- Email: [email protected]
Does walking in a virtual environment induce unstable gait?
Gait & Posture 26 (2007) 289–294 www.elsevier.com/locate/gaitpost
Does walking in a virtual environment induce unstable gait? An examination of vertical ground reaction forces§ John H. Hollman a,*, Robert H. Brey b, Tami J. Bang c, Kenton R. Kaufman c b
a Program in Physical Therapy, Mayo Clinic College of Medicine, Rochester, MN, United States Department of Otorhinolaryngology, Mayo Clinic College of Medicine, Rochester, MN, United States c Department of Orthopedics, Mayo Clinic College of Medicine, Rochester, MN, United States
Received 16 May 2006; received in revised form 12 September 2006; accepted 24 September 2006
Abstract Virtual reality (VR) can induce postural instability in standing and walking, as quantified with kinematic parameters. This study examines the effect of a VR environment on kinetic gait parameters. Ten healthy volunteers walked on an instrumented treadmill in a VR environment and a non-VR environment. In the VR environment, a corridor with colored vertical stripes comprising the walls was projected onto a concave screen placed in front of the treadmill. The speed of the moving image was perceptually equivalent to the speed of the treadmill, creating an illusion that subjects walked through the corridor. Vertical ground reaction forces were sampled. Kinetic parameters that reflect gait stability (weight acceptance peak force, weight acceptance rate, push-off peak force and push-off rate) were compared between the VR and non-VR environments. Subjects walked in the VR environment with increased magnitudes and rates of weight acceptance force and with increased rates of push-off force. Variability in weight acceptance rates and peak forces, and variability in push-off peak forces, were also increased in the VR environment. The gait deviations reflect a compensatory response to visual stimulation that occurs in the VR environment, suggesting that walking in a VR environment may induce gait instability in healthy subjects. # 2006 Elsevier B.V. All rights reserved. Keywords: Virtual reality; Locomotion; Gait stability; Treadmill; Ground reaction force
1. Introduction Kinematic gait parameters are typically used to examine gait instability. Decreased velocity and stride length [1], increased step width [2], increased variability in stride velocity [3] and step width [4], and increased mediolateral displacement of one’s center of mass [5] have all been implicated as markers of gait instability. In elderly individuals in particular, increased variability in stride velocity best predicts risk of falling [3]. While kinematic characteristics of gait instability have been fairly well studied, kinetic markers of instability are neither well § A poster of this study was presented at the 2005 Combined Sections Meeting of the American Physical Therapy Association, New Orleans, LA. * Corresponding author at: Program in Physical Therapy, Mayo School of Health Sciences, Siebens 11-04, 200 First Street SW, Rochester, MN 55905, United States. Tel.: +1 507 284 9547; fax: +1 507 284 0656. E-mail address: [email protected] (J.H. Hollman).
0966-6362/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2006.09.075
studied nor well understood. The ability to control braking forces during weight acceptance and/or propulsive forces during push-off is thought to be important for controlling speed and direction of locomotion [6,7] as well as for adapting gait to environmental constraints [8]. One method for examining braking and propulsive forces is through measuring the ground reaction force (GRF) during locomotion. The GRF represents a summation of forces from all segments of the body during locomotion and can provide a representation of center of mass (COM) control. Since postural stability is defined as the body’s ability to maintain its COM within its base of support [9], the GRF may therefore provide information about a person’s postural control during locomotion. Indeed, studies that have investigated GRFs during walking suggest that peak forces during the weight acceptance and push-off phases of the gait cycle increase in magnitude [10] and in variability [11,12] in people assumed to be less stable during locomotion.
290
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
In an emerging area of study, researchers are examining locomotion in virtual reality (VR) [13–16]. With VR-like stimuli, researchers [17,18] have examined the influences of visual, vestibular and somatosensory input in the control of human locomotion. The interaction between the three sensory systems on gait is complex and interdependent. However, Varraine et al. [18] suggest that when visual stimuli are used to induce optic flow (the perceived movement of visual scenes on the retina as one moves through space), the visual system is primarily responsible for control of gait kinematics. In contrast, Varraine et al. [18] suggest that the somatosensory system, while modulated by the visual system, primarily participates in the control of gait kinetics. Regardless, it is evident that visual system function has an important role in influencing locomotion, the hypothesis of which has been confirmed by researchers [14,15] who have shown that visual stimuli provided by VR are capable of disturbing whole-body equilibrium, resulting in impaired balance and altered gait. Yet, VR has been an effective treatment tool for disorders such as fear of flying and fear of heights [19,20]. VR enables a clinician to use graded exposure protocols designed to simulate – in controlled settings – the environments that precipitate patients’ symptoms. Using VR may enable patients to progressively adapt and compensate to the stimuli. VR, therefore, is being examined for its use in rehabilitation in people with vestibular system pathology [21]. Nevertheless, VR environments seem to induce instability during standing, as measured by postural sway [22], and during walking, as measured by kinematic gait [16] and trunk movement parameters [14]. The effect of walking in a VR environment on kinetic parameters of gait, however, is not well understood. Given the relative lack of research investigating kinetic markers of gait instability and given the emerging interest in studying gait instability in VR environments, this study examined vertical GRFs in people walking on a treadmill in VR and non-VR environments. The purpose of the study was to examine whether the magnitude and variability of weight acceptance peak force, weight acceptance rate, push-off peak force, and push-off rate during walking differ between VR and non-VR environments. We hypothesized that these variables would increase when people walked on a treadmill in a VR environment.
