59
Schedae 2010
Prépublication n° 5
|
Fascicule n° 1
Virtual environments and sensory
integration: Effects of aging and stroke
Nicoleta Bugnariu1, Joyce Fung2, 3
1
Faculty of Health Sciences, University of Ottawa, Ottawa, Ontario, Canada
2
School of Physical and Occupational Therapy, McGill University, Montreal, Quebec, Canada
3
Jewish Rehabilitation Hospital, Laval, Quebec, Canada
Research was carried out on the effects of aging and sensory motor deficits following strokes
with respect to the capacity of the central nervous system to resolve sensory conflicts created by
Virtual Reality (VR). The results of this research demonstrate that VR can be a valuable tool for
therapeutic interventions that require an adaptation to complex, multimodal environments. The
rehabilitation protocols include balancing training in virtual environments.
Les études qui ont été menées sur les effets du vieillissement et des déficits sensori-moteurs consécutivement aux accidents vasculaires cérébraux concernent la capacité du système nerveux central à résoudre les conflits sensoriels créés par la Réalité Virtuelle (RV). Les résultats de ces études
ont démontré que la RV peut être un instrument valable en tant qu’intervention thérapeutique ce
qui requiert l’adaptation aux environnements multimodaux complexes. Les protocoles de rééducation comportent des exercices d’équilibration dans les environnements virtuels.
Introduction
Loss of upright balance control resulting in falls represents a major health problem for older
adults and stroke survivors. Postural imbalance may arise not only from motor or sensory
impairments but also from the inability to select and reweight pertinent sensory information.
This chapter discusses the use of virtual environments to investigate basic mechanisms of
sensorimotor integration. In particular, the effects of age and stroke on the ability of the central
nervous system to resolve sensory conflicts and maintain balance will be addressed. Results
from two main studies will be presented to focus on the current use and beneficial potential of
implementing rehabilitation protocols that include balance training in virtual environments.
Nicoleta Bugnariu, Joyce Fung
« Virtual environments and sensory integration: Effects of aging and stroke »
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
60
1 Background
1.1 Balance control
One of the most challenging aspects of rehabilitation is the regain of balance control. The
task of postural control involves controlling the body’s position in space for the dual purposes
of stability and orientation. It requires the integration of sensory information to asses the
position of body in space and the ability to generate forces for controlling body position.
Postural control requires a complex interaction of musculoskeletal and neural systems. Musculoskeletal components include joint range of motion, spinal flexibility, muscle properties and
biomechanical relationships among linked body segments. Neural components essential to
postural control encompass sensory and motor processes, as well as cognitive planning and
higher level processing essential for adaptive and anticipatory aspects of postural control.
Sensory processes encompass the visual, vestibular and somatosensory systems, sensory
strategies that organize these multiple inputs and internal representations important for the
mapping of sensation to action. Motor processes include neuromuscular response synergies
and the selection and sequencing of muscles to be activated. Adaptive or reactive postural
control, also referred to as feedback control, involves modifying sensory and motor systems
in response to changing tasks and environmental demands. Anticipatory aspects of postural
control, also referred to as feedforforward control, pre-tune sensory and motor systems for
postural demands based on previous experience and learning.
The maintenance of upright equilibrium is essentially a sensorimotor integration task.
The central nervous system (CNS) has to generate task-specific and goal-directed complex
motor responses based on the selective and rapid integration of sensory information from
multiple sources. Since each sensory system has its own coordinate framework, specific
time delay and reliability, sensory conflicts may arise and represent situations in which the
CNS has to recalibrate the weight attributed to each particular sensory input. Humans can
tread on changing and uneven terrains without falling because the intact CNS can make
ongoing corrections.
Research has shown that in quiet stance with feet together, balance in the sagittale
plane is achieved mainly by the control of ankle torque whereas balance in the frontal plane
is predominantly controlled by changing hip torque [WIN 96]. Previous research on multidirectional surface translations showed similar control mechanisms with complex central and
peripheral organization of the postural responses [HEN 98 a, b]. Unexpected movement
of a support surface elicits rapid, automatic and coordinated postural responses that are
triggered primarily by somatosensory afferents depending on the velocity and direction of
perturbations [DIE 88; ING 95; RUN 98; FUN 05]. These responses are not merely segmental
reflexes organized at the level of the spinal cord, but rather depend on the integration of
proprioceptive, visual and vestibular information at many levels of the neuraxis [FUN 05;
MAC 99; HOR 96].
