Neurorehabilitation and Neural Repair

http://nnr.sagepub.com/

Stroke Affects Locomotor Steering Responses to Changing Optic Flow Directions Anouk Lamontagne, Joyce Fung, Bradford McFadyen, Jocelyn Faubert and Caroline Paquette Neurorehabil Neural Repair 2010 24: 457 originally published online 12 January 2010 DOI: 10.1177/1545968309355985

The online version of this article can be found at: http://nnr.sagepub.com/content/24/5/457

Published by:

http://www.sagepublications.com

On behalf of:

American Society of

Additional services and information for Neurorehabilitation and Neural Repair can be found at:

Email Alerts: http://nnr.sagepub.com/cgi/alerts

Subscriptions: http://nnr.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://nnr.sagepub.com/content/24/5/457.refs.html

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 Research Articles Neurorehabilitation and Neural Repair Affects Locomotor Steering 24(5) 457­–468 © The Author(s) 2010 Reprints and permission: http://www. Responses to Changing Optic Flow sagepub.com/journalsPermissions.nav DOI: 10.1177/1545968309355985 Directions http://nnr.sagepub.com

Anouk Lamontagne, PhD,1 Joyce Fung, PhD,1 Bradford McFadyen, PhD,2 Jocelyn Faubert, PhD,3 and Caroline Paquette, PhD1

Abstract Background. Stroke patients manifest steering difficulties during walking, which may arise from an altered of visual motion. Objective. To examine the ability of stroke patients to control their heading direction while walking in a virtual environment (VE) describing translational optic flows (OFs) expanding from different directions. Methods. The authors evaluated 10 stroke patients and 11 healthy people while they were walking overground and visualizing a VE in a helmet- mounted display. Participants were instructed to walk straight in the VE and were randomly exposed to an OF having a focus of expansion (FOE) located in 5 possible locations (0°, ±20°, and ±40° to the right or left). The body’s center of mass (CoM) trajectory, heading direction, and horizontal body reorientation were recorded with a Vicon-512 system. Results. Healthy participants veered opposite to the FOE location in the physical world, with larger deviations occurring at the most eccentric FOE locations. Stroke patients displayed altered steering behaviors characterized either by an absence of CoM trajectory corrections, multiple errors in the heading direction, or systematic veering to the nonparetic side. Both groups displayed relatively small CoM trajectory corrections that led to large virtual heading errors. Conclusions. The control of heading of locomotion in response to different OF directions is affected by stroke. An altered perception of heading direction and/or a poor integration of sensory and motor information are likely causes. This altered response to OF direction while walking may contribute to steering difficulties after stroke.

Keywords heading, hemiparesis, kinematics, virtual reality, visual motion, walking

Introduction Individuals who sustained a stroke are especially chal- lenged in their daily activities as they have difficulties The control of steering, or heading direction, is a require- adapting their locomotion to environmental constraints. They ment for goal-directed locomotion, allowing individuals to show disrupted eye and body movement coordination pat- walk in the desired direction while avoiding obstacles along terns and altered walking trajectories when executing their path. Such control depends heavily on vision that pro- preplanned or cued changes of direction while walking, vides information about the characteristics of the surrounding suggesting disrupted steering abilities.5,6 The loss of bal- world and self-motion. Optic flow (OF) is a predictable pat- ance from executing a sudden turn of the head or body can tern of visual motion generated at the moving eye during result in a fall.7 Although motor deficits undoubtedly impair self-motion.1 Together with the perceived goal direction, it is one of the visual cues that can be used to control heading 1School of Physical and Occupational , McGill University and direction while walking.2-4 When walking straight, the OF Jewish Rehabilitation Hospital (Feil & Oberfeld/CRIR) Research Center, Montréal, Canada describes a radial pattern of expansion with a focus of 2 4 Laval University and Québec Rehabilitation Research Institute (CIRRIS), expansion (FOE) located in the direction of heading. When Québec, Canada a change of direction is needed, the OF theory stipulates that 3School of Optometry, University of Montréal, Montréal, Canada the FOE is realigned with the desired heading direction.1 In support of this theory, it has been shown that offsetting the Corresponding Author: Anouk Lamontagne, Jewish Rehabilitation Hospital (Feil & Oberfeld/ FOE of OF with prisms or through virtual reality causes CRIR) Research Centre, 3205 Place Alton-Goldbloom, Laval (Québec), healthy individuals to veer from their initial trajectory when H7V 1R2, Canada instructed to walk straight ahead.2-4 Email: [email protected]

