Impact of Yoga on Postural Stability in Stroke

Deficits in motor capacity following stroke poses a potential threat to postural stability. The purpose of the study was to assess the impact of yoga training on postural alignment, anticipatory postural adjustments, and compensatory postural adjustments in individuals with chronic stroke deficits. This study is a subset analysis from a prospective, randomized control trial using therapeutic yoga. Twenty one individuals with chronic motor deficits from stroke participated in an hour-long group yoga sessions twice a week for 8 weeks focusing on progressively challenging postures, breathing, and meditation in sitting, standing, and supine positions. Postural symmetry, compensatory postural adjustment, and anticipatory postural adjustment from the NeuroCom SMART Balance Master® were used to evaluate changes in postural stability. The Berg Balance Scale was used to evaluate correlations to changes in dynamic balance control. Prior to participation in yoga, significant weight distribution asymmetry was observed during full upright standing. After 8 weeks of yoga training, there were no significant differences in weight distribution between right and left legs at 0 degrees. Subjects demonstrated greater than 50% improvement in symmetry alignment for different standing positions. No significant changes in compensatory or anticipatory postural adjustments were seen following yoga training. Changes in functional BBS scores were significantly correlated to changes in upright standing symmetry. Yoga training may improve postural symmetry enhancing postural control decreasing the risk for falls in individuals with deficits from chronic stroke.


Introduction
Annually, approximately 795,000 individuals sustain a stroke in the United States [1]. The high prevalence, negative stroke squeal and the subsequent disability after stroke make it the most commonly treated physical disability by rehabilitation therapists [2]. An important aspect of stroke recovery is the improvement in postural stability or balance. Postural stability has been defined as the ability to control the center of gravity within a given base of support including both static and dynamic conditions [3]. Balance is commonly impaired after stroke and often continues to be impaired into the chronic (>6 months) phases of recovery; [4][5][6][7][8][9][10] up to 83% of individuals demonstrate poststroke balance deficits [11].
Limitations in postural stability after stroke result from lesions in the brain that impact the sensory/motor systems [12,13] and are correlated with: greater physical impairments and disability, decreased ability to perform activities of daily living (ADLs), and increased fall risk [5,11,[14][15][16][17][18] Up to 76% of individuals who have survived a stroke sustain a fall, much more frequently than the healthy elderly, and the elevated fall rate is associated with depression, fractures, and mortality [5,9,10].
Postural stability emerges from a synchronized integration of sensory and motor function [19]. Deficits in motor subsystems that control balance such as postural symmetry and anticipatory and compensatory postural adjustments (APA & CPA) could negatively impact standing postural stability, increasing the risk of falls [18], decreasing individual confidence with daily activities, and altering the level of social engagement [10]. Standing postural symmetry refers to an individual's ability to maintain an equal distribution of weight over both lower extremities in upright standing [20]. Asymmetry is a common problem following stroke that is associated with hemiparesis [21,22]. Standing asymmetry often results in greater weight distribution to the non-paretic limb resulting in paretic limb weakness, instability, and gait abnormalities [22,23].
Altered postural symmetry also decreases the potential for successful balance control when presented with unexpected challenges (i.e. perturbation) [24]. Under perturbation, an individual uses anticipatory and compensatory postural adjustments to control their balance [25]. Anticipatory postural adjustments (APA) are a motor subsystem that function to offset the potential effects on balance control made by planned motor activities [25]. In situations where the individual is aware of the expected balance challenges, the healthy APA system will alter selected patterns of motor recruitment to maintain stability. In contrast, compensatory postural adjustments (CPA) function to maintain balance following unexpected center of mass perturbation [26]. Motor unit recruitment is based upon the required force needed to counter act the displacement of the center of mass caused by the disturbance [27]. The more effective APA control creates greater stability and the less need for CPA input [27].
Deficits in motor capacity following stroke poses a potential threat to postural stability in quiet standing; changes in alignment, as well as poor APA and CPA responses, decreases the ability of an individual to accept unexpected balance challenges [24]. Identifying intervention techniques effective at improving symmetrical postural alignment as well as improving APA and CPA postural responses for chronic stroke survivors is imperative. However, at this time, there are no specific evidence-based balance training recommendations for these deficits [4].
In addition, the stroke literature confirms improvement in functional activities when linked to intensive activity specific training environments that are progressive in nature; however, few studies have investigated a focused approach emphasizing balance or the mechanisms of balance control [28]. Thus, there is a need for further development and assessment of progressively challenging interventions focusing on balance training and their potential impact on postural symmetry, anticipatory control, and compensatory mechanism [29].
There is preliminary evidence that yoga may be a progressively challenging exercise intervention that could be used to improve poststroke postural stability [30][31][32][33]. The active mind-body connection introduced through a combination of physical postures, breathing, and meditation in yoga may produce different effects considered more beneficial and therapeutic than traditional exercise programs [34][35][36]. Furthermore, the emphasis on postural alignment and static standing balance control found within yoga training could produce significant changes to individual weight distribution in standing increasing the ability to manage control under perturbation. An emphasis on postural strength and control may also impact the individual's ability to prepare for and react appropriately when challenged by a perturbation. Recently, we studied 47 people with chronic stroke following 8 weeks of group yoga and found significant improvement in functional dynamic balance [33]. The purpose of this study was to assess a subset of individuals for: 1) changes in postural alignment in chronic stroke survivors following yoga training; 2) determine the ability for individuals with chronic stroke deficits to apply APAs and CPAs under perturbation; and 3) measure the change in APAs and CPAs following a yoga training.