1.2 Contribution of sensory systems and their interactions
Effective postural control requires more than the ability to generate and apply forces for
controlling the body’s position in space. In order to know when and how to apply restoring
forces, the CNS must have accurate information of where the body is in space and whether
it is stationary or in motion. Somatosensory afferents, including mechanoreceptors, pressure
receptors, muscle spindles, Golgi tendon organs and joint receptors, provide information
regarding the body’s position and motion in space with respect to the support surface. Visual
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
61
inputs provide information concerning the position and motion of the head with respect to
the surrounding environment. With the semicircular canals sensing angular accelerations of
the head in different planes and the otoliths signaling linear position and acceleration of the
head with respect to gravity and inertial forces, the vestibular system provides a gravito-inertial
frame of reference for postural control. Somatosensors, being distributed throughout the
body, are critical for determining the orientation and configuration of the body segments.
In contrast, visual and vestibular receptors are located in a head that moves independently
from the trunk and limbs. Therefore, body configuration and orientation information from
head-based sensors is either limited, or must be derived. Somatosensory information from
the lower extremities and trunk is particularly important for maintaining balance when the
subject maintains contact with a large, rigid, and stable support surface [HOR 96]. Recent
modeling studies have provided a scheme of coordinate transformations such that head-based
sensors are mapped downwards from neck muscles to leg muscles, whereas somatosensory
afferents from the feet and legs are mapped upwards to the trunk [MER 97].
It was once thought that visual inputs were too slow to have any effect on the rapid
response due to sudden perturbations of stance, since the sensation of motion induced by
moving visual fields has a relatively long latency (-1 s), and the influence on body sway is too
slow [NAS 78]. Subsequent experiments have shown that visual information that conflicts with
those arising from other sensory channels can have a rapid and profound effect on postural
responses [VID 82; KES 04a]. For instance, when visual inputs are stabilized with respect to the
head during the time of the perturbation only, the initial triceps surae burst can be significantly
attenuated and forward sway increases. In contrast, merely closing the eyes during a postural
perturbation has no effect on the early evoked response or the performance, suggesting that
sensory context is an important factor in shaping the strategy for postural responses [VID
82]. Absence of vision under these conditions does not compromise postural performance
since other sensory channels provide sufficient information. In contrast, visual information
that conflicts with that from other sensory channels can have a rapid and profound effect on
postural responses. An example of such sensory conflict can be experienced when riding
a bus while watching the traffic outside the window and for a brief period not knowing for
certain if it is the bus that is moving forward or the traffic moving in the opposite direction.
If the bus is moving, certain postural muscle activation commands are necessary in order
to maintain balance. If the bus is stationary and the illusion of movement comes from the
moving traffic, then the same postural commands would be inappropriate. The influence
of moving visual fields on postural stability depends on the characteristics not only of the
visual environment, but also of support surface, including size of the base of support and
its rigidity or compliance [STR 06; KES 04b].
Within physiological limits, a central recalibration process exists to produce appropriate responses even in the presence of sensory conflicts, as for example when the visual
information is discordant with the information provided by the somatosensory system. The
following studies will present data form experiments that used virtual reality technology
in order to study the mechanisms of sensory integration, the ability of the CNS to resolve
sensory conflicts and maintain balance.
1.3 Virtual reality and rehabilitation
The current state of knowledge in rehabilitation emphasizes the need for intense task-related
practice to promote the re-acquisition of balance skills. Motor learning is promoted by factors
such as changing environmental contexts, alterations in the physical demands, problem
solving, random presentation of practice tasks, sufficient practice and patient empowerment
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
62
[WIN 91]. Virtual reality (VR) refers to a range of computing technologies that present a 3D
interactive simulation that occurs in real-time and presents artificially generated sensory
information (from vision and proprioception) in a form that can be perceived as similar to
real-world objects and events. Virtual reality is a valuable tool for therapeutic interventions
that require adaptation to complex, multimodal environments [KES 04b]. VR technology
with the capacity of simulating environments offers a new and safe way to not only increase
practice time but also to offer the varied environments and constraints needed to maximize
learning. Moreover, the “hi-tech” nature of the practice itself can be a further motivation
for the subjects. The term virtural ‘environment’ or VE is used to describe the simulation of
a visual 3-D environment presented to the subject via a monitor, a large screen, or through
a helmet-mounted display (HMD). In a VE, the simulated objects and events are not only
sensed, but the user can anticipate and react to them as though they were real. The user
often feels, at least to some degree, “present” in the simulated world, and this feeling of
presence is arguably the defining feature of the VR experience. VR systems have been
applied to the training of upper and lower extremities after stroke [HOL 99; DEU 01], to
improve mobility in persons with impaired spatial abilities or to train balance control [McC
02, SVE 04] and in vestibular rehabilitation [WHI 02].