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 458 Neurorehabilitation and Neural Repair 24(5) steering abilities, sensory deficits and altered sensorimotor Setup and Procedures integration are also likely to be involved.5,6 The ability to process different types of visual motion was shown to be Participants were evaluated while walking overground at altered by lesions in the occipitoparietal, parietotem- their comfortable speed in a large open space (12 m × 8 m), poral, and posterior parietal regions possibly involving the with their walking aid when applicable (see walking aid MT (V5) complex.8-11 It would also be affected by a lesion description, Table 1). They were visualizing, in a helmet- of the frontal lobe.12 Whether such altered perception of OF mounted display (HMD) unit, a VE representing a large (40 direction translates into poor steering abilities after stroke, m × 25 m) and symmetrical room created in Softimage XSI however, is not known. (Figure 1A). The virtual room was created to provide a rich- The goal of this study was to examine the ability of textured VE and had pillars at its extremities,3 which individuals with stroke to control their heading direction were separated by 8 m. The HMD (Kaiser ProView while walking in a virtual environment (VE) describing XL50) had a field of view of 50° diagonal, 30° vertical by translational OFs expanding from different directions. It 40° horizontal. Three-dimensional body coordinates were was hypothesized that these individuals would display an recorded using a 10-camera Vicon-512 motion capture altered control of their heading direction in response to the system (Denver, CO). Passive reflective markers were shifts in the foci of expansion, as reflected either by located on specific body landmarks according to the Plug- reduced adjustments and/or directional errors in their In-Gait Model from Vicon, except for the head that was walking trajectories. represented by a 3-marker model with markers located on the front, left side, and right side of the HMD. Location of the body’s center of mass (CoM) was calculated based on a Methods 15-segment kinematic model and the anthropometric char- Participants acteristics of the participants. The 3D head coordinates were tracked in real time by the Tarsus real-time engine A convenience sample of 10 patients with stroke and 11 from Vicon and fed to the Caren-2 virtual reality system healthy persons participated in the study (Table 1). The 2 from Motek (Motek Medical, Amsterdam). This feedback groups were of the same age (t test, P = .12) and presented system synchronized the virtual scene in real time with similar anthropometric characteristics. The inclusion crite- head motions through the physical space (Figure 1B). When ria for the patients were the following: a first unilateral applicable, a visual perturbation could also be superim- stroke in the middle cerebral artery (MCA) territory with posed onto the motion of the scene caused by the head residual deficits, confirmed by computerized tomography; displacements of the participants. The delay in updating the walking independently over 10 m, with or without a walk- virtual reality display in the HMD is less than 10 ms. The ing aid, at a walking speed inferior to 1.0 m/s; and presenting sampling frequency was set at 120 Hz. with significant motor deficits of the affected lower limb, as As they were walking forward, participants were exposed reflected by a Chedoke-McMaster leg and foot impairment to a translational OF that was expanding from a FOE located inventory score of less than 7.13 Exclusion criteria for the in 5 possible locations (Figure 1C): -40° and -20° to the left patients were the following: cognitive problems compro- or nonparetic side, 0°, as well as +20° and +40° to the right or mising the ability to give informed consent and to follow paretic side. A FOE angular offset (Q) was obtained by trans- instructions as well as the presence of a visual field defect, lating the scene laterally (x) based on the participant’s as assessed by the Goldman visual field and/or equivalent forward displacement (y), such that x = ytan(Q). Based on the computerized test. Patients with stroke were also free of forward walking speed of the participants, this corresponds visual spatial neglect, as assessed by the Bell’s Test14 or Star to average rates of mediolateral translation of the scene of Cancellation Test,15 except for 2 of them (Table 1, patients 0.32 m/s (±40° FOE) and 0.14 m/s (±20° FOE) for the S3 and S4) who were enrolled in the study before neglect patients and to 0.64 m/s (±40° FOE) and 0.28 m/s (±20° was confirmed. When appropriate, the behavior of these 2 FOE) for the healthy controls. Participants were instructed to patients is described separately. Both healthy participants walk straight in the VE, that is, to “walk straight with respect and those with stroke were free of visual problems not cor- to the scene they were visualizing in the HMD,” as opposed rected by eyewear as well as of other neurological or to the physical world. For a FOE located at 20° to the right, orthopedic conditions interfering with locomotion. All par- for instance, participants should perceive their virtual head- ticipants signed an informed consent document, and the ing direction as translating forward and toward the right project was approved by the Montréal Center for Interdisci- (Figure 1D). In an attempt to walk straight, they should veer plinary Research in Rehabilitation (CRIR) establishments’ to the left in the physical world, such that their walking tra- Research Ethics Board. jectory remains around neutral (0°) in the virtual world. They

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 a a

2 1 3 2 1 3 3 2 1

None Group

ed psule, psule,

c oparietal, oparietal, oparietal,

capsulethalamic thalamus Location of ovascular Accident ovascular oparietal junction, oparietal junction,

postinternal capsule region, insular cortex basal ganglia caudate basal ganglia; spar and thalamus and lenticular nucleus Cerebr Capsulothalamic Parietal Subcortical, Subcortical, Parietal periventricular