Design
This is a secondary analysis from a randomized pilot study [33] measuring clinically applied balance assessments found on the NeuroCom SMART Balance Master® Assessment suite. Individuals with chronic stroke de icits were initially allocated to yoga, a yoga plus home program, and a control group. Subjects randomly assigned to the yoga groups received twice weekly yoga in a group format for 8 weeks (16 sessions). A detailed description of the yoga treatment can be found in the pilot study manuscript [33]. Those individuals assigned to the control group received no intervention. Since no signi icant differences were found between the two treatment groups; yoga data were combined to form one group. Control group participants did not have access to the balance assessments so no comparisons were made. The 8 week yoga training occurred at a university based, integrative therapy lab. Outcome assessments were completed at baseline and immediately following training.

Participants
The secondary analysis involves a sampling of 21 of the 47 chronic stroke survivors who were allocated to the yoga training groups. Participants were included if they met the following criteria: 18 years and older; sustained a stroke > 6 months ago; had completed all stroke related rehabilitation; were able to stand with or without a device; scored a > 4 out of 6 on the short 6 item Mini-Mental State Examination [37] (MMSE); and gave written informed consent to participate. Subject were excluded if they were receiving palliative care, were unable to ensure transportation to the sessions, or had a selfreported medical contraindication (serious cardiac conditions; serious chronic obstructive pulmonary disease or oxygen dependence; severe weight bearing pain; a history of significant psychiatric illness; uncontrollable diabetes with recent weight loss) or current enrollment in another research trial. This study was approved by the Indiana University Institutional Review Board and Veterans Administration Medical Center Research and Development Board. Subjects provided written informed consent prior to participation.

Intervention
The study intervention consisted of hour-long group yoga sessions twice a week for 8 weeks (16 sessions). A standardized intervention protocol was developed with input from the rehabilitation research team and the yoga therapist. The yoga therapist taught all sessions and the intervention included progressively challenging postures, breathing, and meditation in sitting, standing, and supine positions. There was a focus on balance improvement, thus chosen postures focused on hip and ankle flexibility and strength [38].

Outcome assessments
Assessments were completed face-to-face by a trained research assistant at baseline and after the 8 week intervention. Self-report of demographics and stroke related characteristics were collected which included: time since stroke; type of stroke (ischemic, hemorrhagic, brainstem, or do not remember); and stroke related disability (Modified Rankin Scale, (mRs). The mRs is a validated measure of the degree of disability and dependence after stroke; [39,40] and defines functional independence with a score of 0-2 (slight to no disability) and dependence as 3-5 (moderate to severe disability) [41,42] Balance measurement was performed on the NeuroCom SMART Balance Master® with safety harnesses to prevent falling to the ground.

Postural stability
Measures of postural stability were assessed using a computerized dynamic balance measurement, the NeuroCom SMART Balance Master®. Postural orientation was measured using the weight bearing squat assessment. This test measured each participant's postural alignment in standing in four conditions with progressively increasing knee flexion. Consistency between subjects for knee position was maintained by goniometric assessment. Quantitative information on postural symmetry was calculated for each position by measuring the percentage of weight distributed through each lower extremity for the four conditions. Improved postural alignment was represented by symmetrical distribution (50%/50%) between right and left lower extremities (Table 1).

Toes Up
Change in force required to control sway during five repeated perturbations where subject is challenged with backward motion

Toes Down
Change in force required to control sway during five repeated perturbations where subject is challenged with forward motion Weight Bearing Squat Test*

Condition 1 -0° of knee flexion (%)
Measure of the percentage of weight distributed between right and left lower extremities  APAs and CPAs were assessed using the motor adaptation test. The motor adaptation test assesses automatic postural responses, which are dependent on efficient sensory/motor information processing within central nervous system structures. The automatic postural responses were tested through a series of perturbations in the sagittal plane by moving the platform in the following directions: 'toes up' and 'toes down' . The test series was repeated five times in both directions. The amount of force required to recover from the perturbation induced sway was recorded for each trial. The unexpected nature of the initial perturbation establishes a clinical measure of the CPAs invoked to maintain postural alignment. The subject was then aware the same perturbation would be repeated four additional times. Consequently, an intact APA would set the system for an improved response under the second perturbation. The change in response following the initial perturbation would represent the force applied to the static postural position from the APA system providing a clinical measure of APA response. Falls were recorded following perturbation if the subject required a stepping strategy or physical assistance from a researcher was needed to control their balance. Each of the NeuroCom tests is outlined in Table 1.