2 Methodology
The common methodology described below was used in both studies to investigate the
effects of aging and stroke on the capability of the CNS to select appropriate sensorimotor
strategies and regulate balance while under conditions of sensory conflicts created by VR.
We also aimed to determine the effect of repeated exposures to VR on the ability of the
CNS to resolve sensory conflicts (Figure 1).
Procedure. During quiet stance, subjects were exposed to random visual and/or surface
perturbations consisting of ramp-and-hold tilts of 8o (peak velocity of 36o/s) in each direction
of the pitch and roll planes. Visual perturbations were induced by sudden movements of
a virtual environment (VE) viewed through a helmet-mounted display (HMD, Kaiser Optics
ProViewTM XL50, with a field of view of 50° diagonal, 30° vertical x 40° horizontal).
The VE consisted of a 3D rendered computer-simulated room generated by SoftImage
XSI on a CAREN 2.3.0 Workstation (Motek Inc) and complete with windows, columns, flooring
and ceiling textures. The perturbations consisted of: 1) visual-only, the VE was tilted but
the surface was stationary; 2) surface-only, the surface was tilted but the VE was fixed; 3)
discordant, visual perturbation was combined with synchronized surface perturbation in the
same direction, and 4) concordant, visual perturbation was combined with synchronized
surface perturbation in the opposite direction resembling the real life perception.
The support surface was mounted over a six-degree-of-freedom motion base servocontrolled by six electrohydraulic actuators [FUN 98]. The initial stance posture consisted
of weight evenly distributed between feet placed on 2 force-plates (AMTI OR6-7), heels 15
cm apart and feet oriented in a 15o toe-out position. Subjects were instructed to maintain
balance to the best of their abilities without taking steps where possible. If a step was taken,
the subjects were instructed to resume the initial stance posture. A total of 72 perturbation
trials were completed for a minimum of 1 hour VE immersion.
Data recording. A six-camera VICON 512 system (Oxford Metrics) was used to capture
3D position data at 120 Hz from 36 retroreflective markers placed over anatomical landmarks
and 4 markers placed on the movable platform. Ground reaction forces and moments from
the force-plates were acquired at 1,080 Hz.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
63
Fig. 1. Virtual environment viewed through a helmet-mounted display and
experimental setup used to create four sensory conditions
Eight bilateral muscles were instrumented with bipolar Ag-AgCl surface electrodes
to record electromyographic (EMG) signals using a Noraxon system: tibialis anterior (TA),
gastrocnemius medialis (MG), vastus lateralis (VL), semitendinosus (ST), tensor fascia latae
(TFL), erector spinae (ES) at the L3 level, neck extensor (NE) and neck flexor sternocleidomastoideus (SCM). EMG signals were amplified, digitized, band-pass filtered (10–400 Hz
low-pass) and sampled at 1,080 Hz. EMG signals were further full-wave rectified and lowpass
filtered at 100 Hz during offline analysis. Functional balance and mobility in terms of gait
velocity, ability to maintain tandem stance, timed repeated sit-to-stand, as described in a
physical performance battery [GUR 00] were assessed before and after VE exposure and
perturbation trials.
Data analysis. A biomechanical model (Plug-In Gait) was used in conjunction with
kinematic data and anthropometric measures to calculate the displacement of the body’s
center of mass (COM). Resultant center of pressure (COP) in the horizontal plane was
calculated as the weighted sums from the vertical force and anteroposterior (A/P) and
mediolateral (M/L) moments from the individual force plate. Inertial components in forces
and COP data due to movement of the support surface were corrected as described
previously [PRE 04]. Muscle latencies were determined as the first burst that exceeded
a threshold of two standard deviations above the background mean level and lasting at
least 50 ms, with an activation probability of at least 50%. Data from different trials of
each individual were ensemble-averaged across each one of the four testing conditions
(vision only, surface only, discordant and concordant) and each one of the directions of
pitch (toes-up/down) and tilt (left/right-down). These averages were pooled to produce a
population average (young, old, stroke) for each direction of perturbations and condition
of testing. To estimate the ability to adapt, average responses form the first 10 and last 10
trials were also calculated and compared.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
64
3 Effects of aging
Aging is associated with the decline in the integrity of many postural regulating systems
and these age-related changes in musculoskeletal, sensory systems, neural processing and
conduction of information could all potentially impact on postural control in older adults [HOR
89; MAC 89; MAK 96]. The health care costs associated with seniors’ falls are estimated at 1
billion dollars annually. The effects of aging and repeated exposures on the capability of the
CNS to select pertinent sensory information and resolve sensory conflicts were investigated
with virtual reality in the present study.
3.1 Subjects
Ten young adults (5 males and females of mean age 26±5.1 y.o.) and 10 older adults (5 males
and females of mean age 72 ±3.3 y.o.) participated in this study. Subjects were healthy with
no neurological problems, musculoskeletal conditions or motion sickness.