alcohol or caffeine on the day of testing were excluded from the subject pool. All subjects signed an informed consent form approved by the Mayo Foundation Institutional Review Board, Rochester, MN. 2.2. Instrumentation The virtual reality environment was created using an Elumens VisionStation1 1024 (Elumens Corporation, Durham, NC) projection system. The VisionStation1 system consisted of a 162-cm diameter concave screen that spanned 1608 of the subject’s field of vision. A projector with a modified lens to disperse light in an arc of 1808 was used to project the image on the screen. The VisionStation1 system was mounted above and in front of a Type 9810S1 GaitwayTM Instrumented Treadmill (Kistler Instrument Corporation, Amherst, NY) equipped with tandem piezoelectric force plates beneath the treadmill belt (Fig. 1A). The force plates measured vertical GRFs and GaitwayTM software used the GRF data to calculate center of pressure data. Data were sampled at 100 Hz. The GaitwayTM software discriminated between right- and left-footsteps and allowed raw data to be exported for further analysis. 2.3. Procedures A virtual corridor with patterned walls was projected onto the screen to create the VR environment (Fig. 1B). The image of the corridor was programmed to rotate at an angular velocity such that the linear translation of optic flow at the midpoint of the corridor was perceptually equivalent to the speed of the treadmill. Doing so created the illusion that the subject was walking through the corridor while walking on the treadmill. In the non-VR environment, the corridor was displayed on the screen, but the image was stationary. Subjects walked on the treadmill for 3 min in both environments (VR and non-VR) at a speed of 1.3 m/s— considered a typical walking velocity in individuals represented by the study sample [23]. The order of testing was counterbalanced across conditions to minimize potential
2. Methods 2.1. Subjects Ten healthy adults (three males, seven females) participated in the study (mean age = 25 3 years). All subjects were self-reported to be in good health. All reported an ability to walk over even and uneven terrain without the use of assistive devices. Potential subjects with a history of lower-extremity surgery or history of neuromuscular or cardiovascular pathology were excluded from participating in the study. Additionally, individuals having consumed
Fig. 1. (A) Set-up of the instrumented treadmill and VisionStation1 projection system and (B) a close-up view of the virtual corridor (adapted from Hollman et al. [16]).
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
291
The dependent variables (weight acceptance peak force, weight acceptance rate, push-off peak force, push-off rate, and their respective measures of variability) were compared between the VR and non-VR environments. Since the research hypothesis was directional in nature (i.e., we hypothesized that magnitude and variability of the dependent variables would increase in the VR environment), the tests were conducted as one-tailed paired t-tests. Statistical significance was established at p < 0.05 for all tests. Data were analyzed with SPSS 10.0 statistical software (SPSS Inc., Chicago, IL). Fig. 2. Example of the vertical ground reaction force through a single footfall. Weight acceptance peak force is defined as the peak force measured during weight acceptance phase of the gait cycle. Weight acceptance rate is defined as the rate of change (DF/Dt) of the ground reaction force between initial contact and weight acceptance peak force. Push-off peak force is defined as the peak force during the push-off phase of the gait cycle. Pushoff rate is defined as the rate of change (DF/Dt) of the ground reaction force between push-off peak force and toe-off. GaitwayTM software calculates these parameters between 10% and 90% of the peaks.
3. Results Peak forces and rates of force application at weight acceptance and push-off, and variability in the GRFs, were generally greater in the VR environment as compared to the non-VR environment (Fig. 3). Detailed results follow. 3.1. Weight acceptance
for order effects to confound the study’s results. During each trial, data were sampled continuously for a 20-s interval after subjects achieved steady-state ambulation. 2.4. Data reduction and analysis To measure gait instability through changes in kinetic data, the following time-series GRF variables were examined: weight acceptance peak force, weight acceptance rate, push-off peak force, and push-off rate (Fig. 2). The following variability data were calculated: variability of weight acceptance peak force, variability of weight acceptance rate, variability of push-off peak force, and variability of push-off rate. Variability data for each subject were calculated as the percentage coefficient of variation (% CV) across all strides collected within each 20-s sampling period. CV data were calculated with the equation: S:D: % CV ¼ ¯ 100 X where S.D. is the standard deviation and X¯ is the mean of distribution.
Weight acceptance peak forces (Fig. 4) were significantly greater in the VR environment (mean = 758 187 N) than in the non-VR environment (mean = 738 181; t9 = 3.718, p = 0.005). Variability in weight acceptance peak force was significantly greater in the VR environment (3.1% CV) than the non-VR environment (2.5% CV; t9 = 1.973, p = 0.040). Weight acceptance rates (Fig. 5) were significantly greater in the VR environment (mean = 5739 891 N/s) than in the non-VR environment (5421 1214 N/s; t9 = 2.210, p = 0.027). Variability in weight acceptance rate was significantly greater in the VR environment (12.1% CV) than in the non-VR environment (9.9% CV; t9 = 2.093, p = 0.033). 3.2. Push-off Push-off peak forces (Fig. 4) were greater in the VR environment (mean = 762 181 N) than the non-VR environment (mean = 753 176 N). The differences, however, were not statistically significant (t9 = 1.639, p = 0.078). Variability in push-off peak force was significantly greater in
Fig. 3. Average vertical ground reaction forces obtained from a subject walking in the normal (non-VR) and virtual reality (VR) environments. Increased variability in the ground reaction forces occurred while subjects walked in the VR environment (note standard deviation bars).
292
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
Fig. 4. Weight acceptance and push-off peak forces (mean and S.E.) in the normal (non-VR) and virtual reality (VR) environments (*p < 0.05).