Tempor Frontotempor Frontotempor external ca Parietal, Parietal

enchymal Etiology hemorrhage hemorrhage hemorrhage Ischemia Intrapar Ischemia, lacunar Ischemia, Ischemia Ischemia Ischemia + small Ischemia Massive Ischemia + small Ischemia Ischemia

L L L L L R R R R R — — — — Lesion Side of

Y Y N N N N N N N N — — — — (Y/N) Neglect

e

— —

Strok

3.0 Months 1.0 Months 2.2 Months 3.4 Months 2.9 Months 2.1 Months Time Since 1-51 Months 27.7 Months 14.0 Months 19.0 Months 51.0 Months 12.6

b

(m/s) Speed

0.20 (Quad) 0.15 (Quad) 0.34 0.20 (Cane) 0.80 0.35 (Cane) 0.18 (Quad) 0.3 (Cane) 0.60 (2 Canes) 0.66 0.44 0.23 0.20-0.80 0.73 0.18 0.50-1.00

70 72 90 50 75 73 68 48 60 96 70 15 79 13 (kg) 48-96 63-104 Weight

12 11 172 178 154 162 173 176 152 155 152 185 166 169 (cm) Height 152-185 155-183

F F M M M M M M M M — — — (F/M) 8M/2F 7M/4F Gender

6 79 71 73 68 50 79 66 73 79 51 69 11 63 Age 50-79 50-72 (years)

11)

viation: SD, standard deviation. standard SD, viation:

= 10) (n =

The quad cane is a 4-legged cane. Initial scans were negative. A subsequent report states the presence of a subcortical infarct in this patient having a contralateral sensorimotor involvement of the hemibody with mild dysarthria and a contralateral sensorimotor involvement in this patient having of a subcorticalA subsequent report states the presence infarct negative. Initial scans were Individuals with stroke were assigned to different groups depending on their steering behaviors, as described in the Results section. (1) Minimal Veering Group; (2) Incongruous Veering Group; and (3) Group; Veering (2) Incongruous Group; Veering (1) Minimal as described in the Results section. depending on their steering behaviors, groups assigned to different were Individuals with stroke S10

S3 S8 S9 (n Table 1. Participant Characteristics Table Stroke S1 SD Range a b c S2 S4 S5 S6 S7 Mean dysphagia.

Nonparetic Veering Group. Veering Nonparetic Controls SD Range Mean Abbre

459

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 460 Neurorehabilitation and Neural Repair 24(5)

Figure 1. This figure shows (A) the virtual room used for the experiment, (B) the control algorithm of the virtual environment (VE) is based on real-time tracking of head coordinates, (C) the different locations of the focus of expansion (FOE) used in the present experiment, and (D) a schematic representation of perceived, physical, and virtual walking trajectories. were provided with 3 to 5 practice trials, with and without were recorded, for a total of 15 walking trials. Because perturbations, before the experiments. We also confirmed of their limited endurance, participants with stroke suc- verbally that they understood the instruction, both prior to ceeded in performing 2 to 3 trials per FOE location. An and after the data collection. Three trials per FOE location assistant was standing next to or walking along with the

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 Lamontagne et al. 461

Figure 2. Individual trials representing body center-of-mass (CoM) trajectory in the physical world (left panel) as well as heading orientation in the physical (middle panel) and virtual worlds (right panel) in (A) 1 healthy young individual and (B) 1 elderly control. CoM trajectories are illustrated in the anteroposterior (A/P) direction as a function of the mediolateral (M/L) direction. Heading angular orientations are illustrated as a function of CoM displacement along the A/P axis. Note that CoM and heading traces illustrated in this figure are referred to as positive when the participants are walking to the right. Trials at ±40° in the healthy young individual were recorded only for a short distance because of large M/L deviations that were limited by the physical constraints of the room. Because of the forward and backward oscillations in body CoM position at gait initiation that lead to inconsistent heading orientations, the first 50 frames of each trials are not represented, so heading traces do not start at neutral position. participants at all times, and no falls occurred during the means that participants overcorrected their walking trajec- experiment. tory. Within- and between-participant variabilities were Outcome variables retained for analysis included CoM obtained using standard deviation values calculated across trajectory and horizontal orientations of head and feet with trials and participants, respectively. Virtual heading errors and respect to the physical and virtual world’s coordinates. Dis- within-participant variability were compared across groups placement and angular orientation toward the right (controls) (stroke vs controls) and conditions (FOE location, 5 levels) or the paretic side (stroke) are referred to as positive. In using a 2-way analysis of variance for repeated measures. addition, heading direction was computed as the instanta- The relationships of the virtual heading errors with gait neous angular orientation of body CoM trajectory along the speed and age were assessed using Pearson correlation mediolateral (x-axis) versus the anteroposterior (y-axis) coefficients. Statistics were performed in Statistica with a direction. For every trial, heading orientation in the virtual level of significance accepted at P < .05. world, also termed as virtual heading error, was estimated as the mean heading orientation measured between 2 m and 3 m of forward walking. When 3 m of forward walking Results could not be reached, the mean value between 2 m and max- Examples of CoM trajectory modulation of healthy partici- imal forward walking distance was used instead. Mean pants in response to different FOE locations are illustrated virtual heading errors were thereafter averaged across trials in Figure 2. To help understand what a typical profile of and participants. By convention, a negative heading error modulation looks like, an example of a healthy elderly con- refers to an underestimation of the correction needed to reach trol is contrasted with that of a healthy young individual. a virtual heading of 0°, whereas a positive heading error For both participants, the CoM trajectory and heading