3.2 Results
3.2.1 Kinematics
Representative examples of COP and COM traces (thin and thick lines, respectively,
left-sided graphs), as well as group averages of COM (right-sided bar graphs) during
pitch and roll perturbations are shown in Figures 2 and 3, respectively. During visual-only
perturbations, minimal displacements of COP and COM were observed in both subject
groups. Surface perturbations, with or without visual perturbations, provoked displacements of COP and COM first in the direction of the perturbation, and then reversed to
oppose the perturbation for balance adjustment. Changes in COP always preceded and
encompassed those of COM. In young adults, COP and COM displacements were smallest
during perturbation where the visual and somatosensory stimuli were concordant. The
displacements were markedly larger in older adults and took longer duration to return to
the neutral positions. In older adults, during surface-only and discordant perturbations,
COM and COP excursions increased substantially and took longer or never return to a
neutral or stable position.
During visual-only perturbations, similar and minimal (2-10 mm) postural sway was
observed in both the young (gray) and old (black) groups of subjects (bar graphs, Figures 2
and 3). In general, across all other conditions of testing, older adults displayed significantly
larger COM peak-to-peak excursions (20 to 40 mm more than young adults). The presence
of sensory conflicts in surface-only and discordant perturbations induced significantly larger
COM excursions than concordant pertubations in both young and old adults. However, the
presence of sensory conflicts required a larger correction in older adults. During surface-only
perturbations, mean COM excursions for older subjects compared to young were 10-15 mm
larger in pitch (Figure 2) and 15-25 mm in roll (Figure 3) perturbations. During conditions
of discordant perturbations, COM excursions were 30-50 mm and 20-40 mm larger in old
subjects compared to young during pitch (Figure 2) and roll (Figure 3) perturbations. During
conditions of concordant perturbations, COM excursions were not significantly different
between old and young subjects. The average COP values (not shown) displayed similar
trends, although always larger in range as compared to the COM.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
65
Fig. 2. Pitch plane COP and COM responses in different sensory conditions. Representative example of
individual traces of COP (thin lines) and COM (thick lines) from one young and one old subject (left panel)
exposed to toes-up tilt of the surface. Bar graphs on the right panel show COM peak-to-peak excursions
(mean ± S.D.) averaged across 10 young subjects (gray bars) and 10 older subjects (black bars) in both toes-up
(left column) and toes-down (right column) directions.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
66
Fig. 3. Roll plane COP and COM responses in different sensory conditions. Representative example of individual
traces of COP (thin lines) and COM (thick lines) from one young and one old subject (left panel) exposed to
right-down roll of the surface. Bar graphs on the right panel show COM peak-to-peak excursions (mean ±
S.D.) averaged across 10 young subjects (gray bars) and 10 older subjects (black bars) in both left-down (left
column) and right-down roll (right column) directions.
3.2.2 EMG activity
The average EMG latencies of ventral muscles responding in the toes-up pitch direction during
surface-only, discordant and concordant perturbations in young and old adults are shown in
Figure 4. Young subjects showed similar latencies from the first 10 to the last 10 trials, and
thus are averaged across all trials of similar conditions (Figure 4, gray circles). Old subjects
showed increasingly earlier activations from the first 10 (Figure 4, black diamonds) to the last
10 trials (Figure 4, open triangles). No muscle activation was observed in young adults during
visual-only perturbations. In old adults, sporadic muscles activations were present during
visual-only perturbation but the activation probabilities of 50% was not reached, and thus
were not included further in the analysis. In young adults, muscle recruitment generally
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
67
Fig. 4. Single-session adaptation of muscle responses. EMG latencies (mean ± S.D.) of ventral muscles responding to toes-up pitch surface perturbations across young (gray circles) and old subjects. Note the decrease
in the latencies in old subjects from the first 10 (black diamonds) to the last 10 (open triangles) trials.
followed a distal-to-proximal sequence, regardless of perturbation direction or sensory
conflicts. The ankle muscles, TA and MG, were first to be activated during toes-up and
toes-down tilt, respectively, at a latency of 80-100 ms. VL and ST were activated 110-130
ms, followed by TFL and ES approximately 30-50 ms later, while the neck muscles SCM and
NE were activated slightly earlier at 150-170 ms.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
68
In older adults, the distal-to-proximal sequence of EMG activation was les consistent,
especially under sensory conflicts, during surface-only and discordant perturbations (Figure 4, black diamonds), where a reverse sequence was observed. In some older subjects,
following discordant visual and somatosensory stimuli, activation of neck muscles preceded
distal leg muscles by 25-70 ms. Generally, the EMG onset latencies of older adults, which
were already delayed as compared to young adults, were further prolonged in conditions
of sensory conflicts. However, adaptation occurred during the one-hour session such that
ankle EMG latencies were 20 ms shorter in the last 10/72 perturbations as compared to
the first 10 trials.