Fig. 5. Weight acceptance and push-off rates (mean and S.E.) in the normal (non-VR) and virtual reality (VR) environments (*p < 0.05).
the VR environment (3.6% CV) than in the non-VR environment (2.2% CV; t9 = 2.653, p = 0.013). Push-off rates (Fig. 5) were significantly greater in the VR environment (6924 1693 N/s) than in the non-VR environment (6582 1583 N/s; t9 = 3.549, p = 0.003). Variability in push-off rate was greater in the VR environment (6.4% CV) than in the non-VR environment (5.1% CV). The difference in variability, however, was not statistically significant (t9 = 1.480, p = 0.087).
4. Discussion Results support our hypothesis that kinetic gait parameters indicative of gait instability change in a VR environment. Weight acceptance peak forces, weight acceptance rates and push-off rates all increased significantly in the VR environment, as did variability in weight acceptance rate, push-off peak force and push-off rate. It is clear that multiple changes in kinetic gait parameters occurred in the VR environment. In contrast to the assertion that optic flow has little role in controlling force control during locomotion [18], our data suggest that visual stimuli from a VR environment, even one intended to provide an optic flow that is perceptually consistent with walking speed, induces changes in kinetic parameters of gait. The increased weight acceptance forces, increased weight acceptance and push-off rates, and in particular the variability of the kinetic parameters may reflect a motor response necessary to compensate for an optic flow-induced threat to stability.
Few studies have examined kinetic gait parameters related to gait instability. Giakas et al. [11] reported significantly greater variability in weight acceptance peak forces and push-off peak forces in subjects with scoliosis as compared to control subjects. They also reported significantly higher variability for medial–lateral and anterior–posterior shear forces. Among their conclusions, the researchers reported that people with scoliosis presented with balance dysfunction, exhibited by higher variability in the kinetic gait parameters than a control population. Similarly, White et al. [24] reported higher variability in weight acceptance and push-off peak forces in the gait of patients with cerebral palsy. Results of the present study have similarities to those reported by Giakas et al. [11] and White et al. [24] implying that walking in a VR environment induces gait instability. Postural stability is defined as an ability to maintain the body’s center of mass (COM) over the base of support [9]. Instability during walking may therefore manifest as increased displacement of the body’s COM within a constantly changing base of support. The increased COM displacement that occurs in unstable gait may necessitate that people compensate by making adjustments to their base of support. Therefore, the classic spatiotemporal gait deviations that characterize gait instability including reduced stride length, increased step width, and increased variability in stride velocity and step width may occur [16]. The kinetic control required to make such adjustments may be observed by changes in the GRF during locomotion. For example, if the body’s COM displaces anteriorly over the base of support, downward and forward momentum may increase [10]. To compensate for a forward-falling tendency, the magnitudes of weight acceptance and push-off peak forces would increase with concomitant increases in the push-off and weight acceptance rates. Without center of mass data, it is difficult to confirm these hypotheses. However, based on interpretations of previous studies [10,11,24], the increases in magnitude and variability of the weight acceptance and push-off forces observed in the present study are indicative of gait instability induced by the VR environment. Results also suggest that VR influences gait at a central, more than peripheral, level of processing. While vestibular, visual and somatosensory systems all contribute to one’s ability to maintain postural stability during locomotion, the sensory stimulus manipulated in this study was visual only. As per the design of this study, somatosensory cues were not manipulated nor were vestibular cues. It is possible, however, that a conflict between peripheral (somatosensory) and central (visual) sensory systems was induced. The VR environment used in this study was presented as a curved corridor. While subjects were required to walk along a straight path on the treadmill, they likely perceived themselves to be walking along a curved path. The somatosensory and visual systems may have been providing a different reference frame for the control of gait in this
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
study. Regardless, whether the visual system in isolation or in interaction with the somatosensory system induced the gait changes, it is evident that VR influences gait at a central level of processing. Regarding the clinical relevance of the present study, people often have impaired balance and impaired gait stability in visually stimulating environments following vestibular system injury [17]. While in its infancy, VR is being investigated for its use in rehabilitation for persons with vestibular system pathology who often report difficulty walking while simultaneously moving their heads to attend to peripheral activities [21,25]. For example, patients may describe difficulty walking down a grocery store aisle while looking back and forth at products or when walking in a crowded environment. One method used in vestibular rehabilitation is to create situations that progressively stimulate patients’ symptoms in order to promote habituation and more rapid progress through their rehabilitation. Whitney et al. [21] were among the first investigators to examine VR for patients with vestibular pathology. They studied postural stability in standing activities only. Nevertheless, their study represents a clinical use of VR for rehabilitation purposes. Potentially, the use of VR with a treadmill, as used in the present study, may allow therapists to target situations that cause patients’ symptoms during walking – in a controlled clinical setting – to promote habituation. Some limitations of the study may restrict interpretations of the results. First, gait in able-bodied, healthy adults was examined. The results obtained with this sample may or may not reflect the effect of a VR environment on individuals with vestibular pathology or other disorders that affect gait. Second, treadmills may not necessarily mimic overground walking. While some research suggests that treadmill and overground walking are comparable and that a treadmill is a valid laboratory instrument to measure gait [26], other evidence suggests that gait differs between treadmill and overground walking [27]. It is not certain how results from the present study might translate to overground walking conditions. Third, the force plates in the treadmill bed were only capable of measuring vertical GRF data. Because shear forces (anterior–posterior, medial–lateral) were not measured, the force data may not entirely represent a subject’s kinetic walking patterns. Finally, the use of VR is relatively unexplored in gait studies—‘‘normal’’ gait in VR environments is still, for the most part, undetermined. Despite these limitations, the evidence strongly suggests that the VR environment used in the present study has a significant effect on kinetic gait parameters in healthy individuals walking on a treadmill. Changes in the weight acceptance and push-off phases of the gait cycle observed in the VR environment likely represent compensations for unstable gait. Future related research should examine the effect of VR environments on gait in individuals with vestibular system pathology, the responsiveness of gait stability markers as patients with vestibular system pathology progress through
293
vestibular rehabilitation programs, and the efficacy of VR as a rehabilitation tool. Additional future work may involve different virtual environments, including scenarios that better represent daily activities, such as a VR environment that more closely simulates walking down a grocery store aisle. Such studies may shed light on potential interventions for vestibular dysfunction.