Dynamic balance
Dynamic functional balance was assessed using the Berg Balance Scale (BBS); a 14-item physical performance measure. Previously, we showed balance significantly improved following yoga training for chronic stroke survivors [43]. BBS data are included in this study to assess potential correlations with changes in underlying postural mechanisms.

Statistical Analysis
Descriptive statistics (means, standard deviations, frequencies, and proportions as appropriate) are included to describe the demographics and stroke characteristics of those individuals included in these analyses. To further describe the sample, a comparison of demographics between those with and without NeuroCom analysis was compared. Normality of data was assessed with the Shapiro-Wilks test. While we did so in the primary paper, we did not carry missing data points forward due to the small sample size and exploratory nature of the present analyses. Data derived from the NeuroCom SMART Balance Master® were compared across the two time points (pretest and following 8 weeks of yoga). Postural symmetry and adaptation were evaluated using paired samples t-tests. Mean scores were evaluated for clinical relevance. Relationships between balance and postural symmetry were evaluated using Pearson correlation.

Results
Demographic and stroke characteristic data for the entire study cohort (n=47) and those included in these analyses (n=21) are found in Table 2. Baseline demographic or stroke characteristic data were not statistically different for individuals with or without electronic balance assessment (21 vs 26) except gender. 43% of the participants in this secondary analysis were female compared to only 9% for the overall study. 57% of the subjects rated themselves independent on the modified Rankin scale.

Postural symmetry
Twenty-one study participants completed balance pre/post balance assessment with the NeuroCom SMART Balance Master®. Prior to participation in yoga, significant weight distribution asymmetry (8%, p=0.031) was observed during standing with full knee extension. In addition, a trend towards weight distribution asymmetry was observed with subjects squatting at 60 and 90 degrees (p=0.05) of knee flexion at baseline (Table 3). After 8 weeks of yoga training, there were no significant differences in weight distribution between right and left legs at 0 degrees.   Subjects demonstrated greater than 50% improvement in symmetry alignment for all 4 test positions (Figure 1). A subgroup analysis based on hemiparesis revealed a significant baseline asymmetry in distribution of weight for patients with left hemiparesis at 0, 60, and 90 degrees of knee flexion. Post test results revealed normal weight distributions for all positions within this subgroup, except at 0 degrees.

Postural Adjustment
Trail 1 of the motor adaptation test was used to assess unexpected perturbation response (CPA); no significant differences were found when comparing mean group forces (lbs.) recruited prior to and following training (Table 4). 76% of individuals were able to control their posture during pretest unpredictable perturbation, following training this number increased to 81%. A decrease in force required to maintain stability in trial 2 when compared to trial 1 was quantified as a measure of APA response effectiveness. Mean change in force recruitment to control perturbation induced sway was compared between the first trial and second trial at pretest and posttest for both toes down (TD) and toes up (TU). No significant difference between pretest and posttest force production was found for either the TD or TU condition (Table 4). Mean scores revealed that on average during the TU condition, more force was recruited to control a posterior center of mass displacement.  A subset of participants, 7 during TD and 5 during the TU condition, when comparing pretest trial 1 to trial 2 experienced an increase in force production. Post training analysis revealed that 43% of these subjects in the toes down condition and 60% in the toes up condition experienced a decrease in force production when comparing trial 2 to trial 1. Three individuals experienced a fall during initial perturbation (Trial 1) for the TU condition; two of the three also fell during the TD initial trial. Following training, one of the three individuals was able to maintain their initial balance without falling.
Correlational analyses were run evaluating the eight week BBS score with postural symmetry and postural adjustment mechanisms for the 21 subjects (Table 5). Pretest postural assessments demonstrated a significant inverse correlation with the posttest BBS at 0 (p=0.002), 30 (p=0.005), and 60 (p=0.006) degrees of knee flexion as well as the posttest 0 (p=0.026) degrees of flexion condition. The BBS change score exhibited a significant positive correlation with the posttest 0 (r=0.523, p=0.015) degree knee flexion symmetry score. CPA and APA outcomes were not correlated to BBS posttest outcomes.

Discussion
With this study, we attempted to determine the impact of yoga treatment on postural stability mechanisms in individuals with chronic stroke deficits. Yoga treatment focuses on the awareness of the individual's body alignment and orientation while performing different postures. In addition, participants were asked to actively engage in relaxation while working through directed postural control activities. Based upon the mechanisms of the treatment approach, it was hypothesized that participants in this study would experience positive changes in both postural symmetry and adjustment.

