Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 5 , Pages 793-802, May 2009

Kinematics of Pointing Movements Made in a Virtual Versus a Physical 3-Dimensional Environment in Healthy and Stroke Subjects

Presented as an abstract to the Society for Neuroscience, October 14–18, 2006, Atlanta, GA; and the World Conference of Physical Therapy, June 2–6, 2007, Vancouver, BC, Canada.

  • Luiz A. Knaut, MSc

      Affiliations

    • School of Rehabilitation, University of Montreal, Montreal, QC, Canada
    • Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain, Montreal, QC, Canada
    • ,
  • Sandeep K. Subramanian, MSc

      Affiliations

    • Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain, Montreal, QC, Canada
    • School of Physical and Occupational Therapy, McGill University, Montreal, QC, Canada
    • ,
  • Bradford J. McFadyen, PhD

      Affiliations

    • Department of Rehabilitation, Laval University, Quebec City, QC, Canada
    • Centre for Interdisciplinary Research in Rehabilitation and Social Integration, Quebec, QC, Canada
    • ,
  • Daniel Bourbonnais, PhD, OT

      Affiliations

    • School of Rehabilitation, University of Montreal, Montreal, QC, Canada
    • Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain, Montreal, QC, Canada
    • ,
  • Mindy F. Levin, PhD, PT

      Affiliations

    • Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain, Montreal, QC, Canada
    • School of Physical and Occupational Therapy, McGill University, Montreal, QC, Canada
    • Corresponding Author InformationCorrespondence to Mindy F. Levin, PhD, PT, School of Physical and Occupational Therapy, McGill University, 3630 Promenade Sir William Osler, Montreal, QC, Canada, H3G 1Y5

Article Outline

Abstract 

Knaut LA, Subramanian SK, McFadyen BJ, Bourbonnais D, Levin MF. Kinematics of pointing movements made in a virtual versus a physical 3-dimensional environment in healthy and stroke subjects.

Objective

To compare kinematics of 3-dimensional pointing movements performed in a virtual environment (VE) displayed through a head-mounted display with those made in a physical environment.

Design

Observational study of movement in poststroke and healthy subjects.

Setting

Motion analysis laboratory.

Participants

Adults (n=15; 4 women; 59±15.4y) with chronic poststroke hemiparesis were recruited. Participants had moderate upper-limb impairment with Chedoke-McMaster Arm Scores ranging from 3 to 6 out of 7. Twelve healthy subjects (6 women; 53.3±17.1y) were recruited from the community.

Interventions

Not applicable.

Main Outcome Measures

Arm and trunk kinematics were recorded in similar virtual and physical environments with an Optotrak System (6 markers; 100Hz; 5s). Subjects pointed as quickly and as accurately as possible to 6 targets (12 trials/target in a randomized sequence) placed in arm workspace areas requiring different arm movement patterns and levels of difficulty. Movements were analyzed in terms of performance outcome measures (endpoint precision, trajectory, peak velocity) and arm and trunk movement patterns (elbow and shoulder ranges of motion, elbow/shoulder coordination, trunk displacement, rotation).

Results

For healthy subjects, precision and trajectory straightness were higher in VE when pointing to contralateral targets, and movements were slower for all targets in VE. Stroke participants made less accurate and more curved movements in VE and used less trunk displacement. Elbow/shoulder coordination differed when pointing to the lower ipsilateral target. There were no group-by-environment interactions.

Conclusions

Movements in both environments were sufficiently similar to consider VE a valid environment for clinical interventions and motor control studies.

Key Words: Rehabilitation, Stroke, Upper extremity, Virtual reality, exercise

List of Abbreviations: CRIR, Centre for Interdisciplinary Research in Rehabilitation of Greater Montreal, FOV, field of view, HMD, head-mounted display, IRED, infrared emitting diode, ROM, range of motion, 3D, three-dimensional, 2D, two-dimensional, UE, upper extremity, VE, virtual environment, VR, virtual reality

 

STROKE IS ONE OF THE MAJOR causes of physical disability in adults worldwide,1 and stroke survivors are the largest consumers of rehabilitation services in the United States.2 In a large proportion of patients after stroke, the hemiparetic UE remains an obstacle to the re-establishment of autonomy.3, 4 The decrease in autonomy has a significant impact on the quality of life of patients and their families.2

Motor and sensory impairments like spasticity, abnormal synergies, incoordination, weakness, and loss of sensation affect motor patterns used to produce movement (eg, range of motion, trunk compensation) and therefore motor performance variables (eg, endpoint precision, velocity, trajectory).5, 6, 7, 8, 9, 10, 11 Such sensorimotor deficits can be attenuated by experience-dependent plasticity in the central nervous system that can be induced through repetitive motor training even in the chronic phase of stroke.12, 13, 14 Recent animal and human studies concur that for training programs to induce experience-dependent plasticity effectively, they should include meaningful specific task practice, repetitive and challenging practice, use of appropriate feedback, and learner motivation.15, 16, 17, 18

Motor learning elements can be productively incorporated into environments created with VR technology. VR is a multisensorial experience in which a person is immersed and can interact with a computer-generated environment.19 In VR, environments and tasks can be individualized to the motor abilities and preferences of the learner as well as to the therapeutic goal. The added value of using VR as a therapeutic medium is that it potentially enhances the degree of interaction between the patient and the therapy to limit boredom, fatigue, lack of enthusiasm, and lack of cooperation, which may negatively affect motor recovery.20

The efficacy of VR as a tool to improve UE motor function in patients with chronic stroke is supported in recent literature.21, 22, 23, 24 However, it is not known whether movements performed in VR are similar to those performed in physical training environments. This information is necessary to better understand the applicability of this tool in clinical rehabilitation and motor control studies. In a previous study,25 reaching and grasping kinematics of movements performed in a 2D VE presented some differences compared with those of similar movements performed in a physical environment. Differences may have been a result of the absence of depth perception and limited FOV of the 2D VE, which was displayed to the participants on a computer monitor. Considering the limitation of the VR system in the previous study, it is appropriate to investigate UE movement kinematics in a 3D immersive VR.