5. Conclusion In summary, walking in a VR environment induces changes in gait that reflect gait instability. Subjects walked in the VR environment with increased weight acceptance peak forces and with increased weight acceptance and pushoff rates. Additionally, subjects walked in the VR environment with increased variability in their weight acceptance rates and with increased variability in the magnitude of their weight acceptance peak and push-off peak forces. Changes in these kinetic gait parameters may be a manifestation of compensatory efforts to control the body’s center of mass over the base of support during locomotion and therefore represent gait instability induced by visual stimulation in a VR environment.
Acknowledgements The authors thank Richard A. Robb, PhD, for software development of the virtual environment image, and Steve Irby, MS, for technical assistance. Donald M. Kendall provided funding to support this study.
References [1] Menz HB, Lord SR, Fitzpatrick RC. Age-related differences in walking stability. Age Ageing 2003;32:137–42. [2] Krebs DE, Goldvasser D, Lockert JD, Portney LG, Gill-Body KM. Is base of support greater in unsteady gait? Phys Ther 2002;82: 138– 47. [3] Maki BE. Gait changes in older adults: predictors of falls or indicators of fear. J Am Geriatr Soc 1997;45:313–20. [4] Bauby CE, Kuo AD. Active control of lateral balance in human walking. J Biomech 2000;33:1433–40. [5] Chou LS, Kaufman KR, Hahn ME, Brey RH. Medio-lateral motion of the center of mass during obstacle crossing distinguishes elderly individuals with imbalance. Gait Posture 2003;18:125–33. [6] Hase K, Stein RB. Analysis of rapid stopping during human walking. J Neurophysiol 1998;80:255–61. [7] Hase K, Stein RB. Turning strategies during human walking. J Neurophysiol 1999;81:2914–22. [8] Varraine E, Bonnard M, Pailhous J. Intentional on-line adaptation of stride length in human walking. Exp Brain Res 2000;130:248–57. [9] Horak FB. Clinical measurement of postural control in adults. Phys Ther 1987;67:1881–5. [10] Hsiang SM, Chang CC. The effect of gait speed and load carrying on the reliability of ground reaction forces. Saf Sci 2002;40:639–57. [11] Giakas G, Baltzopoulos V, Dangerfield PH, Dorgan JC, Dalmira S. Comparison of gait patterns between healthy and scoliotic patients
294
[12]
[13]
[14]
[15]
[16]
[17] [18]
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294 using time and frequency domain analysis of ground reaction forces. Spine 1996;21:2235–42. Giakas G, Baltzopoulos V. Time and frequency domain analysis of ground reaction forces during walking: an investigation of variability and symmetry. Gait Posture 1997;5:189–97. Buekers M, Montagne G, de Rugy A, Laurent M. The regulation of externally paced human locomotion in virtual reality. Neurosci Lett 1999;275:171–4. Richards JT, Mulavara AP, Bloomberg JJ. Postural stability during treadmill locomotion as a function of the visual polarity and rotation of a three-dimensional virtual environment. Presence-Teleop Virt 2004;13:371–84. Mulavara AP, Richards JT, Ruttley T, Marshburn A, Nomura Y, Bloomberg JJ. Exposure to a rotating virtual environment during treadmill locomotion causes adaptation in heading direction. Exp Brain Res 2005;16:210–9. Hollman JH, Brey RH, Robb RA, Bang TJ, Kaufman KR. Spatiotemporal gait deviations in a virtual reality environment. Gait Posture 2006;23:441–4. Redfern MS, Yardley L, Bronstein AM. Visual influences on balance. J Anxiety Disord 2001;15:81–94. Varraine E, Bonnard M, Pailhous J. Interaction between different sensory cues in the control of human gait. Exp Brain Res 2002;142: 374–84.
[19] Rothbaum BO, Hodges LF, Kooper R, Opdyke D, Williford JS, North M. Effectiveness of computer-generated (virtual reality) graded exposure in the treatment of acrophobia. Am J Psychiatry 1995;152:626–8. [20] Rothbaum BO, Hodges L, Smith S, Lee JH, Price L. A controlled study of virtual reality exposure therapy for the fear of flying. J Consult Clin Psychol 2000;68:1020–6. [21] Whitney SL, Sparto PJ, Brown KE, Furman JM, Jacobson JL, Redfern MS. The potential use of virtual reality in vestibular rehabilitation: preliminary findings with the BNAVE. Neurol Rep 2002;26:72–8. [22] Akiduki H, Nishiike S, Watanabe H, Matsuoka K, Kubo T, Takeda N. Visual-vestibular conflict induced by virtual reality in humans. Neurosci Lett 2003;340:197–200. [23] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower extremity kinematics during level walking. J Orthop Res 1990;8:383–92. [24] White R, Agouris I, Selbie RD, Kirkpatrick M. The variability of force platform data in normal and cerebral palsy gait. Clin Biomech 1999;14:185–92. [25] Viirre E. Virtual reality and the vestibular apparatus. IEEE Eng Med Biol Mag 1996;15(41–43):69. [26] Belli A, Bui P, Berger A, Geyssant A, Lacour JR. A treadmill ergometer for three-dimensional ground reaction forces measurement during walking. J Biomech 2001;34:105–12. [27] Alton F, Baldey L, Caplan S, Morrissey MC. A kinematic comparison of overground and treadmill walking. Clin Biomech 1998;13:434–40.
