MEDIOLATERAL (ML) MOVEMENT of the center of

636 ORIGINAL ARTICLE Coordination of Dynamic Balance During Gait Training in People With Acquired Brain Injury Ross Allan Clark, PhD, Gavin Williams, PhD, Natalie Fini, Grad Dip Physio, Liz Moore, BAppSci Physio, Adam Leigh Bryant, PhD ABSTRACT. Clark RA, Williams G, Fini N, Moore L, Bryant AL. Coordination of dynamic balance during gait training in people with acquired brain injury. Arch Phys Med Rehabil 2012;93:636-40. Objective: To investigate movement of the center of mass (COM) during different gait training methods in people with neurologic conditions. Design: Coordination of the gait cycle, represented by mediolateral COM displacement amplitude, timing, and stability, was assessed during a variety of gait training methods performed in a single session. Setting: Gait laboratory. Participants: People who were unable to walk unassisted due to an acquired brain injury (n 17) and healthy control subjects (n 25). Interventions: The participants performed 7 alternative gait training methods in a randomized order. These were therapist manual facilitation, the use of a gait assistive device, treadmill walking with handrail support, and 4 variations of body weight support treadmill training with combinations of handrail and/or therapist support. Main Outcome Measures: Mediolateral COM movement was analyzed in terms of displacement amplitude (overall range of motion), timing (relative to stride time), and stability (steadiness of the movement). Normative values for these measures were acquired from 25 healthy participants walking at a selfselected comfortable pace. Results: Body weight support treadmill training without any additional support resulted in significantly (P.05) greater amplitude, altered timing, and reduced movement stability compared with nonpathologic gait. Allowing handrail support or therapist facilitation reduced this effect and resulted in treadmill training ( body weight support) having lower movement amplitudes when compared with the other training methods. Therapist manual facilitation most closely matched nonpathologic gait for timing and stability. Conclusions: In the context of overall dynamic gait coordination, no single method of training provides the optimal stimulus. A training program that uses a variety of techniques may provide a beneficial rehabilitation response. From the Department of Physiotherapy, The University of Melbourne, Melbourne (Clark, Williams, Bryant); Epworth Hospital, Melbourne (Williams, Moore); and Caulfield Hospital, Alfred Health, Melbourne (Fini), VIC, Australia. Supported by a grant from the Royal Automobile Club of Victoria. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Ross Allan Clark, PhD, Dept of Physiotherapy, The University of Melbourne. 202 Berkeley St, Parkville, VIC, 3010, Australia, e-mail: raclark@unimelb.edu.au. In-press corrected proof published online on Feb 13, 2012, at www.archives-pmr.org. 0003-9993/12/9304-00682$36.00/0 doi:10.1016/j.apmr.2011.11.002 Key Words: Brain injuries; Gait; Rehabilitation. 2012 by the American Congress of Rehabilitation Medicine MEDIOLATERAL (ML) MOVEMENT of the center of mass (COM) during gait is representative of dynamic balance 1 and is often significantly impaired in people with neurologic deficits. 2,3 In isolation, spatial measures of ML COM displacement, such as total range of motion, provide important balance-related motor skill information. However, assessing these measures in combination with other aspects of movement, such as phase timing and pattern analysis, provides greater insight into overall gait coordination. 4 In the context of ML COM movement, a nonpathologic pattern will possess consistent interstride amplitudes and timing of peak displacement 5 and will resemble a wave moving in time with the stride duration. 6 However, in those with neurologic conditions, muscle weakness, spasticity, or interlimb asymmetries may disturb this consistent cyclical pattern. For example, recent evidence suggests that in people with Parkinson s disease, a greater percentage of the COM displacement occurs outside the natural frequency of the movement. 7 The reduced consistency of the COM movement would place greater demands on the neuromotor control systems needed to maintain balance, resulting in a less stable gait pattern. It is highly likely that similar cyclicality deficits would occur in response to a variety of neurologic conditions. Although there is only limited research in the area, it is logical that these factors should be taken into consideration to effectively retrain the gait of someone who possesses a neurologic deficit that has impacted their functional mobility. Comparing the different components of dynamic balance-related gait coordination during a variety of ambulation training strategies, and then contrasting these results with those recorded during nonpathologic gait, may provide insights into the efficacy of each strategy. Commonly performed protocols for training gait fall into 3 main categories: (1) facilitation or assistance provided by a therapist; (2) gait assistive devices, such as sticks, frames, and crutches; and (3) partial body weight support (BWS) provided from a harness suspended overhead. However, as yet, no particular condition has been demonstrated to be superior. 8-12 The aim of this study was to assess the 3 different temporal and spatial components of ML COM displacement during a variety of combinations of these commonly implemented gait training strategies in people with acquired brain injury (ABI). ABI BWS COM ML List of Abbreviations acquired brain injury body weight support center of mass mediolateral