Conflicting visual and somatosensory stimuli modulate automatic postural responses in
both healthy young and old adults but the presence of sensory conflicts had a larger impact
on the selection of appropriate strategies for balance control in older adults. When the VE
was manipulated to provide distorted visual perception, i.e. during surface-only or discordant
perturbations, older adults took more steps, had longer EMG onset latencies as well as larger
COP and COM excursions. Postural instability as measured by COM excursions increase
markedly especially during discordant perturbations. Similar age-related postural instability
was also reported by Mahboobin et al. [MAH 05] who showed that optic flow induced larger
postural responses in old subjects than in subjects who had adapted from unilateral loss of
vestibular function. It is plausible that delayed or diminished vestibular and somatosensory
inputs in older adults increases their sensory thresholds to complex multimodal stimuli,
thereby inducing a greater reliance on visual inputs and making it more difficult for them to
respond selectively to visual and physical destabilization. Excessive reliance on visual input
may be a natural compensatory strategy to cope with poor balance in seniors, but it can be
problematic when the visual information is not reliable.
Visual-only perturbations elicit minimal postural responses in both young and old subjects
with only sporadic activations of muscles in older adults, suggesting that under normal
circumstances when there is a stable support surface, the somatosensory information is
weighted more in regulating upright posture. The use of HMD to deliver the visual perturbation might have also limited the influence of visual inputs. Postural responses coupled with
optic flow are less frequent when the optic flow is delivered in a central field of view, like
the HMD, as compared to BNAVE display with a full field of view [WHI 02; SPA 06]. The
influence of moving visual fields on postural stability depends on the characteristics not only
of the visual environment, but also of the support surface, including the size of the base of
support and its rigidity or compliance [AMB 89]. Somatosensory information from the lower
extremities and trunk is particularly important for maintaining balance when the subject
maintains contact with a large, rigid, and stable support surface [HOR 96]. Surface-only
perturbations present a postural challenge for both young and old. EMG onset latencies
can be delayed and prolonged, while postural sway increases, as compared to concordant
visual and somatosensory perturbations. Similar effects of visual stabilization were observed
on initial bursts of ankle muscles [KES 04a].
4 Effects of stroke
Sensory and motor impairments following cerebral vascular accidents can lead to poor
control of balance and mobility. The motor areas of the cerebral cortex are involved in
the preparation of voluntary movement, however, their role in postural mechanisms, and
in particular, in response to unexpected perturbations of stance is uncertain [MAS 92].
Nevertheless, EMG analysis has revealed disordered patterns of muscle activity, including
abnormal timing and sequencing of muscle activation as well as excessive cocontraction in
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
69
hemiplegic subjects reacting to external perturbations [DUN 87]. Some evidence has been
provided on misrepresentation of trunk orientation in subjects with neurologic injury [LUY 97].
Hence, the central problem may reside in sensorimotor integration which involves coordinate
transformations leading to a misinterpretation of the body in space. Stroke patients manifest
altered postural adjustments to voluntary head motions during standing [LAM 03]. Stroke
subjects have difficulty maintaining balance when exposed to perturbations during standing
or walking and theirs responses are not modulated by the direction of perturbation or by
the task requirement [FUN 03]. The Approximately two thirds of stroke patients are above
65 years old; therefore aging may introduce confounding factors for balance control after
stroke. We examined the effects of aging and sensory motor deficits following stroke on the
capability of the center nervous system to select pertinent sensory information and resolve
sensory conflicts created by virtual reality.
4.1 Subjects
Six stroke patients 62 to 76 years old and six age-matched healthy old adults participated
in this study. The patients were 7 to 27 months post stroke, had no neglect, and were able
to ambulate independently, with an assistive device (cane) with a mean speed of 0.57 m/s
(range 0.35 to 0.85 m/s).
4.2 Results
Representative examples of COP and COM traces (light and dark blue/red lines, for control/stroke patients respectively, left-sided graphs) are shown on Figure 5. Bar graphs on the
middle panels show COM peak-to-peak excursions (mean ± S.D.) averaged across groups
of subjects in both directions of perturbations. Bar graphs on the right panels show data
differences between first and last 10 trials. The displacements of the center of pressure
(COP) and body’s center of mass (COM) increased in the presence of sensory conflicts and
neurological injury (p < 0.05).