Does walking in a virtual environment induce unstable gait? An examination of vertical ground reaction forces§ John H. Hollman a,*, Robert H. Brey b, Tami J. Bang c, Kenton R. Kaufman c b
a Program in Physical Therapy, Mayo Clinic College of Medicine, Rochester, MN, United States Department of Otorhinolaryngology, Mayo Clinic College of Medicine, Rochester, MN, United States c Department of Orthopedics, Mayo Clinic College of Medicine, Rochester, MN, United States
Received 16 May 2006; received in revised form 12 September 2006; accepted 24 September 2006
Abstract Virtual reality (VR) can induce postural instability in standing and walking, as quantified with kinematic parameters. This study examines the effect of a VR environment on kinetic gait parameters. Ten healthy volunteers walked on an instrumented treadmill in a VR environment and a non-VR environment. In the VR environment, a corridor with colored vertical stripes comprising the walls was projected onto a concave screen placed in front of the treadmill. The speed of the moving image was perceptually equivalent to the speed of the treadmill, creating an illusion that subjects walked through the corridor. Vertical ground reaction forces were sampled. Kinetic parameters that reflect gait stability (weight acceptance peak force, weight acceptance rate, push-off peak force and push-off rate) were compared between the VR and non-VR environments. Subjects walked in the VR environment with increased magnitudes and rates of weight acceptance force and with increased rates of push-off force. Variability in weight acceptance rates and peak forces, and variability in push-off peak forces, were also increased in the VR environment. The gait deviations reflect a compensatory response to visual stimulation that occurs in the VR environment, suggesting that walking in a VR environment may induce gait instability in healthy subjects. # 2006 Elsevier B.V. All rights reserved. Keywords: Virtual reality; Locomotion; Gait stability; Treadmill; Ground reaction force
1. Introduction Kinematic gait parameters are typically used to examine gait instability. Decreased velocity and stride length [1], increased step width [2], increased variability in stride velocity [3] and step width [4], and increased mediolateral displacement of one’s center of mass [5] have all been implicated as markers of gait instability. In elderly individuals in particular, increased variability in stride velocity best predicts risk of falling [3]. While kinematic characteristics of gait instability have been fairly well studied, kinetic markers of instability are neither well § A poster of this study was presented at the 2005 Combined Sections Meeting of the American Physical Therapy Association, New Orleans, LA. * Corresponding author at: Program in Physical Therapy, Mayo School of Health Sciences, Siebens 11-04, 200 First Street SW, Rochester, MN 55905, United States. Tel.: +1 507 284 9547; fax: +1 507 284 0656. E-mail address: [email protected] (J.H. Hollman).
0966-6362/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2006.09.075
studied nor well understood. The ability to control braking forces during weight acceptance and/or propulsive forces during push-off is thought to be important for controlling speed and direction of locomotion [6,7] as well as for adapting gait to environmental constraints [8]. One method for examining braking and propulsive forces is through measuring the ground reaction force (GRF) during locomotion. The GRF represents a summation of forces from all segments of the body during locomotion and can provide a representation of center of mass (COM) control. Since postural stability is defined as the body’s ability to maintain its COM within its base of support [9], the GRF may therefore provide information about a person’s postural control during locomotion. Indeed, studies that have investigated GRFs during walking suggest that peak forces during the weight acceptance and push-off phases of the gait cycle increase in magnitude [10] and in variability [11,12] in people assumed to be less stable during locomotion.
290
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
In an emerging area of study, researchers are examining locomotion in virtual reality (VR) [13–16]. With VR-like stimuli, researchers [17,18] have examined the influences of visual, vestibular and somatosensory input in the control of human locomotion. The interaction between the three sensory systems on gait is complex and interdependent. However, Varraine et al. [18] suggest that when visual stimuli are used to induce optic flow (the perceived movement of visual scenes on the retina as one moves through space), the visual system is primarily responsible for control of gait kinematics. In contrast, Varraine et al. [18] suggest that the somatosensory system, while modulated by the visual system, primarily participates in the control of gait kinetics. Regardless, it is evident that visual system function has an important role in influencing locomotion, the hypothesis of which has been confirmed by researchers [14,15] who have shown that visual stimuli provided by VR are capable of disturbing whole-body equilibrium, resulting in impaired balance and altered gait. Yet, VR has been an effective treatment tool for disorders such as fear of flying and fear of heights [19,20]. VR enables a clinician to use graded exposure protocols designed to simulate – in controlled settings – the environments that precipitate patients’ symptoms. Using VR may enable patients to progressively adapt and compensate to the stimuli. VR, therefore, is being examined for its use in rehabilitation in people with vestibular system pathology [21]. Nevertheless, VR environments seem to induce instability during standing, as measured by postural sway [22], and during walking, as measured by kinematic gait [16] and trunk movement parameters [14]. The effect of walking in a VR environment on kinetic parameters of gait, however, is not well understood. Given the relative lack of research investigating kinetic markers of gait instability and given the emerging interest in studying gait instability in VR environments, this study examined vertical GRFs in people walking on a treadmill in VR and non-VR environments. The purpose of the study was to examine whether the magnitude and variability of weight acceptance peak force, weight acceptance rate, push-off peak force, and push-off rate during walking differ between VR and non-VR environments. We hypothesized that these variables would increase when people walked on a treadmill in a VR environment.