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 462 Neurorehabilitation and Neural Repair 24(5)

Figure 3. Individual examples of body center-of-mass (CoM) trajectory in the physical world (left panel) as well as heading orientation in the physical (middle panel) and virtual worlds (right panel) in participants with stroke from the (A) Minimal Veering Group, (B) Incongruous Veering Group and (C) Nonparetic Veering Group. Axis labels and abbreviation are the same as in Figure 2, with the exception of positive CoM and heading traces indicating a displacement to the paretic side. Participants belonging to each group are identified as per their participant number in Table 1 (S1 to S10). Note that S2 is not shown in this figure and displayed a steering behavior similar to healthy controls. orientation in the physical world were characterized by a smaller heading corrections in the physical world, resulting deviation in the direction opposite to the FOE location (left in larger heading errors in the virtual world. and middle panels). The amplitudes of the CoM trajectory In comparison with healthy controls, stroke patients and the heading corrections were scaled to the magnitude of presented with a variety of steering behaviors, including the perturbation, being larger for the 40° than for the 20° heading corrections that were of the wrong magnitude and/ and 0° perturbations. The healthy young individual dis- or in the wrong direction (Figure 3). Based on the steering played large yet incomplete heading corrections in the behavior, patients were classified into 3 different groups. physical world (middle panel), resulting in heading errors The Minimal Veering Group (n = 3) presented with little or that were biased toward the FOE location in the virtual no modulation of their CoM trajectory and heading orienta- world (right panel). This participant’s heading error, how- tion in the physical world when exposed to FOE shifts. As ever, decreased in the virtual world and converged toward a result, they displayed large heading errors in the virtual the midline (0°) as he continued walking forward. In com- world that were approaching the magnitude of the FOE parison, the healthy elderly control presented with much shifts. The Incongruous Veering Group (n = 3) displayed

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 Lamontagne et al. 463

Figure 4. Individual trials representing CoM trajectory as well as head and foot horizontal orientation in the physical and virtual worlds for 1 healthy control and 1 patient (S4) with visuospatial neglect. multiple and inconsistent directional errors, with heading a healthy control and a stroke patient with visuospatial corrections being made erratically sometimes toward the neglect. The CoM trajectory of the healthy control veered in paretic and sometimes toward the nonparetic side for similar the direction opposite to the FOE location in the physical FOE perturbations. The Nonparetic Veering Group (n = 3) world, but the correction remained incomplete, leading to veered to their nonparetic side in the physical world for some deviation error in the virtual world. The feet were most trials, regardless of the FOE location. This means that aligned with the CoM trajectory in the physical world, but when the FOE was on the paretic side, participants cor- the head showed few changes in its orientation with FOEs rected their heading in the expected direction by veering to of changing location. This head and foot orientation pattern their nonparetic side. When the FOE was shifted to the non- is typical of what was observed in the healthy control group paretic side, however, they veered in the wrong direction, as well as in most participants with stroke. Alternatively, that is, the ipsilateral nonparetic side. This resulted in larger the patient with visuospatial neglect represented in Figure 4 virtual heading errors for the paretic than the nonparetic invariantly veered to the nonparetic side, irrespective of the side. Note that the 2 participants with stroke who had visuospa- FOE location. As for the other patients in group 2, a direc- tial neglect (S3 and S4) were included in the Minimal Veering tional error arose when the FOE was located toward the Group and the Nonparetic Veering Group, respectively. One nonparetic side, which did not cause the patient to veer remaining participant with stroke (S2), not included in the 3 toward the paretic side in the physical world, as would be groups, displayed behavior identical to that of the healthy expected normally. For the patient shown in Figure 4, his controls. head and nonparetic foot were oriented toward the nonpa- An example of CoM trajectory modulation as well as of retic side in the physical world, whereas the paretic foot head and foot orientation is further detailed in Figure 4 for remained in a more toes-out and widened position.