4.3 Clinical measures
With repeated exposures to VR-induced sensory conflicts, a general training effect associated
with less stepping responses and improved ability to maintain balance was observed in older
adults. In healthy older adults, the last 10 out of 72 perturbation trials had significantly reduced
COP and COM excursions as compared to the first 10 perturbations. The differences between
first and last 10 trials, although present were less evident in stroke subjects. The ability to
anticipate the physical constraints of the environment and adapt the balance behavior accordingly is the result of intricate sensorimotor integration. The cognitive processes range from
correctly perceiving and interpreting information from different body sensors (somatosensory,
vestibular and visual) to planning and coordinating the effectors appropriately to produce the
desired movement. The average number of steps taken by older subjects also decreased from
3 during the first 10 trials to one in the last 10 trials (p < 0.001). At the end of the 1h immersion
in the VE and repeated exposures to sensory conflicts perturbations, four old adults scored
1-2 points higher on their ability to maintain tandem stance (p < 0.05). A loss of stepping
response and reduced training effect was observed in stroke patients who did not step but
relied on external support (harness). However, five out of six stroke patients improved by 10
sec the ability to maintain independent tandem stance position at the end of the 1h immersion
in the VE and repeated exposures to sensory conflicts perturbations. It has to be noted that
the standard test position during experiments was side by side and not tandem.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
70
Fig. 5. Pitch plane COP and COM responses in stroke and control subjects. Representative example of
individual traces of COP (thin lines) and COM (thick lines) from one control healthy and one stroke subject
(left panels) exposed to pitch plane tilt (toes-up). Bar graphs on the middle panels show COM peak-to-peak
excursions (mean ± S.D.) averaged across groups of subjects in both directions of perturbations. Bar graphs
on the right panels show data differences between first and last 10 trials.
Maintaining tandem stance is a challenge for seniors and populations with decreased balance, and inability to do so is correlated with increased risk of falls [MAK 96]. Therefore,
the improvements noted have clinical significance. No change was observed in gait speed
and timed repeated sit-to-stand. All subjects tolerated well the use of the HMD with the
exception of one older subject who reported mild discomfort due to the tight fitting. All
subjects were able to complete the 1h immersion in the VE with no reports of nausea.
4.4 EMG activity
Representative examples of muscle activation sequence following toes up rotation in a healthy
control subject (left panel) and in stroke patient non paretic side (middle panel) and paretic
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
71
Fig. 6. Muscle responses to perturbations in different sensory conditions. A: Example of muscle activation
sequence following toes up rotation in a healthy control subject (left panel) and in stroke patient non paretic
side (middle panel) and paretic side (right panel). B: Group Mean ± SD for muscles onset latencies following
pitch rotations (toes-up) for healthy control subjects (blue) and stroke subjects (red) in different sensory
conditions: surface only, concordant and discordant in left, middle and right panels, respectively.
side (right panel) are showed in Figure 6A. Stroke subjects show under-activated responses
on the paretic leg and exaggerated responses on the non-paretic limb.
The average EMG latencies of ventral muscles responding in the toes-up pitch direction during surface-only, discordant and concordant perturbations in stroke (red circles)
and healthy old adults (blue diamonds) are shown in Figure 6B. Visual only perturbations
produced sporadic muscles activations in both old adults and stroke patients. In general,
aging disrupted the distal-to-proximal muscle recruitment sequence and the presence
of sensory conflicts and stroke exacerbated the inconsistencies. Neck muscles were the
first activated in both stroke and healthy age-matched subjects. EMG latencies of ankle
and hip muscles that were delayed in stroke subjects as compared to healthy older adults
during surface perturbations alone, were further prolonged by 40-60 ms in conditions
of sensory conflicts.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
72
Fig. 7. COM gain in different sensory conditions. Group Mean ± SD for COM gain relative to visual (filled
icons) and surface stimulus (open icons) in four different sensory conditions, for healthy young (blue) and old
subjects (red) and stroke patients (orange)
Appropriate muscles need to be activated in order to change the forces exerted through
the body and limbs in contact with the support surface. These motor responses are based
on selective and rapid integration of visual, vestibular and somatosensory information [VAN
99]. As the sensory conditions changed from one trial to another, and sensory conflicts
were present, an update of sensory weights to current conditions was necessary, so that
motor commands were based on the most precise and reliable sensory information available [MER 05; JEK 06]. Reweighting mechanisms used to resolve sensory conflicts are not
entirely understood. How the weights given to each sensory input are determined and how
various factors influence this process remain open questions. It has been suggested that
weighting depends on the sensory context (inconsistency between sensory signals) as well
as on the subject [VAN 99; BRI 07]. A commonly used technique to study sensory integration
and reweighting is to stimulate postural sway with visual and somatosensory stimuli and
to determine the frequency-response function which described gain and phase. Linear,
spectral analysis was performed for each trial by computing the individual Fourier transforms
of the time series of COM postural displacements and of the stimuli motion, either visual
or surface. For each one of the four sensory conditions, the frequency-response function
was computed by dividing the transform of the estimated COM by the transform of the
stimulus. The gain is the absolute value of the frequency-response function. For example,
a surface gain value of one will indicate that the displacement of body in space is equal to
the displacement of the support surface.
