Lower Limb and Walking

Sitting, Standing and Balance Training

Most individuals with incomplete SCI have the potential to recover some degree of mobility and many functional activities of daily living (ADL) through rehabilitation (McKinley et al. 1999). Proper balance control is important not only for mobility and ambulation, but in fact underlies daily home and community-based functional activities in sitting and standing (Huxham et al. 2001).

SCI is accompanied by changes in sensation, loss of muscle strength, decreased cognitive reserve and spasticity amongst a host of other pathological changes that may lead to balance and gait impairments. Balance and gait impairments in turn, may lead to falls. The incidence of falls in people with SCI has been reported to be as high as 75% with loss of balance being the primary perceived factor contributing to falls in incomplete SCI (Brotherton et al. 2006). Moreover, falls are a major contributor of SCI with falls being the most common cause of SCI in individuals > 60 years old (Dohle and Reding, 2011). It is not currently known the extent to which deficits in balance may affect people with SCI.

The central nervous system (CNS) maintains balance by integrating information from visual, vestibular and sensorimotor systems (Horak and Macpherson, 2006). Greater understanding of the principles underlying the neural control of movement and functional recovery following neurological injury has resulted in increased efforts to design rehabilitation strategies based on task-specific training (Wolpaw and Tennisen, 2001; Behrman and Harkema, 2007; Fouad and Telzaff, 2012). This concept has been translated to various rehabilitation interventions, such as those targeting walking outcomes (e.g. body-weight supported treadmill training) (Mehrholz et al. 2008; Van Hedel and Dietz, 2010; Wessels et al. 2010; Harkema et al. 2012) or arm and hand function (e.g. constraint-induced movement therapy) (Taub et al. 1999; Wolf et al. 2002). For balance, task specific rehabilitation similarly focuses on the achievement of the three main functional goals encompassing balance: 1) maintaining an antigravity posture such as sitting and standing, 2) anticipatory postural control during voluntary self-initiated movements and 3) reactive postural control during an unexpected perturbation (Berg, 1989).

There has been a great deal of focus on the effectiveness of gait training in SCI rehabilitation research (Mehrholz et al., 2008; Van Hedel and Dietz, 2010; Wessels et al. 2010; Harkema et al. 2012), but there has been relatively little attention on the impact of interventions specifically targeting balance outcomes. In other neurologic populations, there is some evidence that task-specific balance training can be effective for improving functional outcomes. A systematic review in people with stroke found moderate evidence that balance could be improved with exercises such as challenging static/dynamic balance ability and practice of balance in different functional tasks, including sitting, standing, walking, and stair climbing (Lubetzky-Vilnai and Kartin, 2010). In recent years, and with the rapid development in technology (e.g., exoskeletons, virtual-reality), there has been more data available about balance outcomes following gait training (Dobkin et al. 2006; Wu et al. 2012; Harkema et al. 2012), as well as specific sitting (Boswell-Ruys et al. 2010; Harvey et al. 2011) or standing (Alexeeva et al. 2011) balance interventions in people with SCI.

Sitting Balance

Acute (< 6 months) SCI

One good quality RCT examined the effect of an additional 3 weeks of task specific exercises on sitting balance in individuals with acute SCI following 6 weeks of standard inpatient rehabilitation consisting of practice of activities of daily living (N=32, AIS A=29, AIS B=2, AIS C=1) (Harvey et al. 2011). Participants were mainly motor complete paraplegics with a median time since injury of 11 weeks. Both experimental and control groups received 6 weeks of standard inpatient rehabilitation consisting of practice of activities of daily living. Despite receiving more training sessions, there was no additional benefit to the experimental group compared to the control group on functional outcomes of sitting balance.

Chronic (> 1 year) SCI

There were 5 studies that investigated the effects of various interventions (i.e. kayak ergometry, task specific exercises in unsupported sitting) on sitting balance. The majority of the participants had motor-complete SCI (N=76, AIS A=51, AIS B=23, AIS C=2). Sitting balance was significantly improved with kayak ergometer training in two Level 4 evidence trials with substantial transfer effects to functional tests in the wheelchair (Bjerkefors and Thorstensson, 2006; Bjerkefors et al. 2007). No significant effect was reported of 8 weeks open water kayak training vs. able-bodied control group who did not train (Grigorenko et al. 2004).