COORDINATION DURING GAIT TRAINING, Clark 637 METHODS Participants Seventeen people (10 men; mean age SD, 38.7 15.3y; mean height SD, 175.0 8.6cm; mean body mass SD, 72.4 22.7kg) with an ABI and who could not walk without assistance were recruited from the rehabilitation units at the Epworth Hospital. Other inclusion criteria were the ability to fully weight bear (for those who had sustained an associated lower-limb fracture) and ability and willingness to provide informed consent. Eleven participants had sustained extremely severe (length of posttraumatic amnesia 28d) 13 traumatic brain injuries (mean, 91.6d). Five participants had sustained a stroke and 1 had multiple sclerosis. The length of time postinjury or diagnosis varied considerably from acute (1mo) to chronic (10y) with a median time of 9 months. 14 Normative data were collected on a population of 25 participants (16 men; mean age SD, 27.8 7.4y; mean height SD, 74.3 7.0cm; mean body mass SD, 70.9 9.8kg) with no prior history of injuries or medical conditions that would negatively influence their gait patterns. Ethical approval was received from the University of Melbourne Human Ethics Review Committee and the Epworth Hospital s Human Research Ethics Committee, and all participants provided informed consent. Data Collection The ABI participants performed 7 gait training methods in a randomized order, summarized as follows: (1) Therapist facilitation (THERAPIST) (2) Gait assistive device (GAIT AID) (3) Treadmill training (TREADMILL) (4) Treadmill training with BWS (BWSTT) (5) Treadmill training with BWS plus therapist assistance (BWSTT T) (6) Treadmill training with BWS plus self-support using their upper limbs (BWSTT UL) (7) Treadmill training with BWS plus self-support using their upper limbs and therapist assistance, that is, methods 5 and 6 combined (BWSTT T UL). BWS was performed using a LiteGait system a and was set at 30% of body weight, because this is a training condition commonly reported in previous literature. 15 Familiarization trials for each gait training method were performed, and for the treadmill-based trials the speed was increased in 0.1m/s increments from 0.1m/s until the participant identified their comfortable walking speed. A single experienced therapist (G.W.) provided all of the therapist facilitation (method 1) and/or assistance (methods 5 and 7) to the participants. A rest period was provided between each method to reduce the effect of fatigue on performance. Normative data collected from the healthy control subjects were acquired during overground walking trials at a self-selected comfortable speed. Similar to previous studies, 16,17 COM was represented by a passive marker located over the sacrum and tracked in 3 dimensions at 120Hz. b Five gait cycles for each condition were extracted for analysis and were normalized to 100 data points per stride to equalize the duration of the gait cycle between conditions. Data Analysis The 3 aspects of ML COM movement examined in this study were the displacement amplitude, timing, and stability. Examples of these measures are presented graphically in figure 1. All analyses were performed for the more affected limb. The ML COM displacement per stride was deemed the mean distance between the most leftward and rightward positions of the COM during ground contact for the more affected limb, with this value representing the displacement amplitude. The timing measure was deemed the time between ground contact and peak ML COM displacement with respect to the stride duration for that limb, a form of discrete relative phase analysis, and was expressed in degrees with 360 representing a full cycle. The zero point for this measure was ground contact for the more affected limb, with higher values indicative of peak ML COM displacement occurring later in the gait cycle. The intrastride stability of ML COM displacement with respect to the stride duration was assessed using frequency analysis. In nonpathologic gait, the ML displacement of the COM follows a pattern that oscillates at the stride frequency, referred to as the fundamental intrinsic harmonic. 6 Displacement at higher frequencies is representative of a less stable gait pattern. 