Postural symmetry
Individuals with chronic stroke deficits in this study exhibited significantly deviated postural alignment in upright standing prior to training. Following yoga, postural asymmetry was reduced by as much as 80% for some individuals. Previous studies have indicated that asymmetrical alignment contributes to impaired balance and decreased ambulatory ability [44]. Limitations in both balance and gait speed have been correlated with deficits in social participation [45]. Although participants in this study were in a chronic state of recovery, training with yoga based therapy demonstrated an ability to alter motor strategies resulting in more equalized postural distribution. These findings suggest it is possible with training for an individual in a chronic state of recovery to alter their postural symmetry.
Functional balance measurement using the BBS was found to be inversely related to asymmetry measures prior to and following training in upright standing. Higher posttest BBS scores were related to decreases in asymmetrical alignment. Specifically, 23% of the variance in the BBS posttest score could be explained by posttest upright symmetrical alignment. Postural orientation as measured by symmetry seems to play a role in functional balance stability. A similar finding looking at gait parameters was found by Hendrickson et al. [23].

Postural adjustments
To avoid falling, an individual must compensate for unexpected challenges and adjust to further expected perturbations. Prior to training in this study, a majority (76%) of the subjects demonstrated compensatory postural activation to overcome unpredictable challenges to standing postural stability. The ability of individuals to maintain balance while challenged with unexpected center of mass translation both anterior and posterior is encouraging. CPAs are needed to avoid falling in static and dynamic conditions. Following yoga training, the average amount of force recruited to maintain balance under either condition did not significantly change. This, in part, might be attributed to the effective application of CPA mechanism prior to training. In addition, yoga training does involve a task specific training paradigm that would like impact CPA control. Two individuals, who fell during CPA pre-testing, demonstrated an ability to maintain upright balance control at posttest. The improvement in CPA effectiveness is clinically relevant as it suggests these two individuals following training may be more likely to withstand an unexpected challenging preventing a fall. Postural training focusing on orientation and upright control appeared to improve these individuals' ability to control balance during unexpected challenges.
Participants in this study also exhibited the ability to make effective anticipatory adjustments to their postural control mechanisms prior to the start of the study. This result is also encouraging considering that many of the subjects reported limited activity levels. While overall changes in APA activity were not significant following training, responses did show improvement for several individuals who initially exhibited ineffective APA response (increased force production) during trial 2 of the pretest. This finding suggests the possibility for an individual to improve anticipatory postural adjustments control with training. Changes in CPA and APA mechanism were not found to be correlated to the BBS. The BBS measures deliberate dynamic functional movements while the postural adjustment tests look at perturbation management; the differences may account for the lack of clinical correlation. In addition, CPA and APA mechanisms are governed by automatic, unconscious motor pathways while dynamic balance activities within the BBS are predominately controlled by volitional motor pathways. The difference in neurological control may also account for the limited connection.

Clinical relevance
Limitations in postural alignment and adjustment predispose an individual to balance control and increase risk for falls. Yoga training provides a task specific approach focusing on the awareness of postural alignment and symmetry as it relates to each pose. The emphasis on postural awareness and control has been shown to be effective in enhancing balance control in previous studies [46]. The translation of this training resulted in significant improvement in postural alignment that was related to functional balance outcomes in upright standing. A potential advantage is the ability to translate this approach to a home/ lifestyle program. Yoga represents a therapeutic approach that could be integrated into daily life that often does not occur with more traditional therapeutic exercise home programs. The opportunity to develop a consistent training program could significantly limit postural atrophy, decreasing fall risk and medical necessity. Research evidence supports a link between alignment and dynamic balance and gait. Maintaining strong postural symmetry could provide a key foundation for continued functional independence and societal participation. The lack of correlation between dynamic balance and postural adjustments suggests the need for developing task specific training for both mechanisms since training in one paradigm is not likely to translate to improvements in the other.

Limitations
Limitations of this study include: unblinded outcome assessment; small sample size (reduced from parent study due to less people completing NeuroCom balance assessment); all participants are from the same urban area; all participants had to be able to access the university campus, thus individuals not able to drive or use public transportation were excluded. Yoga based postural stability exercises used within this treatment regimen were not task specific for dynamic motor capacity. Further dosing and positional selection that would match the intensity requirements for motor capacity improvement may need to be explored.

Conclusion
The results indicate that yoga is likely a viable option as a complement to rehabilitation and recovery after stroke for improvements in postural stability. Specifically, yoga had a positive influence for this cohort on postural symmetry, which is a foundational building block for more dynamic movement capacity. Yoga training should be further explored in a large randomized controlled trial to better understand the potential impact. The longterm impact of consistently applied home programming is worth exploring to determine an impact on ambulatory ability and functional independence.