A good quality RCT assessed sitting balance in chronic SCI using the same task specific exercises as the study by Harvey et al (Harvey et al. 2011) in unsupported sitting for 6 weeks vs. a control group who received no training (N=30, AIS A=25, AIS B=15) (Boswell-Ruys et al. 2010). Overall improvements in both the training and control groups were reported. The addition of task-specific exercises using a rocker board to conventional physical therapy for 4 weeks yielded significant improvements in sit and reach tests as well as COP measures (N=12, AIS A=11, AIS B=1) (Kim et al. 2010). However, this was a relatively small study (N = 12) and it did not appear that participants were randomly assigned to the interventions.

Table 2. Sitting Balance

Author Year Country Score Research Design Total Sample Size	Methods	Outcome
Sitting Balance-Acute
Harvey et al. 2011; Australia/Bangledesh RCT PEDro=8 N=32	Population: 32 individuals- 30 males and 2 females; chronic SCI; motor level T1 – L1; 29 AIS A, 2 AIS B, 1 AIS C; age range= 24-31y; years post injury= 8-17 weeks Treatment: In the control group, individuals received 6 weeks standard in patient rehabilitation. In the experimental group, participants received 6 weeks standard in patient rehabilitation + 3 additional 30-minute sessions/wk of 84 task specific exercises with 3 levels of difficulty (252 exercises) in unsupported sitting. Outcome Measures: Maximal Lean Test (Maximal Balance Range), Maximal Sideward Reach Test.	The mean between-group differences for the Maximal Lean Test, Maximal Sideward Reach Test and the Performance Item of the COPM were –20 mm, 5% arm length, and 0.5 points respectively.
Sitting Balance-Chronic (>1 year SCI)
Boswell-Ruys et al. 2010; Australia RCT PEDro=8 N=30	Population: 30 participants- 25 males and 5 females; 25 AIS A, 15 AIS B; level of injury: T1-12; mean age=45y; mean years post injury= 14.5y Treatment: Participants in the experimental group receieved 1hr of 84 task specific exercises with 3 grades of difficulty in an unsupported sitting 3 times a week for 6 weeks. The control group did not receive any intervention. Outcome Measures: Primary measures were: Upper Body Sway Test, Maximal Balance Range Test; Secondary measures were: Alternating Reach test (supported and unsupported), Seated Reach Test 45°to right, Coordinated Stability Test (Version A), Upper Body Sway Test (lateral and antero-posterior components).	The between-group mean difference for the maximal balance range was 64mm.
Kim et al. 2010; Korea Prospective Controlled Trial Level 2 N=12	Population: 12 individuals- 9 males and 3 females; 11 AIS A, 1 AIS B; level of injury: T6-12. mean age= 40.86y Treatment: The control group received conventional PT. The experimental group received conventional PT and goal-oriented training on a rocker board. The patients sat on a stable surface with their legs straight on the floor. Reach forwarrd, left and right, were all measured. Sessions were 5 sets of 10 reps 5 times a week for 4 weeks. Outcome Measures: Modified Functional Reach Test, sway area and sway velocity using the Balance Performance Monitor	There was an increase in the MFRT distance in the experimental group. The experimental group showed a decrease in sway area with both opened and closed eyes after training. The experimental group showed a significant difference before and after training compared to the control, as shown by MFRT distance and swaying area.
Bjerkefors et al. 2006; Sweden Pre-post Level 4 N=10	Population: 10 individuals- 7 males and 3 females; 7 AIS A, 2 AIS B, 1 AIS C; level of injury between T3-12; mean age= 37.6 ± 12y; median years post-injury= 11.5y Treatment: Participants paddled a modified kayak ergometer for 60 minutes 3 times a week for 10 weeks. Outcome Measures: sit and reach tests	Sit and reach tests significantly increased from 3.5cm at baseline to 5.8cm at the end of 10 weeks.
Bjerkefors et al. 2007; Sweden Pre-post Level 4 N=10	Population: 10 individuals- 7 males and 3 females; 7 AIS A, 2 AIS B, 1 AIS C; level of injury between T3-12; mean age= 37.6 ± 12y; median years post-injury= 11.5y Treatment: Participants paddled a modified kayak ergometer for 60 minutes 3 times a week for 10 weeks. Outcome Measures: anterior-posterior (A/P), medio-lateral (M/L) angular and linear and twisting (TW) displacements on support surface translations – forward (FWD), backward (BWD) and lateral (LAT); Kinematic Responses include: I-onset of acceleration (unpredictable), II-constant velocity, III-deceleration (predictable), IV-end of deceleration	A/P angular and linear and TW angular during LAT translations for all kinematic responses were significantly decreased except II for A/P angular M/L angular displacements during LAT translations-significant decrease for kinematic response IV. M/L linear displacement during LAT translations-no significant effects for all kinematic responses.
Grigorenko et al. 2004; Sweden Pre-post Level 4 N=24	Population: Experimental group: 12 individuals- 9 males and 3 females; chronic SCI; 6 AIS A, 5 AIS B, 1 AIS C; level of injury: T2-11; mean age=40y; median years post-injury= 17y; Control group: 12 able bodied participants who did not train Treatment: Participants were involved in 2-3 modified kayak sessions on open water per week for 8 weeks. Outcome Measures: sitting quietly on a force plate-standard deviation (SD), median velocity, median frequency	Small effects in all 3 variables except on the median frequency in the sagittal plane (opposite to becoming normal) Before training and comparing to the control group, all variables differed. Small effects on balance variables-no significant effect.