2 This study used a wavelet-based ratio protocol similar to that of Marghitu and Nalluri 18 and Deval et al 19 to detect irregularities in the gait pattern. This ratio was created by comparing the percentage of total signal energy contained within the frequency band of the signal that contains the intrinsic harmonic ( 1.88 times stride frequency) with the remaining signal energy, which represents unnecessary and inefficient movement. Therefore, higher percentage values represent a more stable gait pattern. Statistical Analysis Multiple analysis of variance tests were performed comparing each of the gait training protocols. All analyses were performed with a significance level set at P.05, with Bonferroni adjustment of the P value performed for post hoc comparisons. RESULTS All participants could complete the THERAPIST, BWSTT T, BWSTT UL, and BWSTT T UL trials; however, 4, 2, and 2 participants were unable to perform the BWSTT, GAIT AID, and TREADMILL protocols, respectively, due to either physical inability or risk to safety. Five participants were responsible for these 8 incomplete trials, and therefore the data for these participants were excluded from further analysis. The ML COM displacement amplitude results are provided in figure 2A. A significant main effect (F 9.039, P.001, observed power.993) was observed, with the ABI results significantly higher than nonpathologic gait results. The values for all training strategies were at least twice that of the nonpathologic gait pattern. The BWSTT condition recorded the highest value and was significantly higher than the BWSTT UL, BWSTT T UL, and TREADMILL conditions. The training method with the most support (BWSTT T UL) resulted in the lowest mean COM displacement and was significantly lower than all conditions except BWSTT UL and TREAD- MILL. The timing results are provided in figure 2B. A significant main effect was observed (F 5.212, P.003, observed power.932); however, nonpathologic gait was only significantly different from the BWSTT and BWSTT UL conditions, which were markedly lower (ie, peak COM displacement occurred earlier in the gait cycle). Both BWSTT and BWSTT UL conditions resulted in timing values significantly out of phase with all other training conditions. The stability results are provided in figure 2C. A significant main effect (F 4.808, P.010, observed power.839) was observed, with significantly higher values observed during

638 COORDINATION DURING GAIT TRAINING, Clark Fig 1. ML displacement amplitude, timing, and stability of the COM. Examples of the 3 outcome measures in isolation using simulated data (A, B, and C) and observed in nonpathologic (D) and ataxic (E) gait. (A) Amplitude: larger movements representative of reduced dynamic balance. (B) Timing: an out of phase pattern indicative of poor coordination. (C) Stability: unstable movement pattern with high frequency movement reflective of unsteady displacement of the COM. (D) Nonpathologic gait, collected from a woman participant. This pattern possesses a relatively consistent amplitude and timing and is stable, predominantly fluctuating at a constant frequency with little movement variability. (E) BW- STT data collected during a trial performed by a patient with ataxia due to a traumatic brain injury. This pattern has large displacement, inconsistent timing, and poor stability. nonpathologic gait compared with all gait training strategies with the exception of THERAPIST and TREADMILL. The THERAPIST condition most closely matched the nonpathologic gait pattern, with significantly higher values compared with the BWSTT and BWSTT UL conditions. The BWSTT condition scored worst but was only significantly lower than the THERAPIST and nonpathologic gait conditions. Interestingly, BWSTT UL also resulted in significantly lower values compared with the THERAPIST and TREADMILL conditions. DISCUSSION This study examined coordination of the COM during a variety of gait training methods in people with an ABI and compared these results to people with nonpathologic gait. Of the training strategies included in this study, BWSTT with no other support or facilitation performed poorly. It resulted in large COM displacements, poor timing of COM movement within the gait cycle, and reduced stability. These timing and stability results are consistent with the study of Kyvelidou et al, 20 who observed increased gait variability in healthy people during BWSTT. The overall poor results for the BWSTT condition may also be a factor in the contrasting results of studies assessing the efficacy of BWSTT. 8,21 In regard to the COM displacement results, the treadmill-based protocols that allowed upper-limb support (TREADMILL, BWSTT UL, and BWSTT UL T) were superior to the other training methods. This may provide justification for allowing patients to support themselves using handrails during treadmill-based gait training; however, little is known about how handrail support effects functional outcome. 22 Only the THERAPIST training protocol was not significantly different from nonpathologic gait for timing and stability of COM measures, and it resulted in mean scores within 1SD of nonpathologic gait values, suggesting that this training condition may provide a task-specific training stimulus. The results of this study suggest that treadmill training ( BWS) with handrail support appears to provide a suitable method for task-specific training of dynamic balance because it replicates ML COM amplitudes found in nonpathologic gait. Although it remains unclear whether rehabilitation programs that include treadmill training ( BWS) improve functional mobility more than other methods of gait retraining, 14 the evidence suggests that BWSTT allows for a greater distance to be walked, at an increased gait velocity, and with a higher intensity of activity. For example, in the subacute stroke setting, patients walked distances approximately 5 times greater per 30-minute session when performing BWSTT compared with overground walking in the first week of an intervention. 23 Additionally, during stroke rehabilitation, walking intensity in the absence of BWS is predominantly below the threshold needed to induce cardiorespiratory benefits. 24,25 These factors are strong reasons for the use of treadmill-based training strategies, particularly early in the rehabilitation process. However, if timing and movement stability parameters are considered important components of gait coordination, treadmill training ( BWS) with handrail support may not provide the optimal