alcohol or caffeine on the day of testing were excluded from the subject pool. All subjects signed an informed consent form approved by the Mayo Foundation Institutional Review Board, Rochester, MN. 2.2. Instrumentation The virtual reality environment was created using an Elumens VisionStation1 1024 (Elumens Corporation, Durham, NC) projection system. The VisionStation1 system consisted of a 162-cm diameter concave screen that spanned 1608 of the subject’s field of vision. A projector with a modified lens to disperse light in an arc of 1808 was used to project the image on the screen. The VisionStation1 system was mounted above and in front of a Type 9810S1 GaitwayTM Instrumented Treadmill (Kistler Instrument Corporation, Amherst, NY) equipped with tandem piezoelectric force plates beneath the treadmill belt (Fig. 1A). The force plates measured vertical GRFs and GaitwayTM software used the GRF data to calculate center of pressure data. Data were sampled at 100 Hz. The GaitwayTM software discriminated between right- and left-footsteps and allowed raw data to be exported for further analysis. 2.3. Procedures A virtual corridor with patterned walls was projected onto the screen to create the VR environment (Fig. 1B). The image of the corridor was programmed to rotate at an angular velocity such that the linear translation of optic flow at the midpoint of the corridor was perceptually equivalent to the speed of the treadmill. Doing so created the illusion that the subject was walking through the corridor while walking on the treadmill. In the non-VR environment, the corridor was displayed on the screen, but the image was stationary. Subjects walked on the treadmill for 3 min in both environments (VR and non-VR) at a speed of 1.3 m/s— considered a typical walking velocity in individuals represented by the study sample [23]. The order of testing was counterbalanced across conditions to minimize potential
2. Methods 2.1. Subjects Ten healthy adults (three males, seven females) participated in the study (mean age = 25 3 years). All subjects were self-reported to be in good health. All reported an ability to walk over even and uneven terrain without the use of assistive devices. Potential subjects with a history of lower-extremity surgery or history of neuromuscular or cardiovascular pathology were excluded from participating in the study. Additionally, individuals having consumed
Fig. 1. (A) Set-up of the instrumented treadmill and VisionStation1 projection system and (B) a close-up view of the virtual corridor (adapted from Hollman et al. [16]).
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
291
The dependent variables (weight acceptance peak force, weight acceptance rate, push-off peak force, push-off rate, and their respective measures of variability) were compared between the VR and non-VR environments. Since the research hypothesis was directional in nature (i.e., we hypothesized that magnitude and variability of the dependent variables would increase in the VR environment), the tests were conducted as one-tailed paired t-tests. Statistical significance was established at p < 0.05 for all tests. Data were analyzed with SPSS 10.0 statistical software (SPSS Inc., Chicago, IL). Fig. 2. Example of the vertical ground reaction force through a single footfall. Weight acceptance peak force is defined as the peak force measured during weight acceptance phase of the gait cycle. Weight acceptance rate is defined as the rate of change (DF/Dt) of the ground reaction force between initial contact and weight acceptance peak force. Push-off peak force is defined as the peak force during the push-off phase of the gait cycle. Pushoff rate is defined as the rate of change (DF/Dt) of the ground reaction force between push-off peak force and toe-off. GaitwayTM software calculates these parameters between 10% and 90% of the peaks.
3. Results Peak forces and rates of force application at weight acceptance and push-off, and variability in the GRFs, were generally greater in the VR environment as compared to the non-VR environment (Fig. 3). Detailed results follow. 3.1. Weight acceptance
for order effects to confound the study’s results. During each trial, data were sampled continuously for a 20-s interval after subjects achieved steady-state ambulation. 2.4. Data reduction and analysis To measure gait instability through changes in kinetic data, the following time-series GRF variables were examined: weight acceptance peak force, weight acceptance rate, push-off peak force, and push-off rate (Fig. 2). The following variability data were calculated: variability of weight acceptance peak force, variability of weight acceptance rate, variability of push-off peak force, and variability of push-off rate. Variability data for each subject were calculated as the percentage coefficient of variation (% CV) across all strides collected within each 20-s sampling period. CV data were calculated with the equation: S:D: % CV ¼ ¯ 100 X where S.D. is the standard deviation and X¯ is the mean of distribution.
Weight acceptance peak forces (Fig. 4) were significantly greater in the VR environment (mean = 758 187 N) than in the non-VR environment (mean = 738 181; t9 = 3.718, p = 0.005). Variability in weight acceptance peak force was significantly greater in the VR environment (3.1% CV) than the non-VR environment (2.5% CV; t9 = 1.973, p = 0.040). Weight acceptance rates (Fig. 5) were significantly greater in the VR environment (mean = 5739 891 N/s) than in the non-VR environment (5421 1214 N/s; t9 = 2.210, p = 0.027). Variability in weight acceptance rate was significantly greater in the VR environment (12.1% CV) than in the non-VR environment (9.9% CV; t9 = 2.093, p = 0.033). 3.2. Push-off Push-off peak forces (Fig. 4) were greater in the VR environment (mean = 762 181 N) than the non-VR environment (mean = 753 176 N). The differences, however, were not statistically significant (t9 = 1.639, p = 0.078). Variability in push-off peak force was significantly greater in
Fig. 3. Average vertical ground reaction forces obtained from a subject walking in the normal (non-VR) and virtual reality (VR) environments. Increased variability in the ground reaction forces occurred while subjects walked in the VR environment (note standard deviation bars).
292
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
Fig. 4. Weight acceptance and push-off peak forces (mean and S.E.) in the normal (non-VR) and virtual reality (VR) environments (*p < 0.05).
Fig. 5. Weight acceptance and push-off rates (mean and S.E.) in the normal (non-VR) and virtual reality (VR) environments (*p < 0.05).
the VR environment (3.6% CV) than in the non-VR environment (2.2% CV; t9 = 2.653, p = 0.013). Push-off rates (Fig. 5) were significantly greater in the VR environment (6924 1693 N/s) than in the non-VR environment (6582 1583 N/s; t9 = 3.549, p = 0.003). Variability in push-off rate was greater in the VR environment (6.4% CV) than in the non-VR environment (5.1% CV). The difference in variability, however, was not statistically significant (t9 = 1.480, p = 0.087).