Standing Balance

Acute (< 6 months) SCI

There was one lower quality RCT that compared BWSTT (experimental group) vs. over ground gait training (control group) in acute (<8 weeks post-injury) incomplete SCI (N=146, N=45 at 6 months; AIS C=38, AIS D=7, 12 weeks of training) (Dobkin et al. 2006) (Table 3). There were no significant differences in balance scores following training between the two groups. However, there was a large median difference between baseline and 6-months post-training in both groups, indicating that for people who were able to walk at six months (N=45), both types of interventions resulted in considerable improvements in balance scores.

Chronic (> 1 year) SCI

Virtual Reality

There were 3 trials that assessed standing balance in people with chronic incomplete SCI using virtual reality (VR) (N=38, AIS C=6, AIS D=20, Table 3). Two studies performed similar VR interventions consisting of standing on a force plate and performing task specific exercises while the center of pressure (COP) position signal was used for visual biofeedback for 4 and 8 weeks respectively (Sayenko et al. 2010; Tamburella et al. 2013). Pre-post studies support the feasibility and reported positive effects on balance function using this approach (Sayenko et al. 2010; Villiger et al. 2013). In the study by Tamburella et al, participants were randomized to receive active training with or without the visual biofeedback. After 8 weeks of training, only the experimental group showed significant improvements in BBS (Baseline= 26.0±10.69 to post-training=41.0±7.8) and Timed Up and Go test (Baseline=21.70±10.70 to post-training=15.22±6.14) (Tamburella et al. 2013).

Body Weight Support Training (BWST)

There were 4 trials that measured standing balance in incomplete SCI following body-weight supported treadmill training. One good quality RCT (Alexeeva et al. 2011), and three studies of level 4 evidence (Musselman et al. 2009; Fritz et al. 2011; Wu et al., 2012) utilized BWST in addition to various interventions such as treadmill, over ground, physical therapist (PT) skills training or a combination of these. After 8 weeks of training, PT skills training resulted in greater balance improvements than BWSTT and BWST on a track (Alexeeva et al. 2011). Small to medium effect sizes for the BBS (0.31-0.67) were reported when BWST was combined with 10 days of intensive mobility training (Fritz et al. 2011). An overall improvement in BBS was found when combining 8 weeks of BWSTT with resistance or assistance but no significant difference between the 2 interventions was reported (Wu et al., 2012).

Four level 4 evidence pre-post trials from the same clinical setting spanning reported significant improvements in balance scores following a program of 3-5 days per week of treadmill- progressing to overground-based BWST (NRN protocol; Behrman et al. 2012; Buehner et al. 2012; Harkema et al. 2012; Lorenz et al. 2012). In addition, this data set demonstrated that the Berg Balance Scale scores were significantly correlated to the severity of injury (Lorenz et al. 2012). Among the pre-post studies, effect sizes were generally small for interventions involving BWST regardless if they were combined with other types of therapy when assessed using the BBS (d=0.18-0.47) with the exception of 1 case-control trial (d=0.88) (Musselman et al. 2009).

There are some drawbacks of measuring balance with only functional outcomes. A ceiling effect was observed in BBS scores of SCI participants receiving BWST with assistance and resistance (Wu et al. 2012). Only one trial was able to use predictable and unpredictable perturbations to assess balance reactions after kayak ergometry training (Bjerkefors et al. 2007). Nonetheless, functional measures are quick, cost effective and easy to apply in both the research and clinical setting and have the added benefit of being validated for the SCI population (Lemay and Nadeau, 2009). It would be optimal that when assessing balance, where feasible, reactions to sudden movements are included in order to give a more comprehensive understanding of balance capacity in persons with SCI.