The objective of this study was to compare the kinematics of pointing movements performed in a 3D fully immersive VE, displayed through an HMD, with those of movements performed in a physical environment in healthy subjects and in subjects with arm motor deficits caused by stroke-related brain damage. Our purpose was not to determine the differences in pointing movements between healthy subjects and stroke survivors, because these differences have been previously well documented.5, 8

Because a 3D immersive VE provides stereoscopic vision (depth perception), we hypothesized that there would be no differences in the kinematics of pointing movements performed in a 3D immersive VE and a similar physical environment in healthy subjects or in subjects with stroke.

Back to Article Outline

Methods 

Study Sample 

Differences in movements made in a physical environment and a VE were investigated in 2 subject groups: stroke survivors and healthy subjects. Fifteen stroke survivors with hemiparesis (11 men and 4 women; age, 59±15.4y; table 1) were recruited from 3 rehabilitation centers associated with the CRIR. Twelve healthy subjects (6 men and 6 women; age, 53.3±17.1y) recruited from the community also participated in the study. Ethics approval was obtained from CRIR, and all subjects signed an informed consent form prior to participating. Participants had no pain, orthopedic, neuromuscular, or neurologic (other than the stroke) problems affecting the arm and trunk and were able to speak and understand English or French. They had no attention deficits or uncorrected visual problems as assessed by standard clinical tests. Although detailed evaluations of attention and visual deficits were not conducted because of time constraints, participants did not show any impact of attention or visual impairments on the performance of activities of daily living or clinical sensorimotor assessments. Stroke survivors were included if they had experienced a stroke at least 6 months previously (ie, chronic stroke) and had Chedoke-McMaster Arm Scores ranging from 3 to 6 out of 7, indicating moderate hemiparesis.26 Finally, stroke survivors were excluded if they had a lesion in the cerebellum or the occipital lobe and if they had marked apraxia or aphasia as determined by standard clinical tests.

Table 1. Demographic and Clinical Description of Participants With Stroke
SubjectSexAge (y)Site of LesionDuration (mo)Hand DominanceSide of HemiparesisCM ScoreFM ScoreCS Index
P1F75Temporal parietal24RightLeft5594
P2M49Basal ganglia11RightRight6655
P3F60Subcortical21RightRight5538
P4F80MC artery14RightRight5617
P5M57Subcortical14RightRight5576
P6M67Parietal26LeftRight45410
P7M77MC artery73LeftLeft4474
P8M40MC artery30RightLeft5498
P9F30MC artery101RightRight44112
P10M45Basal ganglia12RightRight4508
P11M70MC artery40RightLeft3297
P12M78Subcortical32RightLeft3198
P13M45Subcortical13RightLeft6584
P14M51MC artery18RightRight5575
P15M61Parietal64RightLeft6597

Abbreviations: CM, arm section of the Chedoke McMaster score (7=normal arm activity); CS, Composite Spasticity Index (4=normal tone); F, female; FM, upper-limb section of the Fugl-Meyer score (66=normal arm function); M, male; MC, middle cerebral.

Clinical Evaluations 

Prior to the experiment, all subjects with stroke underwent a 30-minute clinical assessment by research clinicians to determine the level of motor impairment of their affected UE.

The motor recovery level of the hemiparetic UE was evaluated with the Fugl-Meyer Upper Limb Scale.27 This evaluation measures the capacity to produce movements voluntarily, selectively, and in a coordinated fashion. According to this scale, UE movement was considered normal if the subject reached a maximal score of 66 points.28

Spasticity of the elbow muscles was assessed using the Composite Spasticity Index. This valid29 and reliable test measures spasticity using a combined score of the resistance felt during stretch of the passive elbow flexors, the excitability of the biceps brachii tendon reflex, and the amount of wrist clonus. A score of 4 out of 16 indicates normal tonus, while a score of 16 out of 16 means severe spasticity.30

Kinematic Recording 

Movement kinematics were recorded using an Optotrak Motion Analysis System (Northern Digital Corp, Type 3020)a at a sampling frequency of 100Hz. To record arm and trunk movements, IREDs were placed on the tip of the index (distal phalange of the index finger, ie, endpoint), wrist (styloid process at the head of radius), elbow (lateral epicondyle), ipsilateral and contralateral shoulders (acromion processes), and trunk (middle of sternum). For each trial, data were recorded for 5 seconds.

Physical Environment 

To be comparable, the physical environment and VE were created to be as similar as possible to each other. In the physical environment, six 60×60mm square targets were attached to an adjustable support and arranged in a 2-row by 3-column grid numbered 1 to 6 (fig 1). The distance between the centers of adjacent squares was 260mm. The grid was positioned in front of the participant such that the middle squares (ie, targets 2 and 5) were aligned with the sternum and the midline between the top and bottom squares was aligned with the shoulders. The distance between the subject and the midpoint of the grid was equal to the length of the subject's arm (ie, from the acromion to the tip of the index) plus 50mm. Fifty millimeters was added to the arm length in order to avoid physical contact of the fingertip with the target to ensure that the conditions in the physical environment were similar to those in the VE, in which no haptic information from the target was available.

  • View full-size image.
  • Fig 1. 

    Target arrangement in the coronal (A) and transverse (B) planes for both the physical and virtual environments for participants reaching with the right arm. Abbreviations: LC, lower contralateral; LI, lower ipsilateral; LM, lower middle; UC, upper contralateral; UI, upper ipsilateral; UM, upper middle.

Virtual Reality Environment 

The VE, previously described in Subramanian et al,31 consisted of a 3D environment generated by a personal computer (Dual Xeon 3.06GHz, 2GB RAM, 160GB hard drive) and displayed to the user through an HMD.b Head position and orientation in virtual space were tracked by a 6 IRED rigid body attached to the HMD. The endpoint representation in the VE was indicated by a blue dot, obtained from the IRED on the index. This was the only body cue indicated to the participants when they were immersed in the VE. The data created by these interfaces were integrated by CAREN software.c The system also included a dual-head Nvidia Quatro FX3000 graphics card (70 Hz) providing high-speed stereoscopic representation of the environment that was created on SoftImage XSI.d

The scene in the VE consisted of 6 targets of the same dimensions and displayed in the same array as that described for the physical environment, except that they appeared as elevator buttons arranged on a virtual elevator wall. This environment was specifically created so that the movements required of the users could be directly compared with those made in the physical environment. In addition, it was designed to be easily understood and to reproduce functional pointing movements. Although it differed from VEs commonly used in rehabilitation, it had a gaming context because users earned points for each movement performed in a fast and accurate way, and the game score was visible to the participant. The scene was calibrated so that the target locations in 3D space were reproduced exactly as in the physical environment with respect to the distance from the participant's body.31

Experimental Procedure 

Participants were comfortably seated on a chair with the knees and hips in approximately 90° of flexion and the feet supported on the floor. Prior to each trial, participants placed the tip of the index finger on the xiphoid process so that the arm was in approximately 50° of shoulder abduction, 0° of shoulder flexion, 120° of elbow flexion, and the forearm and wrist were in the neutral position.