COORDINATION DURING GAIT TRAINING, Clark 639 the smooth and stable progression of each individual s COM displacement during gait. These findings may also have been due to the 2-point harness system used, because in patients with poor dynamic balance the ML COM displacement would be abruptly restrained by the harness, therefore causing an earlier and attenuated peak in addition to altering its cyclical nature. An additional limitation is the normative healthy control participant data, which was not individually matched for age, height, weight, or sex to the ABI participants. We did, however, choose a cohort of healthy control subjects who covered a similar range of height and weights to the ABI population, and with a similar sex ratio. We felt that this was sufficient because the aim of this healthy control data was not to attempt to determine what a single participant s COM movement should be, but to provide a range of typical values for each outcome measure. Another limitation is the ABI participant group, which consisted of a variety of neurologic conditions. While this increases the heterogeneity of the cohort, it may more accurately reflect the typical population that is performing gait training in neurologic rehabilitation clinics. Additionally, although this article was written with the assumption that reducing COM displacement amplitude and variability and improving timing that is, attempting to recreate a nonpathologic gait pattern is beneficial during training, it remains unclear whether this actually leads to improved functional mobility. For example, increased lateral displacement of the COM is associated with an increased width of the base of support, yet these factors do not necessarily limit a person with ABI from attaining faster overground gait speeds. 26 This increased width of the base of support compensates for the increased lateral displacement of the COM to allow for maintenance of dynamic balance; however, other factors such as the timing of COM displacement and its variability may not be as simply corrected. Although width of the base of support and lateral displacement of the COM have received considerable attention in studies of people with neurologic gait disturbances, 3,27,28 the other components of gait coordination such as the timing and variability of COM displacement may also contribute to the successful achievement of independent walking, and therefore warrant further investigation. Fig 2. Displacement, timing, and stability for each of the gait training protocols. Mean SD results of the ML COM displacement, timing, and stability assessments for each of the gait training protocols. The horizontal dotted line represents the mean result for the nonpathologic gait pattern, with the gray shaded region the SD. (A) Displacement, (B) timing, and (C) stability. stimulus. Given the diverse results for each of the outcome measures, a rehabilitation program that includes a variety of gait training strategies may be required to achieve this aim. These variable results for each of the outcome measures may also be the reason that, despite the number of studies being performed in the area, no single method of gait training has been shown to be superior. 8-12 Study Limitations These findings are limited to a single assessment and therefore cannot be interpreted to represent a training response. Although limiting the direct applicability of these results to a long-term rehabilitation program, this does provide a potential foundation for the design of gait training programs that aim to improve these key components of dynamic balance. Future studies could assess the longitudinal benefits of prescribing a variety of gait training protocols chosen to specifically target CONCLUSIONS These results indicate that treadmill training ( BWS) with handrail support reduces ML displacement of the COM and postural instability during the gait training session. However, this upper-limb support may come at the cost of altering the timing and variability components of the gait pattern. Careful consideration of these factors should be given when implementing a gait rehabilitation program to achieve adequate postural stability for independent gait; however, further research is required to elucidate the relationship between temporal and spatial measures of COM displacement during gait training and functional mobility outcomes. References 1. Pandy MG, Lin YC, Kim HJ. Muscle coordination of mediolateral balance in normal walking. J Biomech 2010;43:2055-64. 2. Hsue BJ, Miller F, Su FC. The dynamic balance of the children with cerebral palsy and typical developing during gait. Part II: instantaneous velocity and acceleration of COM and COP and their relationship. Gait Posture 2009;29:471-6. 3. Kaufman KR, Brey RH, Chou LS, Rabatin A, Brown AW, Basford JR. Comparison of subjective and objective measurements of balance disorders following traumatic brain injury. Med Eng Phys 2006;28:234-9.