Table 3. Standing Balance

Author Year Country Score Research Design Total Sample Size	Methods	Outcome
Standing Balance-Acute
Dobkin et al. 2006; USA RCT PEDro=2 N=146	Population: 146 individuals- 116 males and 30 females; 38 AIS, and 7 AIS D; level of injury: C5-L3; median age= 25y; time since injury= 1.03 months Treatment: The control group had 12 weeks of over ground training. The experimental group participated in 12 weeks of body weight support treadmill training. The sessions were 1 hour long for 45-60 sessions. 5 times a week for 12 weeks Outcome Measures: Berg Balance Scale	There were no differences in balance between the overground training group and the body-weight supported treadmill training group.
Standing Balance-Chronic Virtual Reality
Villiger et al. 2013; Switzerland Pre-post Level 4 N=14	Population: 14 individuals- 9 males and 5 females; chronic SCI; 2 AIS C and 12 AIS D; level of injury: C4-T12. mean age= 53y; median years post-injury= 4y Treatment: Participants received 4-5 45-minute sessions of intensive virtual reality augmented training sessions per week for a total of 16-20 sessions. Outcome Measures: Berg Balance Scale	Significant increases were found for all patients in BBS (16.5% increase post treatment and 13% at follow up).
Sayenko et al. 2010; Canada, Japan Pre-post Level 4 N=6	Population: 6 participants- 5 males and 1 female; chronic SCI; 4 AIS C and 2 AIS D; level of injury: C4-T12; mean age= 41y; median years post-injury= 7y Treatment: Patients participated in 3 60-minute visual feedback training sessions, totalling 12 sessions. During training, participants stood on a force platform and were asked to shift their center of pressure (COP) in the indicated directions as represented by a cursor on the monitor. Outcome Measures: Static standing eyes open and closed as measured by COP displacement; Dynamic standing as measured by voluntary COP displacement.	All participants showed substantial improvements in the scores, which varied between 236±94 and 130±14% of the initial values for different exercises. Improvements were all statistically significant for both eyes open and closed except mean velocity in the medial/lateral direction. The balance performance during training-irrelevant tasks was significantly improved: for example, the area inside the stability zone after the training reached 221±86% of the pre-training values.
Tamburella et al. 2013; Italy Open-case study with retrospective matched controls Level 4 N=18	Population: 18 individuals- 9 males and 9 females; chronic SCI; 6 AIS D; level of injury: T9-L5; mean age= 52y; median time since injury: 2.3y Treatment: The control group participated in overground conventional rehabilitation including BWS standing and stepping on a treadmill and overground, balance exercises. The experimental group participated in 40 min of control group protocol and 20 min of specific vBFB (visual biofeedback task specific balance training). The sessions were 60 minutes long 5 times a week for a total of 8 weeks. Outcome Measures: Berg Balance Scale, COP measures, Timed Up and Go (TUG)	At T4, the experimental group saw an improvement in balance aned gait demonstrated by clinical and instrumental evaluation; the improvement was maintained at follow up examinations. In the experimental group, the enhancement in balance that existed at T1 preceded the improvement in gait, and significant correlations between the improvements in gait and balance were observed. In comparison with H data, vBFB treatment demonstrated a significant higher level of effecvtiveness than conventional rehab.
Standing Balance-Chronic Body Weight Support
Alexeeva et al. 2011; USA RCT PEDro=7 Level 1 N=35	Population: 35 individuals- 30 males and 5 females; chornic SCI; 8 AIS C and 27 AIS D; level of injury: C2-T10. mean age= 38.5y; median years post injury= 4y Treatment: Patients participated in a 13-week training program, with three 1 hour sessions per week. The PT group is a structured rehab program individualized for each participant. The TRK group consisted of body weight supported ambulation on a fixed track. The TM group involved body weight supported ambulation on a treadmill. Outcome Measures: Tinetti Balance Scale	All three training groups showed significant improvements in maximal walking speed, muscle strength, and psychological well-being. A significant improvement in balance was seen for PT and TRK groups but not for participants in the TM group.
Alexeeva et al. 2011; USA RCT PEDro=7 Level 1 N=35	Effect Sizes: Forest plot of standardized mean differences (SMD ± 95%C.I.) as calculated from pre- and post-intervention data
Wu et al. 2012; USA Repeated assessment with crossover PEDro=4 Level 2 N=10	Population: 10 individuals- 8 males and 2 females; chronic SCI; all AIS D; level of injury: C2-T10; mean age= 47y; median time since injury=5y Treatment: Group 1 underwent 4 weeks of assistance training then 4 weeks of resistance training. Group 2 underwent 4 weeks of resistance training first, then 4 weeks of assistance training. Resistance was provided by a cable-driven robotic locomotor training system. Sessions were 45 minutes long, 3 times a week for 8 weeks. Outcome Measures: Berg Balance Scale.	A significant improvement in walking speed and balance in humans with SCI was observed after robotic treadmill training using the cable-driven robotic locomotor trainer.
Buehner et al. 2012; USA Prospective Cohort Study Level 2 N=225	Population: 225 individuals- 167 males and 58 females; chronic SCI; 57 AIS C and 167 AIS D; level of injury was not specified; mean age= 42.5y; median time since injury= 2.45y Treatment: NRN Locomotor Training Program consisting of manual-facilitated BWS standing and stepping on a treadmill and overground. Sessions included 1hr of treadmill training, 30 minutes overground assessment, and 15-30 minutes of community reintegration. Sessions were 5 days per week for non ambulators, 4 days per week for ambulators with pronounced assistance, and 3 days week for independent walkers. Outcome Measures: Berg Balance Scale, AIS classification, lower extremity pin prick, light touch and motor scores, 10MWT, 6MWT	Significant gains occurred in lower extremity motor scores. Final Berg Balance Scale scores and initial lower extremity motor scores were positively related. Although 70% of participants showed significantly improved gait speed after locomotor training, only 8% showed AIS category conversion.