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 464 Neurorehabilitation and Neural Repair 24(5)

Figure 5. (A) Mean (-1 SD) virtual heading errors as well as (B) mean within-participant and (C) between-participant variability of virtual heading responses for the patients and controls across the different locations of the FOE. In A, the negative polarity of the virtual heading errors indicates an undercorrection of the heading trajectory with respect to the virtual world’s coordinates. Statistically significant results are indicated in A and B. No standard deviation error bars or statistical analyses can be computed for between-participant variability in C. Abbreviations: SD, standard deviation; FOE, focus of expansion.

The average heading orientations in the virtual world, or changing direction. Virtual heading errors increased at more heading errors, were compared between patients and healthy eccentric FOE locations, both for patients and healthy par- participants across FOE locations (Figure 5). All partici- ticipants (P < .01). Given the different behaviors observed pants undercorrected their heading in response to FOEs of in the stroke patients, their mean virtual heading errors

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 Lamontagne et al. 465 were found to be similar to those of the healthy participants occipital, parietal, and temporal cortical areas while view- (P > .05), despite a tendency for larger heading errors in the ing radial OFs.17 The human MT (V5) complex, and more patients at ±40° of FOE perturbation. Within-participant and specifically the MST area, were shown to be highly sensi- between-participant variabilities in the heading responses, tive to changes in the characteristics of the OF stimuli.17,18 however, were larger in the patients than in the healthy group Discriminating heading, depending on the nature of the (P < .05, within-participant). Heading errors of patients and visual stimulus, would involve a complex network compris- healthy participants were not correlated with walking speed ing posterior brain areas such as KO and V3a,19-21 the MT (R2 = 0.00-0.07; P > .05) or age (R2 = 0.00-0.19; P > .05). A complex,20,22 temporoparietal and occipitotemporal areas,19,20 qualitative analysis of the heading behavior of patients with and even frontal areas12,22 and subcortical/limbic struc- respect to the lesion site or the use of walking aids did not tures.19 Altered processing of visual motion direction has reveal any identifiable relationships (Table 1). been previously reported in participants with a stroke in the occipitoparietal, parietotemporal, or posterior parietal brain areas, which were all believed to involve the MT com- Discussion plex.8-11 In the present study, the MCA territory lesions This study provides the first evidence that individuals who would spare the posterior visual areas supplied by the pos- sustained a stroke can manifest altered locomotor steering terior cerebral artery, but the MT complex, potentially behaviors when exposed to OFs expanding from different supplied by both the MCA and posterior cerebral artery, locations. Such altered behaviors are reflected by abnor- could not be excluded. Some of our patients also presented mally small adaptations and/or by directional errors in the hemorrhagic events, which potentially extended the lesion participants’ walking trajectories as well as by an increased beyond the MCA territory. It is noteworthy that our patients variability across trials. The present results are consistent who displayed altered steering behaviors had lesions that with the attenuated and/or altered gait speed modulation invariantly involved the parietal lobe and/or thalamus. The responses of participants with stroke when exposed to OFs parietal area is a key site for the integration of multiple of changing speeds.16 sources of sensory information,23,24 and it is involved in the It has been shown that healthy young individuals control of steering.25 The thalamus relays sensory informa- exposed to an artificially shifted FOE while walking veer in tion to different cortical areas, contributing crucially to the direction opposite to the goal in the physical world so motor actions. It seems reasonable to assume that lesions in that a virtual heading error close to zero can be reached.4 such areas affect multisensory integration and can alter We have also shown that the larger the shift in the FOE, the heading behaviors. Some of our patients also displayed larger the walking trajectory deviation in the physical lesion involving the frontotemporal region. There is evi- world, and the larger the virtual heading error.3 In other dence from intact or damaged brain studies that the frontal words, the corrections in the physical walking trajectory are lobe,12 and more specifically the dorsal premotor area,22,25 insufficient to lead to straight walking trajectories in the is involved in heading identification. Those studies involved virtual world, and the heading errors increase at more a ground plane or a radial OF pattern, which can resemble eccentric locations of FOE. Present data collected in healthy the type of stimulus used in the present study. According to controls who are older than the young individuals tested Peuskens et al,22 such activation in the dorsal premotor area previously in other studies do support these observations. could be “ . . . a final stage in a parieto-premotor visuomotor connection, specifically related to transformation of head- ing information . . . into motor schemes.”(p 2460) In light of Brain Lesion and Altered those findings, it appears likely that our patients had brain Heading Perception lesions that altered the perception and/or integration of An absence of response characterized the Miminal Veering heading information. Group, and multiple heading errors were displayed by the Incongruous Veering Group. No qualitative relationship could be established between the patterns of heading Altered Sensorimotor Integration responses and the localization of the brain lesion. It cannot and Visuomotor Control be excluded that the different heading behaviors (absence The present paradigm also involved a sensory scenario in of responses vs errors) may just be different expressions of which visual stimuli were incongruent with proprioceptive the same inability to perceive or use OFs to complete the and vestibular information. The central nervous system has heading task. It is worth investigating, however, whether the ability to reweight relevant sources of sensory informa- the damaged brain areas of our participants could lead to an tion in a context-dependent fashion to maintain postural altered discrimination of visual motion direction and head- equilibrium.26-28 The transient aftereffect of self-motion ing. The intact human brain shows activation in the experienced after treadmill running29 is another indirect