640 COORDINATION DURING GAIT TRAINING, Clark 4. Krasovsky T, Levin MF. Review: toward a better understanding of coordination in healthy and poststroke gait. Neurorehab Neural Repair 2010;24:213-24. 5. Kubo M, Ulrich B. Coordination of pelvis-hat (head, arms and trunk) in anterior-posterior and medio-lateral directions during treadmill gait in preadolescents with/without Down syndrome. Gait Posture 2006;23:512-8. 6. Zijlstra W, Hof AL. Displacement of the pelvis during human walking: experimental data and model predictions. Gait Posture 1997;6:249-62. 7. Lowry KA, Smiley-Oyen AL, Carrel AJ, Kerr JP. Walking stability using harmonic ratios in Parkinson s disease. Mov Disord 2009;24:261-7. 8. Moseley, Stark A, Cameron ID, Pollock A. Treadmill training and body weight support for walking after stroke. Cochrane Database Syst Rev 2005 Oct 19;(4):CD002840. 9. States RA, Salem Y, Pappas E. Overground gait training for individuals with chronic stroke: a Cochrane systematic review. J Neurol Phys Ther 2009;33:179-86. 10. Dickstein R. Rehabilitation of gait speed after stroke: a critical review of intervention approaches. Neurorehabil Neural Repair 2008;22:649-60. 11. Wessels M, Lucas C, Eriks I, de Groot S. Body weight-supported gait training for restoration of walking in people with an incomplete spinal cord injury: a systematic review. J Rehabil Med 2010;42:513-9. 12. Meyer MJ, Megyesi J, Meythaler J, et al. Acute management of acquired brain injury part II: an evidence-based review of pharmacological interventions. Brain Inj 2010;24:706-21. 13. Shores EA, Marosszeky JE, Sandanam J, Batchelor J. Preliminary validation of a clinical scale for measuring the duration of posttraumatic amnesia. Med J Aust 1986;144:569-72. 14. Williams G, Clark R, Schache A, et al. Training conditions influence walking kinematics and self selected walking speed in patients with neurological impairments. J Neurotrauma 2011;28: 281-7. 15. Hesse S. Treadmill training with partial body weight support after stroke: a review. NeuroRehabilitation 2008;23:55-65. 16. Saini M, Kerrigan DC, Thirunarayan MA, Duff-Raffaele M. The vertical displacement of the center of mass during walking: a comparison of four measurement methods. J Biomech Eng 1998; 120:133-9. 17. Eames MH, Cosgrove A, Baker R. Comparing methods of estimating the total body centre of mass in three-dimensions in normal and pathological gaits. Hum Mov Sci 1999;18:637-46. 18. Marghitu DB, Nalluri P. An analysis of greyhound gait using wavelets. J Electromyogr Kinesiol 1997;7:203-12. 19. Delval A, Snijders AH, Weerdesteyn V, et al. Objective detection of subtle freezing of gait episodes in Parkinson s disease. Mov Disord 2010;25:1684-93. 20. Kyvelidou A, Kurz MJ, Ehlers JL, Stergiou N. Aging and partial body weight support affects gait variability. J NeuroEng Rehabil 2008;5:22. 21. Brown TH, Mount J, Rouland BL, Kautz KA, Barnes RM, Kim J. Body weight-supported treadmill training versus conventional gait training for people with chronic traumatic brain injury. J Head Trauma Rehabil 2005;20:402-15. 22. Chen G, Patten C. Treadmill training with harness support: selection of parameters for individuals with poststroke hemiparesis. J Rehabil Res Dev 2006;43:485-97. 23. Ada L, Dean CM, Morris ME, Simpson JM, Katrak P. Randomized trial of treadmill walking with body weight support to establish walking in subacute stroke: the MOBILISE trial. Stroke 2010;41:1237-42. 24. Kuys S, Brauer S, Ada L. Routine physiotherapy does not induce a cardiorespiratory training effect post-stroke, regardless of walking ability. Physiother Res Int 2006;11:219-27. 25. MacKay-Lyons MJ, Makrides L. Cardiovascular stress during a contemporary stroke rehabilitation program: is the intensity adequate to induce a training effect? Arch Phys Med Rehabil 2002; 83:1378-83. 26. Williams G, Morris ME, Schache A, McCrory PR. People preferentially increase hip joint power generation to walk faster following traumatic brain injury. Neurorehab Neural Repair 2010; 24:550-8. 27. Basford JR, Chou LS, Kaufman KR, et al. An assessment of gait and balance deficits after traumatic brain injury. Arch Phys Med Rehabil 2003;84:343-9. 28. Chou LS, Kaufman KR, Walker-Rabatin AE, Brey RH, Basford JR. Dynamic instability during obstacle crossing following traumatic brain injury. Gait Posture 2004;20:245-54. Suppliers a. LiteGait, PO Box 3141, Tempe, AZ 85280. b. Vicon, 14 Minns Business Park, West Way, Oxford OX2 0JB, UK.