Harkema et al. 2012; USA Prospective Cohort Study Level 2 N=196	Population: 169 individuals- 148 males and 48 females; chronic SCI; 66 AIS C and 130 AIS D; level of injury was not specified; mean age= 41y; median time since injury= 0.9y Treatment: NRN Locomotor Training Program consisting of manual-facilitated BWS standing and stepping on a treadmill and overground. Sessions included 1hr of treadmill training, 30 minutes overground assessment, and 15-30 minutes of community reintegration. Sessions were 5 days per week for non ambulators, 4 days per week for ambulators with pronounced assistance, and 3 days week for independent walkers. Outcome Measures: Berg Balance Scale, 6MWT, 10MWT	Outcome measures at enrollment showed high variability between patients with AIS grades C and D. Significant improvement from enrollment to final evaluation was observed in balance and walking measures for patients with AIS grades C and D. The magnitude of improvement significantly differed between AIS groups for all measures. Time since SCI was not associated significantly with outcome measures at enrollment, but was related inversely to levels of improvement.
Lorenz et al. 2012; USA Prospective Cohort Study Level 2 N=337	Population: 337 participants- 255 males and 82 females; chronic SCI; 99 AIS C and 238 AIS D; level of injury: T10 or above; mean age= 40y; median time since injury= 1y Treatment: NRN Locomotor Training Program consisting of manual-facilitated BWS standing and stepping on a treadmill and overground. Sessions included 1hr of treadmill training, 30 minutes overground assessment, and 15-30 minutes of community reintegration. Sessions were 5 days per week for non ambulators, 4 days per week for ambulators with pronounced assistance, and 3 days week for independent walkers. Outcome Measures: Berg Balance Scale, 6MWT, 10MWT.	There was significant improvement on each outcome measure and significant attenuation of improvement over time. Patients varied significantly across groups defined by recovery status and American Spinal Injury Association Impairment Scale (AIS) grade at enrollment with respect to baseline performance and rates of change over time. Time since SCI was a significant determinant of the rate of recovery for all measures.
Behrman et al. 2012; USA Prospective Cohort Study Level 2 N=95	Population: 95 individuals- 75 males and 20 females; chronic SCI; 31 AIS C and 64 AIS D; level of injury= T11 or above; mean age= 43y; median time since injury= 1y; time since injury: <1 yr (n=47), 1-3 yrs (n=24), ³3 yrs (n=24). Treatment: NRN Locomotor Training Program consisting of manual-facilitated BWS standing and stepping on a treadmill and overground. Sessions included 1hr of treadmill training, 30 minutes overground assessment, and 15-30 minutes of community reintegration. Sessions were 5 days per week for non ambulators, 4 days per week for ambulators with pronounced assistance, and 3 days week for independent walkers. Outcome Measures: Berg Balance Scale, 6MWT, 10MWT	Individuals classified within each of the 4 phases of the NRS were functionally discrete, as shown by significant differences in the mean values of balance, gait speed, walking endurance, and the variability of these measurements was significantly reduced by NRS classification. The magnitude of improvements in these outcomes was also significantly different among phase groups.
Fritz et al. 2011; USA Pre-post Level 4 N=15	Population: 15 individuals- 11 males and 4 females; incomplete chronic SCI; Lower functioning group: 10 individuals- 8 males and 2 females; mean age= 38.5y; time since injury: 6.6y; AIS lower extremity score= 24;Higher functioning group: 5 individuals- 3 males and 2 females; mean age= 50.4y; time since injury= 5.7y; AIS lower extremity score= 44 Treatment: Participants received intensive mobility training (IMT) in activities that encouraged repetitive, task specific training of the lower extremities. IMT combines BWSTT, balance exercises, muscle strengthening, coordination and range of motion in a massed intensive therapy. Sessions were 3 hours a day for 3-5 days per week for a total of 10 weeks. Outcome Measures: Berg Balance Scale, Dynamic Gait Index (DGI)	Individuals in the higher functioning ISCI group (BBS score ≥45 and gait speed ≥0.6m/s) spent more time in the intensive therapy on average than individuals in the lower functioning ISCI group. Effect sizes were comparable for changes in balance and mobility assessments between the lower and higher functioning groups, with the largest effect sizes observed for the DGI.
Musselman et al. 2009; Canada Case Series Level 4 N=4	Population: 4 participants- 2 males and 2 females; all AIS C; level of injury: C5-L1; mean age: 44.5y; Gender: median time since injury= 2.7y Treatment: Initial 3 months BWSTT by all 4 patients. Patients 1, 2 received 3 months skills training, followed by 3 months BWSTT. Patients 3, 4 received the training in reverse order. Sessions were 1 hour long, 5 days a week for 3 months. Outcome measures: Berg Balance Scale, Modified Emory Functional Ambulation Profile, 10MWT, 6MWT, Activities- specific Balance Confidence Scale.	Overall improvements in walking speed met or exceeded the minimal clinically important difference for individuals with iSCI (> or = 0.05 m/s), particularly during the skill training phase (skill training: median=0.09 m/s, IQR=0.13; BWSTT: median=0.01 m/s, IQR=0.07). Walking endurance, obstacle clearance, and stair climbing also improved with both types of intervention. Three of the 4 patients had retained their gains at follow-up (retention of walking speed: median=92%, IQR=63%). The findings suggest that skill training was effective in this small group of individuals.