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 466 Neurorehabilitation and Neural Repair 24(5) evidence of such a sensory reweighting process. The ability participants might have had far extrapersonal space neglect to reweight the relevant source of information appears to be that was not picked up by paper-pencil tests.43 reduced in older individuals30-32 and after stroke,33 so that those with stroke may be unable to upregulate vision and resolve the sensory conflict to control their heading. It was Limitations also suggested that gaze reorientation is a fundamental Our results outline the within-individual and interindi- component of a locomotor steering synergy while chang- vidual variability of behavior usually observed in stroke. ing direction34,35 and that active gaze would contribute to Given the exploratory nature of this study, inclusion cri- identifying heading by providing extraretinal cues.36 Pas- teria were purposely kept large, but factors such as walking sive viewing of OFs also triggers eye movements for the capacity, chronicity, and the extent of the brain lesion may purpose of image stabilization.37 This emphasizes the inti- have contributed to the heterogeneity of behaviors. The mate relationship between gaze control, visual perception, nature of the brain imaging tool (CT scans) and variabil- and the control of heading. In those with stroke, the coor- ity in the depth of the information available in the dination of gaze and posture during locomotor steering patients’ charts also limited our interpretation of results was shown to be disrupted.5 This suggests that both per- in terms of the lesion site and its effects on heading ceptual and motor aspects of visuomotor control could behaviors. It is hoped that higher resolution imaging tools contribute to their altered heading responses. will help shed light on the role of specific brain areas in locomotor steering behaviors. Factors such as attention and the sense of presence (sense of “being there,” within Visuospatial Neglect the VE), which have not been quantified in this study, may Perhaps one of the most unexpected findings of the present also differ between those with stroke and healthy partici- study was the behavior observed in the Nonparetic Veering pants and account for some of the differences between the Group, which consisted of walking trajectories consistently 2 groups. It could also be argued that gait speed, which deviating to the nonparetic side in the physical world. was slower in the stroke participants, was partly respon- Closer examination of present data reveals that the direc- sible for their altered behavior. Indeed, walking faster or tional error occurred when the FOE was located on the running, as compared with walking slower, reduces the nonparetic side, which should have caused a turn to the deviations in walking trajectory when walking with shifted paretic side. This is not likely to be a result of biomechani- OFs,44 blindfolded,45 or when presenting a vestibular cal factors because those with stroke tend to walk with a imbalance.46 However, the fact that larger heading horizontal rotational bias of their body toward their paretic errors were not associated with slower gait speeds in side, therefore, favoring a physical turn toward the paretic the present study suggests that speed is not an explana- side.6 It has also been demonstrated that looking to the side tory variable. It was unfortunate that gaze behavior at an object causes a slight curvature in one’s walking tra- could not be monitored in the present study. Although jectory toward the direction of gaze.38 In the absence of this limitation does not invalidate the present results, it reliable vision cues such as when walking blindfolded, a also cannot reveal the presence or absence of an altered walking trajectory deviation contralateral to the eye or head gaze behavior associated with the identification and rotation is observed instead.39 In the present study, head and control of heading direction in stroke participants. foot starting positions were aligned with the room’s coordi- nates before the participants started to walk, but the eyes might have been oriented in one or another direction. Even Conclusions and Future Directions if this was the case, however, the effect would be very lim- The present results show that stroke affects the control of ited as in Jahn et al,39 for instance, a gaze rotation of 35° heading of locomotion in response to translational OFs leads to a small trajectory deviation of 2° to 4°. Results of expanding from different directions. An altered perception our study also show that one of the participants of the Non- of heading direction and/or a poor integration of sensory paretic Veering Group had neglect, whereas the 2 others and motor information are likely causes. Gaze shifts, which were theoretically free of such problems. In neglect, the help in identifying heading direction through extraretinal paretic hemispace is less attended, and the subjective mid- cues, might also be involved, contributing in part to the line would be displaced to the nonparetic or ipsilesional altered heading behaviors. The altered heading responses to side.40 This predicts that those with neglect should steer OF direction observed in the stroke participants may likely toward their paretic side in an attempt to walk straight,41 a contribute to their steering difficulties. Future directions for prediction that is not supported by the present results. In research should include examining the effects of specific reality, there is conflicting evidence as to whether those circumscribed brain lesions on locomotor steering behav- with neglect veer to their paretic or nonparetic sides as iors as well as exploring the role of oculomotor control and they walk.41,42 It also remains possible that some of our the use of other sources of visual information, such as