4. Discussion Results support our hypothesis that kinetic gait parameters indicative of gait instability change in a VR environment. Weight acceptance peak forces, weight acceptance rates and push-off rates all increased significantly in the VR environment, as did variability in weight acceptance rate, push-off peak force and push-off rate. It is clear that multiple changes in kinetic gait parameters occurred in the VR environment. In contrast to the assertion that optic flow has little role in controlling force control during locomotion [18], our data suggest that visual stimuli from a VR environment, even one intended to provide an optic flow that is perceptually consistent with walking speed, induces changes in kinetic parameters of gait. The increased weight acceptance forces, increased weight acceptance and push-off rates, and in particular the variability of the kinetic parameters may reflect a motor response necessary to compensate for an optic flow-induced threat to stability.
Few studies have examined kinetic gait parameters related to gait instability. Giakas et al. [11] reported significantly greater variability in weight acceptance peak forces and push-off peak forces in subjects with scoliosis as compared to control subjects. They also reported significantly higher variability for medial–lateral and anterior–posterior shear forces. Among their conclusions, the researchers reported that people with scoliosis presented with balance dysfunction, exhibited by higher variability in the kinetic gait parameters than a control population. Similarly, White et al. [24] reported higher variability in weight acceptance and push-off peak forces in the gait of patients with cerebral palsy. Results of the present study have similarities to those reported by Giakas et al. [11] and White et al. [24] implying that walking in a VR environment induces gait instability. Postural stability is defined as an ability to maintain the body’s center of mass (COM) over the base of support [9]. Instability during walking may therefore manifest as increased displacement of the body’s COM within a constantly changing base of support. The increased COM displacement that occurs in unstable gait may necessitate that people compensate by making adjustments to their base of support. Therefore, the classic spatiotemporal gait deviations that characterize gait instability including reduced stride length, increased step width, and increased variability in stride velocity and step width may occur [16]. The kinetic control required to make such adjustments may be observed by changes in the GRF during locomotion. For example, if the body’s COM displaces anteriorly over the base of support, downward and forward momentum may increase [10]. To compensate for a forward-falling tendency, the magnitudes of weight acceptance and push-off peak forces would increase with concomitant increases in the push-off and weight acceptance rates. Without center of mass data, it is difficult to confirm these hypotheses. However, based on interpretations of previous studies [10,11,24], the increases in magnitude and variability of the weight acceptance and push-off forces observed in the present study are indicative of gait instability induced by the VR environment. Results also suggest that VR influences gait at a central, more than peripheral, level of processing. While vestibular, visual and somatosensory systems all contribute to one’s ability to maintain postural stability during locomotion, the sensory stimulus manipulated in this study was visual only. As per the design of this study, somatosensory cues were not manipulated nor were vestibular cues. It is possible, however, that a conflict between peripheral (somatosensory) and central (visual) sensory systems was induced. The VR environment used in this study was presented as a curved corridor. While subjects were required to walk along a straight path on the treadmill, they likely perceived themselves to be walking along a curved path. The somatosensory and visual systems may have been providing a different reference frame for the control of gait in this
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294
study. Regardless, whether the visual system in isolation or in interaction with the somatosensory system induced the gait changes, it is evident that VR influences gait at a central level of processing. Regarding the clinical relevance of the present study, people often have impaired balance and impaired gait stability in visually stimulating environments following vestibular system injury [17]. While in its infancy, VR is being investigated for its use in rehabilitation for persons with vestibular system pathology who often report difficulty walking while simultaneously moving their heads to attend to peripheral activities [21,25]. For example, patients may describe difficulty walking down a grocery store aisle while looking back and forth at products or when walking in a crowded environment. One method used in vestibular rehabilitation is to create situations that progressively stimulate patients’ symptoms in order to promote habituation and more rapid progress through their rehabilitation. Whitney et al. [21] were among the first investigators to examine VR for patients with vestibular pathology. They studied postural stability in standing activities only. Nevertheless, their study represents a clinical use of VR for rehabilitation purposes. Potentially, the use of VR with a treadmill, as used in the present study, may allow therapists to target situations that cause patients’ symptoms during walking – in a controlled clinical setting – to promote habituation. Some limitations of the study may restrict interpretations of the results. First, gait in able-bodied, healthy adults was examined. The results obtained with this sample may or may not reflect the effect of a VR environment on individuals with vestibular pathology or other disorders that affect gait. Second, treadmills may not necessarily mimic overground walking. While some research suggests that treadmill and overground walking are comparable and that a treadmill is a valid laboratory instrument to measure gait [26], other evidence suggests that gait differs between treadmill and overground walking [27]. It is not certain how results from the present study might translate to overground walking conditions. Third, the force plates in the treadmill bed were only capable of measuring vertical GRF data. Because shear forces (anterior–posterior, medial–lateral) were not measured, the force data may not entirely represent a subject’s kinetic walking patterns. Finally, the use of VR is relatively unexplored in gait studies—‘‘normal’’ gait in VR environments is still, for the most part, undetermined. Despite these limitations, the evidence strongly suggests that the VR environment used in the present study has a significant effect on kinetic gait parameters in healthy individuals walking on a treadmill. Changes in the weight acceptance and push-off phases of the gait cycle observed in the VR environment likely represent compensations for unstable gait. Future related research should examine the effect of VR environments on gait in individuals with vestibular system pathology, the responsiveness of gait stability markers as patients with vestibular system pathology progress through
293
vestibular rehabilitation programs, and the efficacy of VR as a rehabilitation tool. Additional future work may involve different virtual environments, including scenarios that better represent daily activities, such as a VR environment that more closely simulates walking down a grocery store aisle. Such studies may shed light on potential interventions for vestibular dysfunction.