Conclusion

Only preliminary recommendations can be made from the results of this review as there are relatively few studies that provide information on specific balance outcomes in SCI. Early balance training does not appear to enhance the effects of standard physical therapy in either sitting or standing balance. In people with chronic SCI who cannot stand, sitting balance can be improved with both static and dynamic task specific training. For participants with lower severity injuries (e.g., AIS C and D), BWS over ground training combined with physiotherapist-led task-specific exercises and feedback appear to be more effective to improve standing function than BWSTT alone.

There is Level 2 evidence (Dobkin et al. 2006) that there were no differences in balance whether participants engaged in overground training or Body-weight supported treadmill training, but the participants that completed either training and were able to walk made considerable progress in balance.

There is Level 4 evidence (Sayenko et al. 2010; Tamburella et al. 2013) that visual field feedback training leads to substantial improvements in static and dynamic standing eyes open and closed scores, and improvements in balance performance during training-irrelevant tasks.

Tamburella et al. found that the visual biofeedback task specific balance training group saw improvements in balance and gait, and that it demonstrated a significantly higher level of effectiveness than conventional rehabilitation.

There is one study with Level 1 evidence (Alexeeva et al. 2011) and 3 studies with Level 4 evidence (Musselman et al. 2009; Fritz et al. 2011; Wu et al., 2012) that found that BWSTT in addition to Physical Therapy resulted in greater balance improvements than BWSTT alone. Wu et al. found non-significant differences in balance when combining BWSTT with resistance or assistance training.

There is Level 4 evidence (Behrman et al. 2012; Buehner et al. 2012; Harkema et al. 2012; Lorenz et al. 2012) that treadmill and overground based BWST leads to improvements in balance scores.

There is level 1 evidence (Harvey et al. 2011) that task-specific sitting balance exercises for an additional 3 weeks in acute SCI resulted in no difference on balance outcomes.