Participants in each group performed the same task in both environments. The task consisted of 12 pointing movements made toward each of 6 different targets in each environment (72 trials per environment; in total, 144 trials in the physical environment and VE). The target sequence and the order of presentation of environments were randomized to avoid learning effects. The task was designed so that forward trunk displacement was not necessary because the goal was to point to and not to touch the targets. The targets were arranged to recruit different combinations of shoulder and elbow movements with different degrees of difficulty. For example, the upper row of targets (ie, 1, 2, 3) required more shoulder flexion than the lower row. In addition, the ipsilateral targets required shoulder horizontal abduction combined with elbow extension, and contralateral targets required shoulder horizontal adduction combined with elbow extension. Participants were instructed to execute the movements as accurately and quickly as possible. Subjects with stroke performed the task with the paretic UE, and healthy subjects used the nondominant UE.

The target to be pointed at was indicated at the beginning of each trial by an auditory signal emitted by the computer (eg, 6, meaning “point to target 6”). Information about successful pointing attempts in terms of precision (ie, finger arrived within the 60×60mm target) and speed (within 5s) were indicated to the subject by a ping sound generated by the computer. In addition, in the VE, a concurrent visual command was used to indicate the target to the participant (ie, a change of the target color). To avoid fatigue, the task was separated into 3 blocks of 24 trials, with a 3-minute pause between blocks for each environment. The order of task completion in each environment was balanced across subjects. If needed, additional rest pauses were given to the participants.

Prior to data collection, participants practiced some trials in each condition to familiarize themselves with the environment and equipment. Symptoms of cybersickness (nausea, dizziness, vomiting, visual problems) associated with viewing virtual environments through an HMD were prevented by using a high-speed tracking system with short latencies for data acquisition, which effectively eliminated oscillations of the VE,32 and by wearing the HMD for less than 20 minutes at a time.33

Data Analysis 

Kinematic variables used to compare the movements performed in the physical environment and the VE were separated into performance outcomes and movement pattern outcomes. Performance outcomes included endpoint (ie, tip of index) precision, peak velocity, and trajectory straightness. Movement pattern outcomes were elbow and shoulder ROM, trunk displacement and rotation, and interjoint coordination between elbow extension and shoulder horizontal abduction/adduction.

Tangential velocities of the endpoint and trunk were computed from the magnitude of the velocity vector, obtained by numerical differentiation of the x, y, and z positional data for markers placed on the index finger and sternum, respectively. The beginning and end of movement were defined as the times at which the tangential velocity surpassed and remained above or fell and remained below 10% of the peak velocity of the same trial. Data were analyzed using LabView software.e

Performance Outcomes 

Endpoint precision was calculated in terms of absolute error and was computed as the root-mean-squared distance between the final index finger position and the center of the target. Endpoint peak velocity was calculated from tangential velocity traces. Endpoint trajectory was determined by the index of curvature (ie, the ratio of the actual length of the endpoint path to the length of a straight line joining the initial and final positions), which has been shown to characterize trajectory straightness better than area measurements.34 An ideal straight line has an index of 1, whereas that of a semicircle has an index of 1.57.

Movement Pattern Outcomes 

Elbow flexion/extension ROM was calculated based on the angles formed by 2 vectors between the wrist-elbow IREDs and the ipsilateral shoulder-elbow IREDs. Full elbow extension was defined as 180°. Shoulder flexion/extension ROM was calculated based on the angles between the vectors formed by the elbow-ipsilateral shoulder IREDs and a vertical line through the axis of the ipsilateral shoulder in the sagittal plane. The position with the arm alongside the body was defined as 0°. Shoulder horizontal adduction/abduction ROM was calculated based on the angles between the 2 vectors formed by the elbow-ipsilateral shoulder IREDs and the contralateral shoulder-ipsilateral shoulder IREDs on the horizontal plane. Zero degrees of shoulder horizontal adduction was defined as the arm in a position following the line joining the IREDs placed on the shoulders. Trunk displacement was measured in millimeters and computed from the IRED on the sternum as the distance moved between the beginning and end of trunk movement, in the sagittal plane. Trunk axial rotation was defined as the angle of rotation of the vector joining the 2 shoulder IREDs with respect to the coronal plane. The initial position was defined as 0°. Interjoint coordination was defined as the slope of the elbow extension versus shoulder horizontal abduction/adduction relationship that was computed using quadratic regression analysis. A slope of 1 indicated that both joints contributed equally to the movement, while a slope different from 1 indicated that the movement involved predominantly 1 of the joints. This relationship was chosen as the more complicated coordination compared with that between elbow extension and shoulder flexion in pointing movements, because it involves movements in 2 planes (ie, horizontal, sagittal).6

Statistical Analysis 

Data normality was checked using the Levene test. Pointing movements executed in the physical environment and the VE were compared on mean data using a multivariate 2×2×6 analysis of variance with group (healthy, stroke) and environment (physical environment, VE) as independent variables and target (n=6) as the dependent variable. Data from different targets could not be combined because movements toward each target involved different movement patterns. For example, to point to ipsilateral targets, participants combined elbow extension with shoulder flexion and horizontal abduction, while contralateral targets required elbow extension combined with shoulder flexion and horizontal adduction. Because the goal of the study was to compare the environments and not the groups, between-group analyses were done only to measure group-by-environment interactions. Statistical analyses were done using SPSS 10.0 for Windows software,f and the significance level was set at P less than .05.