Figure 7 illustrates COM gain relative to the surface (open icons) and visual (filled icons)
stimulus in young (blue), healthy old adults (red) and stroke patients (orange). In general,
young adults displayed low visual gains and high values of surface gains. In conditions of
visual-somatosensory conflict, they increased surface gain and decreased visual gain, suggesting that young adults deal with the sensory conflict by either attempting to suppress visual
information altogether, or attributing more weight and increased reliance on somatosensory
feedback. Healthy older subjects and stroke patients displayed higher visual gains than young
adults regardless of the sensory conditions. Moreover, in conditions of sensory conflict, they
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
73
adopted an opposite strategy increasing the visual gain and lowering the surface gain. This
demonstrates an excessive reliance on visual inputs or a need to first stabilize their head, a
common strategy adopted by people with balance impairments.
Conclusions
We examined the effects of aging and sensory motor deficits following stroke on the capability
of the center nervous system to resolve sensory conflicts created by VR. Virtual reality can be
a valuable tool for therapeutic interventions that require adaptation to complex, multimodal
environments. When designing VR protocols involving multisensory modalities one should
be aware of the potential of sensory conflicts and their effects.
Conflicting visual and somatosensory stimuli can modulate automatic postural responses
in both healthy young and old adults. Aging affects the interaction of the somatosensory
and visual systems on the ability of the CNS to resolve sensory conflicts and to maintain
upright stance equilibrium. Compounding effects of age and neurological injury can skew
the sensory recalibration processes required for resolution of sensory conflicts toward an
excessive reliance on visual inputs. The presence of sensory conflicts has a significantly larger
impact on the regulation of balance in stroke subjects. By far the most challenging condition
was created by the discordant perturbations with longest onset latencies of distal muscles,
activation of neck muscles first and increased COM gain relative to the visual stimulus. This
strategy was observed in both stroke and healthy age-matched subjects. It demonstrates
an excessive reliance on visual inputs.
Visual dependence may be a compensatory strategy for coping with poor balance post
stroke. In addition to motor impairment, postural imbalance in patients with hemiparesjs may
be caused not only by elementary sensory impairment (visual, somatosensory, vestibular)
but also by the inability to solve sensory conflicts, to select pertinent sensory information.
Excessive reliance on visual input may be a natural compensatory strategy for coping with poor
balance in patients after stroke. The excessive reliance on vision may become problematic
when the visual information is not reliable.
The resolution of sensory conflicts is affected by aging and stroke but can be enhanced by
training. Repeated exposure to VR-induced sensory conflicts improves balance performance
in all healthy older subjects and to some extent in stroke subjects. Even with a one-hour
immersion in virtual environment and exposure to sensory conflicts, it is possible for the
CNS to recalibrate and adapt to the changes. A training program of longer durations is
needed to confirm sustainable long-term effects. Therefore, preventive and rehabilitation
programs targeting postural control in seniors and stroke patients should take into account
the possible impairment of sensory organization or sensorimotor integration and include
exercises to be performed in different VE under conditions of sensory conflicts.
Acknowledgements
This work was supported in part by the Canadian Institutes of Health Research and the Jewish
Rehabilitation Hospital (JRH) Foundation. N. Bugariu was funded by a Tomlinson postdoctoral
research fellowship award at McGill University. J. Fung is a William Dawson Scholar at the
School of Physical and Occupational Therapy, McGill University and the Director of Research at
the JRH, a site of the Montreal Research Center for Interdisciplinary Research in Rehabilitation
(CRIR). The authors would like to thank all participants, as well as Christian Beaudoin, Luncinda
Hughey, Eric Johnstone and Erika Hasler for their skilful help and assistance.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
74
References
[AMB 85] AMBLARD B., CREMIEUX J., MARCHAND A.R., CARBLANC A., “Lateral orientation and stabilization of
human stance: static versus dynamic visual cues”, Exp Brain Res, vol. 61 (1), pp. 21-37, 1985.
[BRI 07] BRIAN L.D., MICHEL G., “Feedforward versus feedback modulation of human vestibular-evoked
balance responses by visual self motion information”, J Physiol, vol. 582 (Pt 1), pp. 153-161, 2007.