5. Conclusion In summary, walking in a VR environment induces changes in gait that reflect gait instability. Subjects walked in the VR environment with increased weight acceptance peak forces and with increased weight acceptance and pushoff rates. Additionally, subjects walked in the VR environment with increased variability in their weight acceptance rates and with increased variability in the magnitude of their weight acceptance peak and push-off peak forces. Changes in these kinetic gait parameters may be a manifestation of compensatory efforts to control the body’s center of mass over the base of support during locomotion and therefore represent gait instability induced by visual stimulation in a VR environment.
Acknowledgements The authors thank Richard A. Robb, PhD, for software development of the virtual environment image, and Steve Irby, MS, for technical assistance. Donald M. Kendall provided funding to support this study.
References [1] Menz HB, Lord SR, Fitzpatrick RC. Age-related differences in walking stability. Age Ageing 2003;32:137–42. [2] Krebs DE, Goldvasser D, Lockert JD, Portney LG, Gill-Body KM. Is base of support greater in unsteady gait? Phys Ther 2002;82: 138– 47. [3] Maki BE. Gait changes in older adults: predictors of falls or indicators of fear. J Am Geriatr Soc 1997;45:313–20. [4] Bauby CE, Kuo AD. Active control of lateral balance in human walking. J Biomech 2000;33:1433–40. [5] Chou LS, Kaufman KR, Hahn ME, Brey RH. Medio-lateral motion of the center of mass during obstacle crossing distinguishes elderly individuals with imbalance. Gait Posture 2003;18:125–33. [6] Hase K, Stein RB. Analysis of rapid stopping during human walking. J Neurophysiol 1998;80:255–61. [7] Hase K, Stein RB. Turning strategies during human walking. J Neurophysiol 1999;81:2914–22. [8] Varraine E, Bonnard M, Pailhous J. Intentional on-line adaptation of stride length in human walking. Exp Brain Res 2000;130:248–57. [9] Horak FB. Clinical measurement of postural control in adults. Phys Ther 1987;67:1881–5. [10] Hsiang SM, Chang CC. The effect of gait speed and load carrying on the reliability of ground reaction forces. Saf Sci 2002;40:639–57. [11] Giakas G, Baltzopoulos V, Dangerfield PH, Dorgan JC, Dalmira S. Comparison of gait patterns between healthy and scoliotic patients
294
[12]
[13]
[14]
[15]
[16]
[17] [18]
J.H. Hollman et al. / Gait & Posture 26 (2007) 289–294 using time and frequency domain analysis of ground reaction forces. Spine 1996;21:2235–42. Giakas G, Baltzopoulos V. Time and frequency domain analysis of ground reaction forces during walking: an investigation of variability and symmetry. Gait Posture 1997;5:189–97. Buekers M, Montagne G, de Rugy A, Laurent M. The regulation of externally paced human locomotion in virtual reality. Neurosci Lett 1999;275:171–4. Richards JT, Mulavara AP, Bloomberg JJ. Postural stability during treadmill locomotion as a function of the visual polarity and rotation of a three-dimensional virtual environment. Presence-Teleop Virt 2004;13:371–84. Mulavara AP, Richards JT, Ruttley T, Marshburn A, Nomura Y, Bloomberg JJ. Exposure to a rotating virtual environment during treadmill locomotion causes adaptation in heading direction. Exp Brain Res 2005;16:210–9. Hollman JH, Brey RH, Robb RA, Bang TJ, Kaufman KR. Spatiotemporal gait deviations in a virtual reality environment. Gait Posture 2006;23:441–4. Redfern MS, Yardley L, Bronstein AM. Visual influences on balance. J Anxiety Disord 2001;15:81–94. Varraine E, Bonnard M, Pailhous J. Interaction between different sensory cues in the control of human gait. Exp Brain Res 2002;142: 374–84.
[19] Rothbaum BO, Hodges LF, Kooper R, Opdyke D, Williford JS, North M. Effectiveness of computer-generated (virtual reality) graded exposure in the treatment of acrophobia. Am J Psychiatry 1995;152:626–8. [20] Rothbaum BO, Hodges L, Smith S, Lee JH, Price L. A controlled study of virtual reality exposure therapy for the fear of flying. J Consult Clin Psychol 2000;68:1020–6. [21] Whitney SL, Sparto PJ, Brown KE, Furman JM, Jacobson JL, Redfern MS. The potential use of virtual reality in vestibular rehabilitation: preliminary findings with the BNAVE. Neurol Rep 2002;26:72–8. [22] Akiduki H, Nishiike S, Watanabe H, Matsuoka K, Kubo T, Takeda N. Visual-vestibular conflict induced by virtual reality in humans. Neurosci Lett 2003;340:197–200. [23] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower extremity kinematics during level walking. J Orthop Res 1990;8:383–92. [24] White R, Agouris I, Selbie RD, Kirkpatrick M. The variability of force platform data in normal and cerebral palsy gait. Clin Biomech 1999;14:185–92. [25] Viirre E. Virtual reality and the vestibular apparatus. IEEE Eng Med Biol Mag 1996;15(41–43):69. [26] Belli A, Bui P, Berger A, Geyssant A, Lacour JR. A treadmill ergometer for three-dimensional ground reaction forces measurement during walking. J Biomech 2001;34:105–12. [27] Alton F, Baldey L, Caplan S, Morrissey MC. A kinematic comparison of overground and treadmill walking. Clin Biomech 1998;13:434–40.