Back to Article Outline

Results 

Examples of typical endpoint and trunk trajectories of movements made to each of the targets in 1 healthy and 1 stroke subject are shown in figure 2. Analysis of group by environment interactions revealed no differences for all outcomes investigated: endpoint peak velocity (F6,45=.490; P=.813), endpoint precision (F6,45=1.238; P=.305), trajectory straightness (F6,45=.867; P=.526), elbow/shoulder interjoint coordination (F6,45=1.885; P=.104), elbow extension (F6,45=.555; P=.764), shoulder flexion (F6,45=.393; P=.879), shoulder horizontal adduction (F6,45=.447; P=.843), trunk displacement (F6,45=.414; P=.866), and trunk rotation (F6,45=.315; P=.926).

  • View full-size image.
  • Fig 2. 

    Endpoint and trunk trajectories toward the 3 upper (UI, UM, UC) and 3 lower (LI, LM, LC) targets in the physical environment (PE) and the VEs obtained from 1 healthy subject and 1 patient with stroke. Abbreviations: LC, lower contralateral; LI, lower ipsilateral; LM, lower middle; UC, upper contralateral; UI, upper ipsilateral; UM, upper middle.

Performance Outcomes: Healthy Group 

In the healthy subjects, there were few differences in performance or movement pattern outcomes for the pointing movements made in the physical environment and the VE. Small but significant differences between physical environment and VE conditions were observed in endpoint precision for the upper contralateral (physical environment 6.5% more precise) and lower contralateral (physical environment 6.1% more precise) targets as well as in trajectory straightness for the upper contralateral target (physical environment: 5% straighter; P<.05). Overall, movements made in VE to all targets were approximately 35.4% slower than those made in physical environment for all targets (P<.05). Significance levels for all variables are listed in table 2.

Table 2. Comparison Between Movements Made by Healthy Subjects in the Virtual and Physical Environments of Each Variable for Each Target Obtained With Multivariate Analysis of Variance
Healthy Subjects
VariableTarget
UIUMUCLILMLC
Endpoint trajectory (IC)
Mean ± SD
PE1.2±0.11.2±0.11.2±0.0⁎1.1±0.11.1±0.01.1±0.1
VE1.2±0.11.2±0.11.3±0.1⁎1.1±0.11.1±0.01.1±0.1
F1,222.3683.7786.943⁎2.4953.3764.245
P0.1380.0650.015⁎0.1280.0800.051
Endpoint precision (mm)
Mean ± SD
PE260.0±7.042.0±5.0248.0±10.0⁎364.0±11.0262.0±7.0353.0±12.0⁎
VE257.0±16.050.0±24.0265.0±24.0⁎370.0±14.0265.0±8.0374.0±17.0⁎
F1,220.3151.2414.592⁎1.2040.42512.202⁎
P0.5800.2270.043⁎0.2840.5210.002⁎
Endpoint peak velocity (mm/s)
Mean ± SD
PE2242.0±804.0⁎1928.0±682.0⁎1724.0±573.0⁎2077.0±691.0⁎1762.0±627.0⁎1605.0±600.0⁎
VE1638.0±351.0⁎1434.0±322.0⁎1302.0±382.0⁎1525.0±392.0⁎1306.0±303.0⁎1169.0±371.0⁎
F1,225.684⁎5.153⁎4.511⁎5.796⁎5.146⁎4.589⁎
P0.026⁎0.033⁎0.045⁎0.025⁎0.033⁎0.044⁎
Elbow extension (degrees)
Mean ± SD
PE96.7±9.697.1±9.098.9±7.196.4±9.397.2±8.999.6±8.5
VE95.6±13.497.8±11.597.2±11.095.0±13.197.2±10.898.4±11.5
F1,220.0000.1840.0750.0000.1830.000
P1.0000.6720.7870.9920.6730.988
Shoulder flexion (degrees)
Mean ± SD
PE83.3±11.086.3±9.987.2±7.664.2±11.367.9±10.269.8±9.7
VE81.5±12.485.5±9.486.9±8.163.4±11.167.3±9.670.6±9.2
F1,220.1450.0370.0120.0260.0220.045
P0.7070.8490.9150.8730.8830.835
Shoulder horizontal adduction (degrees)
Mean ± SD
PE52.5±17.467.3±14.874.4±16.555.1±18.570.2±14.577.7±16.8
VE51.1±18.067.5±15.179.0±17.054.5±17.569.5±14.983.1±18.0
F1,220.0390.0000.4510.0060.0160.575
P0.8450.9830.5090.9410.9000.456
Trunk displacement (mm)
Mean ± SD
PE26.0±12.034.0±9.062.0±18.022.0±9.029.0±8.0100.0±8.0
VE28.0±16.033.0±11.055.0±20.023.0±13.027.0±9.098.0±12.0
F1,220.0900.0810.7820.0330.1371.923
P0.7660.7780.3860.8570.7150.179
Trunk rotation (degrees)
Mean ± SD
PE9.9±4.612.0±4.619.9±5.89.2±4.611.9±4.355.7±15.8
VE10.3±5.911.6±5.616.3±8.29.4±5.811.5±5.3446.2±17.5
F1,220.0290.0521.5130.0070.0442.200
P0.8660.8220.2320.9320.8360.152
Elbow/shoulder coordination (slope)
Mean ± SD
PE–3.9±3.2–3.1±1.7–2.8±1.2–1.8±2.0–1.8±0.920.6±5.7
VE–3.1±3.2–2.9±2.0–2.0±1.3–1.5±1.6–1.5±1.016.5±7.7
F1,220.4120.0702.8300.2590.5711.470
P0.5280.7940.1070.6160.4580.238

Abbreviations: IC, index of curvature; LC, lower contralateral target; LI, lower ipsilateral target; LM, lower middle target; PE, physical environment; UC, upper contralateral target; UI, upper ipsilateral target; UM, upper middle target.

⁎Significant difference at P <.05.

Movement Pattern Outcomes: Healthy Group 

Elbow flexion/extension, shoulder flexion/extension and shoulder horizontal adduction/abduction ranges of motion as well as the amount of trunk displacement and rotation were similar in physical environment and VE for all targets (P>.05). In addition, there were no differences in the elbow/shoulder interjoint coordination used in each environment across targets (P>.05). Shoulder horizontal adduction contributed more at the beginning of the movement than elbow extension for all targets. This pattern occurred in 100% of the movements in the physical environment and 92% of the movements in the VE.