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 Lamontagne et al. 467 perceived target location, in the perception and control of 9. Vaina LM, Soloviev S. First-order and second-order motion: locomotor heading after a brain lesion. neurological evidence for neuroanatomically distinct systems. Prog Brain Res. 2004;144:197-212. Acknowledgments 10. Braun D, Petersen D, Schonle P, Fahle M. Deficits and recov- We would like to thank the people who participated in this study. ery of first- and second-order in patients We are also thankful to Eric Johnston and Christian Beaudoin for with unilateral cortical lesions. Eur J Neurosci. 1998; their skilful technical assistance. 10:2117-2128. 11. Vaina LM, Cowey A, Eskew RT Jr, LeMay M, Kemper Authors’ Note T. Regional cerebral correlates of global motion percep- Present address for Caroline Paquette: Laboratory, tion: evidence from unilateral cerebral brain damage. Brain. Department of and , McGill University at 2001;124(pt 2):310-321. SMBD-Jewish General Hospital and Lady Davis Institute, Mon- 12. Billino J, Braun DI, Bohm KD, Bremmer F, Gegenfurtner KR. tréal, Canada. Cortical networks for motion processing: effects of focal brain lesions on perception of different motion types. Neuropsycho- Declaration of Conflicting Interests logia. 2009;47:2133-2144. The authors declared no conflicts of interest with respect to the 13. Gowland C, Stratford P, Ward M, et al. Measuring physical authorship and/or publication of this article. impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24:58-63. Funding 14. Gauthier L, Dehaut F, Joanette Y. The Bells test for visual This study was funded by the Québec Rehabilitation Research neglect: a quantitative and qualitative test for visual neglect. Network (REPAR), the Canadian Institutes of Health Research Int J Clin Neuropsychol. 1989;11:49-54. (CIHR) and the Heart and Stroke Foundation of Canada. A. 15. Halligan P, Wilson B, Cockburn J. A short screening test for Lamontagne is supported by a CIHR New Investigator Salary visual neglect in stroke patients. Int Disabil Stud. 1990;12:95-99. Award. Part of the equipment used for this experiment was funded 16. Lamontagne A, Fung J, McFadyen BJ, Faubert J. Modula- by an infrastructure fund (New Opportunities Funds) from the tion of walking speed by changing optic flow in persons with Canada Foundation for Innovation. stroke. J Neuroeng Rehabil. 2007;4:22. 17. Ptito M, Kupers R, Faubert J, Gjedde A. Cortical representa- References tion of inward and outward radial motion in man. Neuroim- 1. Gibson JJ. The visual perception of objective motion and sub- age. 2001;14:1409-1415. jective movement. 1954. Psychol Rev. 1994;101:318-323. 18. Morrone MC, Tosetti M, Montanaro D, Fiorentini A, Cioni 2. Harris JM, Bonas W. Optic flow and scene structure do not G, Burr DC. A cortical area that responds specifically to optic always contribute to the control of human walking. Vision flow, revealed by fMRI. Nat Neurosci. 2000;3:1322-1328. Res. 2002;42:1619-1626. 19. Beer J, Blakemore C, Previc FH, Liotti M. Areas of the 3. Sarre G, Berard J, Fung J, Lamontagne A. Steering behaviour human brain activated by ambient visual motion, indicat- can be modulated by different optic flows during walking. ing three kinds of self-movement. Exp Brain Res. 2002;143: Neurosci Lett. 2008;436:96-101. 78-88. 4. Warren WH Jr, Kay BA, Zosh WD, Duchon AP, Sahuc S. 20. Vaina LM, Soloviev S. Functional of head- Optic flow is used to control human walking. Nat Neurosci. ing perception. In: Vaina LM, Beardsley SA, Rushton S, eds. 2001;4:213-216. Optic Flow and Beyond. New York, NY: Kluwer Academic 5. Lamontagne A, Fung J. Gaze and postural reorientation in the Press; 2004:109-137. control of locomotor steering after stroke. Neurorehabil Neu- 21. Rutschmann RM, Schrauf M, Greenlee MW. Brain activation ral Repair. 2009;23:256-266. during dichoptic presentation of optic flow stimuli. Exp Brain 6. Lamontagne A, Paquette C, Fung J. Stroke affects the coor- Res. 2000;134:533-537. dination of gaze and posture during preplanned turns while 22. Peuskens H, Sunaert S, Dupont P, Van Hecke P, Orban GA. walking. Neurorehabil Neural Repair. 2007;21:62-67. Human brain regions involved in heading estimation. J Neu- 7. Hyndman D, Ashburn A, Stack E. Fall events among peo- rosci. 2001;21:2451-2461. ple with stroke living in the community: circumstances of 23. Meehan SK, Staines WR. Task-relevance and temporal syn- falls and characteristics of fallers. Arch Phys Med Rehabil. chrony between tactile and visual stimuli modulates cortical 2002;83:165-170. activity and motor performance during sensory-guided move- 8. Plant GT, Laxer KD, Barbaro NM, Schiffman JS, Nakayama ment. Hum Brain Mapp. 2009;30:484-496. K. Impaired visual motion perception in the contralateral 24. Macaluso E, Driver J. Multisensory spatial interactions: hemifield following unilateral posterior cerebral lesions in a window onto functional integration in the human brain. humans. Brain. 1993;116(pt 6):1303-1335. Trends Neurosci. 2005;28:264-271.

