoke maximal twitches in right trapezius. The twitch-like increments in force (superimposed twitches) evoked during different strength voluntary contractions were linearly related to voluntary force (r = ?0.82 to ?0.99). Hence, voluntary activation could be quantified by twitch interpolation with this stimulus. Comparison of unilateral and bilateral MVCs showed that maximal voluntary force was greater in unilateral than bilateral efforts (92.7 ± 2.9% and 82.3 ± 5.8% MVC, respectively) but voluntary activation was similar (88.6 ± 9.6% and 91.7 ± 5.2%). Trapezius is commonly affected in work-related musculoskeletal disorders. Measurement of voluntary activation will be a useful technique to demonstrate whether the reduced maximal voluntary force reported in such disorders is due to muscular or neural factors.     

Influence of gravity compensation on muscle activity during reach and retrieval in healthy elderly

IntroductionArm support like gravity compensation may improve arm movements during stroke rehabilitation. It is unknown how gravity compensation affects muscle activation patterns during reach and retrieval movements. Since muscle activity during reach is represented by a component varying with movement velocity and a component supposedly counteracting gravity, we hypothesized that gravity compensation decreases the amplitude of muscle activity, but does not affect the pattern. To examine this, we compared muscle activity during well defined movements with and without gravity compensation in healthy elderly.MethodsTen subjects performed reach and retrieval movements with and without gravity compensation. Muscle activity of biceps, triceps, anterior, middle and posterior parts of deltoid and upper trapezius was compared between the two conditions.ResultsThe level of muscle activity was lower with gravity compensation in all muscles, reaching significance in biceps, anterior deltoid and trapezius (p ? 0.026). The muscle activation pattern did not differ between movements with and without gravity compensation (p ? 0.662).DiscussionGravity compensation only influenced the level of muscle activity but not the muscle activation pattern in terms of timing. Future studies should examine if the influence of gravity compensation is comparable for stroke patients. This may stimulate early and intensive training during rehabilitation.