Performance Outcomes: Stroke Group 

Similar to healthy subjects, there were few differences in performance variables across targets in the stroke group. Endpoint precision to 5 of the 6 targets and peak velocities were similar across targets in both environments (P>.05). Subjects with stroke made less accurate movements in VE than in the physical environment only when pointing to the upper contralateral target (physical environment: 13% more precise; P<.01). Stroke survivors made more curved trajectories in VE than the physical environment for movements executed to both contralateral targets (upper contralateral, physical environment, 21% straighter; and lower contralateral, physical environment, 20.5% straighter) and the lower ipsilateral (physical environment, 14.7% straighter) target (.007<P<.023). Significance levels for all variables are listed in table 3.

Table 3. Comparison Between Movements Made by Stroke Subjects in the Virtual and Physical Environments of Each Variable for Each Target Obtained With Multivariate Analysis of Variance
Subjects With Stroke
VariableTarget
UIUMUCLILMLC
Endpoint trajectory (IC)
Mean ± SD
PE1.3±0.11.4±0.21.4±0.2⁎1.2±0.1⁎1.2±0.21.2±0.2⁎
VE1.5±0.41.7±0.61.7±0.4⁎1.3±0.3⁎1.3±0.31.5±0.4⁎
F1,223.0573.9087.919⁎5.826⁎4.0918.550⁎
P0.0910.0580.009⁎0.023⁎0.0530.007⁎
Endpoint precision (mm)
Mean ± SD
PE268.0±19.095.0±49.0260.0±23.0⁎351.0±32.0245.0±52.0346.0±31.0
VE297.0±52.0134.0±72.0294.0±41.0⁎372.0±33.0263.0±51.0373.0±44.0
F1,223.9842.9387.636⁎3.0490.9733.913
P0.0560.0980.010⁎0.0920.3320.058
Endpoint peak velocity (mm/s)
Mean ± SD
PE1305.0±438.01222.0±444.01091.0±390.01353.0±428.01162.0±432.01040.0±398.0
VE1124.0±405.01034.0±359.0930.0±355.01144.0±411.0986.0±337.0884.0±303.0
F1,221.3871.6111.3951.8611.5461.465
P0.2490.2150.2480.1830.2240.236
Elbow extension (degrees)
Mean ± SD
PE81.8±22.180.4±22.579.3±21.180.4±18.781.1±18.582.0±15.4
VE83.2±27.283.2±23.476.4±21.885.0±25.583.3±20.879.6±17.0
F1,220.0090.0420.3490.0070.0480.546
P0.9250.8390.5600.9320.8280.466
Shoulder flexion (degrees)
Mean ± SD
PE63.2±17.965.2±15.667.3±16.644.7±15.648.1±14.950.9±14.2
VE65.4±17.966.5±14.266.7±13.847.5±16.748.41±14.551.5±12.9
F1,220.1110.0610.0130.2280.0040.016
P0.7410.8070.9120.6370.9490.901
Shoulder horizontal adduction (degrees)
Mean ± SD
PE36.1±17.347.0±18.558.4±20.335.2±16.248.4±21.059.9±19.2
VE38.0±16.049.8±18.062.0±17.240.8±15.450.2±18.166.0±16.7
F1,220.0940.1730.2760.9730.0700.876
P0.7610.6800.6040.3320.7930.357
Trunk displacement (mm)
Mean ± SD
PE53.0±38.071.0±34.0103.0±39.0⁎50.0±38.056.0±28.089.0±33.0⁎
VE50.0±23.060.0±20.076.0±30.0⁎41.0±21.047.0±19.062.0±20.0⁎
F1,220.4761.3114.506⁎0.6780.9627.322⁎
P0.4960.2620.043⁎0.4170.3350.011⁎
Trunk rotation (degrees)
Mean ± SD
PE8.1±3.511.3±4.717.9±6.6⁎8.2±3.512.0±4.220.7±6.1⁎
VE6.8±3.38.7±3.912.2±5.0⁎7.4±2.710.8±4.014.3±4.1⁎
F1,221.0382.7027.232⁎0.4840.69411.462⁎
P0.3170.1110.012⁎0.4930.4120.002⁎
Elbow/shoulder coordination (slope)
Mean ± SD
PE–5.2±5.1–5.0±3.9–2.8±1.7–11.9±20.0⁎–3.3±3.0–2.2±1.5
VE–5.8±4.3–4.8±4.4–2.3±1.6–2.1±2.2⁎–2.5±2.3–2.0±1.5
F1,220.1090.0200.5368.664⁎0.9160.137
P0.7440.8870.4700.006⁎0.3470.714

Abbreviations: IC, index of curvature; LC, lower contralateral target; LI, lower ipsilateral target; LM, lower middle target; PE, physical environment; UC, upper contralateral target; UI, upper ipsilateral target; UM, upper middle target.

⁎Significant differences.

Movement Pattern Outcomes: Stroke Group 

Ranges of elbow and shoulder joint motion were similar in both environments for all targets (P>.05). As well, trunk displacement and rotation were not different in the VE and the physical environment for movements to 4 of the 6 targets (P>.05). Significant differences were observed only for the contralateral targets (upper contralateral and lower contralateral; P<.05), for which subjects with stroke moved the trunk less in VE.

As observed in healthy subjects, subjects with stroke also tended to start the pointing movements using more shoulder horizontal adduction than elbow extension. However, this pattern occurred less often than in healthy subjects, corresponding to 88% of the movements in the physical environment and 81% of the movements in the VE. Elbow/shoulder interjoint coordination was similar between environments for all but the lower ipsilateral target (P<.01). The less negative value of the slope suggests that for the target in the VE, there was a more equal contribution of both joints, whereas in the physical environment, the contribution of elbow extension was greater than that of shoulder horizontal adduction.

Back to Article Outline

Discussion 

This study is an important step in the understanding of how virtual reality technology can be used for motor rehabilitation. Since the 1990s, authors have investigated the applicability of VR to motor rehabilitation.35, 36, 37, 38, 39, 40 However, the similarity of movement kinematics in virtual and physical environments has not been systematically studied. In a previous article by our group,31 the 6-target virtual environment developed for this study was described in detail, and some preliminary comparisons of movement performance variables (endpoint velocity, pointing error, trajectory smoothness) were reported. The present study provides more detailed comparisons of both performance and movement pattern data.