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010 468 Neurorehabilitation and Neural Repair 24(5)

25. Field DT, Wilkie RM, Wann JP. Neural systems in the visual 36. Wilkie R, Wann J. Controlling steering and judging heading: control of steering. J Neurosci. 2007;27:8002-8010. retinal flow, visual direction, and extraretinal information. J 26. Mergner T, Maurer C, Peterka RJ. A multisensory posture Exp Psychol. 2003;29:363-378. control model of human upright stance. Prog Brain Res. 37. Niemann T, Lappe M, Buscher A, Hoffmann KP. Ocular 2003;142:189-201. responses to radial optic flow and single accelerated targets in 27. Keshner EA, Kenyon RV, Langston J. Postural responses humans. Vision Res. 1999;39:1359-1371. exhibit multisensory dependencies with discordant visual and 38. Cutting JE, Readinger WO, Wang RF. Walking, looking support surface motion. J Vestib Res. 2004;14:307-319. to the side, and taking curved paths. Percept Psychophys. 28. Mergner T, Schweigart G, Maurer C, Blumle A. Human pos- 2002;64:415-425. tural responses to motion of real and virtual visual environ- 39. Jahn K, Kalla R, Karg S, Strupp M, Brandt T. Eccentric eye ments under different support base conditions. Exp Brain Res. and head positions in darkness induce deviation from the 2005;167:535-556. intended path. Exp Brain Res. 2006;174:152-157. 29. Durgin FH, Pelah A, Fox LF, Lewis J, Kane R, Walley KA. 40. Karnath HO. Subjective body orientation in neglect and the Self-motion perception during locomotor recalibration: more interactive contribution of neck muscle proprioception and than meets the eye. J Exp Psychol. 2005;31:398-419. vestibular stimulation. Brain. 1994;117(pt 5):1001-1012. 30. Bugnariu N, Fung J. Aging and selective sensorimotor strate- 41. Huitema RB, Brouwer WH, Hof AL, Dekker R, Mulder T, gies in the regulation of upright balance. J Neuroeng Rehabil. Postema K. Walking trajectory in neglect patients. Gait Pos- 2007;4:19. ture. 2006;23:200-205. 31. Mahboobin A, Loughlin PJ, Redfern MS. A model-based 42. Robertson IH, Tegner R, Goodrich SJ, Wilson C. Walking tra- approach to attention and sensory integration in postural con- jectory and hand movements in unilateral left neglect: a ves- trol of older adults. Neurosci Lett. 2007;429:147-151. tibular hypothesis. Neuropsychologia. 1994;32:1495-1502. 32. O’Connor KW, Loughlin PJ, Redfern MS, Sparto PJ. Postural 43. Azouvi P, Bartolomeo P, Beis JM, Perennou D, Pradat- adaptations to repeated optic flow stimulation in older adults. Diehl P, Rousseaux M. A battery of tests for the quantitative Gait Posture. 2008;28:385-391. assessment of unilateral neglect. Restor Neurol Neurosci. 33. Bugnariu N, Fung J. Stroke and selective sensorimotor strate- 2006;24:273-285. gies in the regulation of upright balance. Abstr Soc Neurosci. 44. Jahn K, Strupp M, Schneider E, Dieterich M, Brandt T. Visu- 2006;32:353. ally induced gait deviations during different locomotion 34. Hollands MA, Patla AE, Vickers JN. “Look where you’re going!”: speeds. Exp Brain Res. 2001;141:370-374. gaze behaviour associated with maintaining and changing the 45. Dickstein R, Ufaz S, Dunsky A, Nadeau S, Abulaffio N. direction of locomotion. Exp Brain Res. 2002;143: 221-230. Speed-dependent deviations from a straight-ahead path during 35. Reed-Jones R, Reed-Jones J, Vallis LA, Hollands M. The forward locomotion in healthy individuals. Am J Phys Med effects of constraining eye movements on visually evoked Rehabil. 2005;84:330-337. steering responses during walking in a virtual environment. 46. Brandt T, Strupp M, Benson J. You are better off running than Exp Brain Res. 2009;197:357-367. walking with acute vestibulopathy. Lancet. 1999;354:746.

Downloaded from nnr.sagepub.com at UNIVERSITE DE MONTREAL on August 19, 2010