The use of VR for motor control studies has been criticized on the basis that we lack knowledge about how altered sensory information delivered via VR may affect voluntary movement production.41 The major limitations of sensory experiences in VR are a result of differences in haptic experiences and in perception of the location of objects in 3D space. Indeed, even with the technological advances, certain types of haptic feedback are not yet properly integrated into virtual experiences. Because of this, we studied pointing movements made in nearly identical physical and virtual environments that did not rely on haptic information and found that performance and movement pattern outcomes were similar for healthy subjects and subjects with stroke. Thus, the results of this study suggest that sensory information from a carefully designed 3D VR environment such as the one studied here may be perceptually similar enough to the real world environment to have only minimal impact on movement kinematics.42 Small disparities in movement characteristics observed in our study are likely a result of differences in perception of the location of the target in space, which may account for slower and less accurate movement.43

Movement Kinematics and Perception of the Environment 

Comparisons of the kinematics of pointing or reaching movements made in physical and 3D virtual environments have not been reported previously. For both groups, motor performance and movement patterns were similar in both environments for movements to midline targets. This was also true for 1 (upper ipsilateral) and both ipsilateral targets in subjects with stroke and healthy subjects, respectively, while there were some differences in performance outcomes for movements to contralateral targets. Our results for midline targets are similar to those reported by Viau et al,25 who compared midline movements made in a physical environment and a 2D VE. These similarities suggest that distance perception in the fully immersive VE and physical environment was similar. For accurate distance estimation, stereopsis that was present in the physical environment and provided by binocular visual cues in the VE is necessary,44 especially when relatively fast and skilled movements are made.45 Our results support those of Interrante et al,46 who reported no differences in adults without previous experience in VEs in distance perception in a physical environment and a VE that was a geometrical and photographic exact replica of the physical environment. These authors also suggested that distance compression in immersive VEs may be significantly reduced when the space is known by the users (eg, reproduction of the environment that they presently physically occupy).

How perception may influence action has also been addressed by consideration of cortical activation while subjects are viewing different environments. Differences in cortical activation have been reported when an object-grasping action was observed in 4 different environments: physical environment, 3D VE with a realistic hand representation, 3D VE with a coarse hand representation, and a 2D television-displayed movie.47 In all environments, activation was present in the motor cortex, visual areas, and left posterior parietal cortex and operculum. However, the 2D VE was disadvantaged because activation in the inferior temporal regions, important for object recognition and perceptuo-cognitive representation of action, was absent.48 The lack of activation in this region suggests that the observed action may be meaningless for the person.49

On the other hand, 3D VEs do not exactly reproduce visual stimuli from the physical world.47 Activation of the right inferior parietal cortex occurs when subjects observe self-movement, but not when the same movements are observed in VEs. This is relevant because the right inferior parietal cortex is important for motor planning50 and plays a crucial role in visually guided reaching and manipulation.51 In addition, as observed by Inoue et al,52 the inferior parietal cortex is part of a network formed by the premotor and posterior cingulate cortices and the cerebellum, which are responsible for monitoring self-movement and for integrating visual and proprioceptive information with ongoing motor commands for accurate pointing.

In our study, environmental conditions differed in that participants could see their whole UE in the physical environment but received only visual feedback about endpoint position in the VE. However, this difference may not necessarily influence the results, because vision of the endpoint is most likely used as a reference to guide UE movement.53, 54, 55, 56

In the physical environment, pointing movements in subjects with stroke were made with excessive trunk movements even though the targets were placed within arm's reach. In contrast, subjects with stroke used less trunk displacement and rotation when pointing to contralateral targets in VE. This may be attributed to the effect of the environment itself or to the ergonomic influence of wearing the HMD, restraining head-on-trunk movement especially for movements to the contralateral targets. This was observed only in the stroke group, because healthy subjects did not use trunk movements for task execution.

Field of View 

Healthy subjects but not subjects with stroke made slower movements in the VE than the physical environment for all targets. Slower movements may be related to viewing the environment through an HMD. Reduced gait speed was reported in healthy subjects walking different distances (ie, 3.1, 6.1, 9.1 meters) in the physical environment and the VE.46 Subjects walked approximately 2 seconds slower for all distances when immersed in an HMD-viewed 3D VE than when walking in the physical environment. Wearing the HMD may have imposed some constraints on movement production because of its additional weight (≈1kg) and the reduction in the FOV (30° vertical, 40° horizontal) compared with normal FOV values (120° vertical, 180° horizontal).57 Reduced FOVs (ie, binocular 4° and 16° FOVs) did not affect distance perception during pointing and reaching movements to objects placed at distances between 200 and 400mm away.58 Similarly, distance estimation was not different in restricted (32°–43° vertical, 38°–47° horizontal) and unrestricted FOV conditions when walking toward targets 2 to 15m from their start position.57, 59 On the other hand, Loftus et al58 also found that the restricted FOV decreased precision and peak velocity of pointing movements. A similar reduction in peak velocity during reaching movement was found with FOVs restricted to 11° to 23°.60 Although the FOVs investigated in those studies were considerably smaller than those in our study, our participants may have been disadvantaged in terms of movement speed and precision when wearing the HMD. Similar decreases in speed in subjects with stroke may not have been observed, however, because movement velocities were already diminished in this group.5

Contralateral Reaches 

In addition to speed, differences occurred in both groups in endpoint precision and curvature only for movements made toward contralateral targets (upper contralateral, lower contralateral). This may be explained by the greater difficulty to perform contralateral movements and by perceptual differences for objects located in the peripheral FOV. Pointing to contralaterally located targets is more difficult, especially for subjects with stroke, because the arm has to cross the body midline in a complex coordination of elbow extension with shoulder flexion and horizontal adduction.6 In addition, psychophysical research suggests that people prefer to use the most proximal hand to the target/object when executing pointing and reaching movements.61 Differences in contralateral movements can also be explained by the necessity to rotate the head and trunk together as the hand crosses the midline, which was not necessary for other targets, suggesting that turning both the head and the trunk may be more difficult in the VE.

Back to Article Outline

Conclusions 

We observed that healthy subjects made slower pointing movements in the VE displayed through the HMD than the physical environment. Aside from this difference, in general, movements in a fully immersive VE and physical environment were performed similarly by healthy subjects and subjects with stroke. Guiding the movement and accuracy in VE was higher when healthy participants pointed toward contralateral targets, while for subjects with stroke, trunk movements may have been limited by wearing the HMD. The results of the movement comparison are likely to be specific for the 3D environment created for this study. Our environment was not typical of VEs commonly used in rehabilitation that may have a more complex gaming or functional context.22, 39, 62 Indeed, our environment did not exploit all the attributes of VR such as novelty, adaptability, and competition through more sophisticated forms of gaming.

Future research should verify whether differences in movement kinematics exist when more complex movements involving object manipulation are made in fully immersive VEs. If VEs are to be used to create interventions to improve arm motor performance in patients with stroke, it would also be necessary to investigate whether and how motor learning may be affected by the training environment.

Suppliers

Back to Article Outline

Acknowledgments 

We thank Ruth Dannenbaum-Katz, MSc, Christian Beaudoin, MSc, and Eric Johnstone, BSc.

Back to Article Outline

References 

  1. World Health Organization. 2002. http://www.who.int/en/Accessed January 15, 2009
  2. American Heart Association. 2006. http://www.americanheart.orgAccessed January 15, 2009
  3. Olsen TS. Arm and leg paresis as outcome predictors in stroke rehabilitation. Stroke. 1990;21:247–251
  4. Carod-Artal J, Egido JA, Gonzalez JL, Varela de Seijas E. Quality of life among stroke survivors evaluated 1 year after stroke: experience of a stroke unit. Stroke. 2000;31:2995–3000
  5. Cirstea MC, Levin MF. Compensatory strategies for reaching in stroke. Brain. 2000;123:940–953
  6. Cirstea MC, Mitnitski AB, Feldman AG, Levin MF. Interjoint coordination dynamics during reaching in stroke. Exp Brain Res. 2003;151:289–300
  7. Dewald JP, Sheshadri V, Dawson ML, Beer RF. Upper-limb discoordination in hemiparetic stroke: implications for neurorehabilitation. Top Stroke Rehabil. 2001;8:1–12
  8. Levin MF. Interjoint coordination during pointing movements is disrupted in spastic hemiparesis. Brain. 1996;119:281–293
  9. Mercier C, Bourbonnais D. Relative shoulder flexor and handgrip strength is related to upper limb function after stroke. Clin Rehabil. 2004;18:215–221
  10. Musampa NK, Mathieu PA, Levin MF. Relationship between stretch reflex thresholds and voluntary arm muscle activation in patients with spasticity. Exp Brain Res. 2007;181:579–593
  11. Winward CE, Halligan PW, Wade DT. Current practice and clinical relevance of somatosensory assessment after stroke. Clin Rehabil. 1999;13:48–55
  12. Michaelsen SM, Dannenbaum R, Levin MF. Task-specific training with trunk restraint on arm recovery in stroke: randomized control trial. Stroke. 2006;37:186–192
  13. Nudo RJ. Functional and structural plasticity in motor cortex: implications for stroke recovery. Phys Med Rehabil Clin N Am. 2003;14(1 Suppl):S57–S76
  14. Teasell R, Bayona N, Salter K, Hellings C, Bitensky J. Progress in clinical neurosciences: stroke recovery and rehabilitation. Can J Neurol Sci. 2006;33:357–364
  15. Barker RN, Brauer SG. Upper limb recovery after stroke: the stroke survivors' perspective. Disabil Rehabil. 2005;27:1213–1223
  16. Cirstea MC, Ptito A, Levin MF. Arm reaching improvements with short-term practice depend on the severity of the motor deficit in stroke. Exp Brain Res. 2003;152:476–488
  17. Kleim JA, Jones TA. Principles of experience-dependent neural plasticity: implications for rehabilitation after brain damage. J Speech Lang Hear Res. 2008;51:S225–S239
  18. Wolf SL, Winstein CJ, Miller JP, et al. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA. 2006;296:2095–2104
  19. Schultheis MT, Himelstein J, Rizzo AA. Virtual reality and neuropsychology: upgrading the Current Tools. J Head Trauma Rehabil. 2002;17:378–394
  20. Tinson DJ. How stroke patients spend their days: an observational study of the treatment regime offered to patients in hospital with movement disorders following stroke. Int Disabil Stud. 1989;11:45–49
  21. Sveistrup H. Motor rehabilitation using virtual reality. J Neuroeng Rehabil. 2004;10:1–10
  22. Holden MK. Virtual environments for motor rehabilitation: review. Cyberpsychol Behav. 2005;8:187–219
  23. Foley N, Teasell R, Jutai J, Bhogal S, Kruger E. Evidence-based review of stroke rehabilitation—upper extremity interventions. In:  Teasell R,  Foley N,  Salter K,  Bhogal S,  Jutai J,  Speechley M editor. Evidence-based review of stroke rehabilitation. 10th ed.. 2007;http://www.ebrsr.comAccessed January 15, 2009
  24. Henderson A, Korner-Bitensky N, Levin M. Virtual reality in stroke rehabilitation: a systematic review of its effectiveness for upper limb motor recovery. Top Stroke Rehabil. 2007;14:52–61
  25. Viau A, Feldman AG, McFadyen BJ, Levin MF. Reaching in reality and virtual reality: a comparison of movement kinematics in healthy subjects and in adults with hemiparesis. J Neuroengineering Rehabil. 2004;1:11
  26. Gowland C, Stratford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24:58–63
  27. Fugl-Meyer A, Jaasko L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient I: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:11–17
  28. Duncan P, Goldstein LB, Matchar D, Divine GW, Feussner J. Measurement of motor recovery after stroke: outcome assessment and sample size requirements. Stroke. 1992;23:1084–1089
  29. Nadeau S, Arsenault AB, Gravel D, Lepage Y, Bourbonnais D. Analysis of the spasticity index used in adults with a stroke. Can J Rehabil. 1998;11:219–220
  30. Levin MF, Hui-Chan C. Are H and stretch reflexes in hemiparesis reproducible and correlated with spasticity?. J Neurol. 1993;240:63–71
  31. Subramanian S, Knaut LA, Beaudoin C, McFadyen BJ, Feldman AG, Levin MF. Virtual reality environments for post-stroke arm rehabilitation. J Neuroengineering Rehabil. 2007;4:20
  32. Lo WT, So RH. Cybersickness in the presence of scene rotational movements along different axes. Appl Ergon. 2001;32:1–14
  33. Regan EC, Price KR. The frequency of occurrence and severity of side-effects of immersion virtual reality. Aviat Space Environ Med. 1994;65:527–530
  34. Archambault P, Pigeon P, Feldman AG, Levin MF. Recruitment and sequencing of different degrees of freedom during pointing movements involving the trunk in healthy and hemiparetic subjects. Exp Brain Res. 1999;126:55–67
  35. Satava RM. Medical applications of virtual reality. J Med Syst. 1995;19:275–280
  36. Dumay AC. Medicine in virtual environments. Technol Health Care. 1995;3:75–89
  37. Rose FD, Attree EA, Johnson DA. Virtual reality: an assistive technology in neurological rehabilitation. Curr Opin Neurol. 1996;9:461–467
  38. Wilson PN, Foreman N, Stanton D. Virtual reality, disability and rehabilitation. Disabil Rehabil. 1997;19:213–220
  39. Jang SH, You SH, Hallett M, et al. Cortical reorganization and associated functional motor recovery after virtual reality in patients with chronic stroke: an experimenter-blind preliminary study. Arch Phys Med Rehabil. 2005;86:2218–2223
  40. You SH, Jang SH, Kim YH, et al. Virtual reality-induced cortical reorganization and associated locomotor recovery in chronic stroke: an experimenter-blind randomized study. Stroke. 2005;36:1166–1171
  41. Latash ML. Virtual reality: a fascinating tool for motor rehabilitation (to be used with caution). Disabil Rehabil. 1998;20:104–105
  42. Stoffregen TA, Bardy BG. On specification and the senses. Behav Brain Sci. 2001;24:195–261
  43. Rochat P, Wraga M. An account of the systematic error in judging what is reachable. J Exp Psychol Hum Percept Perform. 1997;23:199–212
  44. Cumming BG, DeAngelis GC. The physiology of stereopsis. Annu Rev Neurosci. 2001;24:203–238
  45. Servos P. Distance estimation in the visual and visuomotor systems. Exp Brain Res. 2000;130:35–47
  46. Interrante V, Anderson L, Ries B. Distance perception in immersive virtual environments, revisited. In: Proceedings of the IEEE Virtual Reality Conference; 2006 March 25-29; Alexandria, VA.
  47. Perani D, Fazio F, Borghese NA, et al. Different brain correlates for watching real and virtual hand actions. Neuroimage. 2001;14:749–758
  48. Janssen P, Vogels R, Orban GA. Selectivity for 3D shape that reveals distinct areas within macaque inferior temporal cortex. Science. 2000;288:2054–2056
  49. Decety J, Grezes J, Costes N, et al. Brain activity during observation of actions: influence of action content and subject's strategy. Brain. 1997;120:1763–1777
  50. Decety J. Do imagined and executed actions share the same neural substrate?. Brain Res Cogn Brain Res. 1996;3:87–93
  51. Sakata H, Taira M, Kusunoki M, Murata A, Tanaka Y. The TINS Lecture: the parietal association cortex in depth perception and visual control of hand action. Trends Neurosci. 1997;20:350–357
  52. Inoue K, Kawashima R, Satoh K, et al. PET study of pointing with visual feedback of moving hands. J Neurophysiol. 1998;79:117–125
  53. Ketcham CJ, Dounskaia NV, Stelmach GE. The role of vision in the control of continuous multijoint movements. J Mot Behav. 2006;38:29–44
  54. Rossetti Y, Desmurget M, Prablanc C. Vectorial coding of movement: vision, proprioception, or both?. J Neurophysiol. 1995;74:457–463
  55. Saunders JA, Knill DC. Humans use continuous visual feedback from the hand to control both the direction and distance of pointing movements. Exp Brain Res. 2005;162:458–473
  56. Sergio LE, Scott SH. Hand and joint paths during reaching movements with and without vision. Exp Brain Res. 1998;122:157–164
  57. Knapp JM, Loomis JM. Limited field of view of head-mounted displays is not the cause of distance underestimation in virtual environments. Presence. 2004;13:572–577
  58. Loftus A, Murphy S, McKenna I, Mon-Williams M. Reduced fields of view are neither necessary nor sufficient for distance underestimation but reduce precision and may cause calibration problems. Exp Brain Res. 2004;158:328–335
  59. Creem-Regehr SH, Willemsen P, Gooch AA, Thompson WB. The influence of restricted viewing conditions on egocentric distance perception: implications for real and virtual indoor environments. Perception. 2005;34:191–204
  60. Gonzalez-Alvarez C, Subramanian A, Pardhan S. Reaching and grasping with restricted peripheral vision. Ophthalmic Physiol Opt. 2007;27:265–274
  61. Helbig CR, Gabbard C. What determines limb selection for reaching?. Res Q Exerc Sport. 2004;75:47–59
  62. Weiss PL, Rand D, Katz N, Kizony . Video capture virtual reality as a flexible and effective rehabilitation tool. J NeuroEng Rehabil. 2004;1:12
  • a NDI International, 103 Randall Dr, Waterloo, ON, N2V 1C5, Canada.
  • b Kaiser XL50; Rockwell Collins, 79 Leonard St, London EC2A 4QS, United Kingdom.
  • c Computer Assisted Rehabilitation Environment, MOTEK Medical BV, Keienbergweg 77, 1101 GE Amsterdam.
  • d Softimage Co, 3510 St Laurent, Montreal, QC, H2X 2V2, Canada.
  • e National Instruments Corp, 11500 N Mopac Expwy, Austin, TX 78759-3504.
  • f SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.

 Supported by the Canadian Foundation for Innovation (project no. 202524).

 We certify that no party having a direct interest in the results of the research supporting this article has or will confer a benefit on us or on any organization with which we are associated, and we certify that all financial and material support for this research (eg, National Institutes of Health or National Health Service grants) and work is clearly identified on the title page of the article.

 Reprints are not available from the author.

PII: S0003-9993(09)00080-X

doi:10.1016/j.apmr.2008.10.030

Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 5 , Pages 793-802, May 2009

Article Tools

* Image Usage & Resolution