[DEU 01] DEUTSCH J.E., LATONIO J., BURDCA G.C., BOIAN R., “Post-stroke rehabilitation with Rutgers ankle
system A case study”, Presence, vol. 10, pp. 416-430, 2001.
[DIE 88] DIENER H.C., HORAK F.B., NASHNER L.M., “Influence of stimulus parameters on human postural
responses”, J Neurophysiol, vol. 59, pp. 1888-1905, 1988.
[DUN 87] DUNCAN P.W., BADKE M.B., “Determinants of abnormal motor control”, in Stroke rehabilitation:
the recovery of motor control, Chicago, Year Book Medical Publishers, pp. 135–159, 1987.
[FUN 98] FUNG J., JOHNSTONE E., “Lost in space. Multi- axial and multi-dimensional surface perturbations
delivered by a novel motion base device”, Soc Neurosci, vol. 24 (1), p. 1158, 1998.
[FUN 99] FUNG J., MACPHERSON J.M., “Attributes of quiet stance in the chronic spinal cat”, J Neurphysiol,
vol. 82, pp. 3056-3065, 1999.
[FUN 03] FUNG J., BOONSINSUKH R., RAPAGNA M., “Postural response triggered by surface perturbations are
task specific and goal directed”, in Progress in motor control, M. Latash, M. Levin (ed.), vol. 3: Effects
of age, disorders and rehabilitation. Human Kinetics, pp. 169-181, 2003.
[FUN 05] FUNG J., HUGHEY L., “Postural responses triggered by multidirectional leg litfs and surface tilts”,
Exp Brain Res, vol. 165, pp. 152-166, 2005.
[GUR 00] GURALNIK J.M., FERRUCCI L., PIEPER C.F., “Lower extremity function and subsequent disability:
Consistency across studies, predictive models, and value of gait speed alone compared with the short
physical performance battery”, J Gerontol A Biol Sci Med Sci, vol. 55A (4), pp. M221-231, 2000.
[HEN 98a] HENNRY S., FUNG J., HORAK F.B., “Control of stance during lateral and anterior/posterior surface
translations”, IEEE Trans Rehab Eng, vol. 6 (1), pp. 32-34, 1998.
[HEN 98b] HENNRY S., FUNG J., HORAK F.B., “EMG responses to maintain stance during multidirectional
surface translations”, J Neurophysiol, vol. 80, pp. 1939-1950, 1998.
[HOL 99] HOLDEN M., TODOROV E., CALLABAN J., BIZZI E., “Virtual environment training improves motor
performance in patients with stroke”, Neurology Report, vol. 23, pp. 57-67, 1999.
[HOR 89] HORAK F., SHUPERT C., MIRKA A., “Components of postural dyscontrol in the elderly: a review”,
Neurobiol Aging, vol. 10, pp. 727-745, 1989.
[HOR 96] HORAK F.B., MACPHERSON J.M., “Postural orientation and equilibrium”, in Handbook of Physiology, section 12: Integration of Motor, Circulatory, Respiratory and Metabolic control during Exercises,
New York, FA Davis, pp. 22-46, 1996.
[ING 95] INGLIS T., MACPHERSON M., “Bilateral labrynthectomy in the cat: effects on postural response to
translation”, J Neurophysiol, vol. 73, pp. 1181-1191, 1995.
[MAS 92] MASSION J., “Movement, posture and equilibrium: Interaction and coordination”, Prog Neurobiol,
vol. 38 (1), pp. 35-56, 1992.
[JEK 06] JEKA J., ALLISON L., SAFFER M., ZHANG Y., KIEMEL T., “Sensory reweighting with translational
visual stimuli in young and elderly adults: the role of state-dependent noise”, Exp Brain Res, vol. 174,
pp. 517-527, 2006.
[KES 04a] KESHNER E.A., KENYON R.V., LANGSTON J., “Postural responses exhibit multisensory dependencies
with discordant visual and support surface motion”, J Vestib Res, vol. 14 (4), pp. 307-19, 2004.
[KES 04b] KESHNER E.A., KENYON R.V., “Using immersive technology for postural research and rehabilitation”, Assistive Technology, vol. 16 (1), pp. 54-62, 2004.
[LAM 03] LAMONTAGNE A., FUNG J., “Postural adjustments to voluntary head motions during standing are
modified following stroke”, Clin Biomech, vol. 18 (9), pp. 832-842, 2003.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf
75
[LUY 97] LUYAT M., OHLMAN T., “Subjective vertical and postural activity”, Acta Psychol, vol. 95 (2),
pp. 181-193, 1997.
[MAC 99] MACPHERSON J.M., FUNG J., “Weight support and balance during stance in the chronic spinal
cat”, J Neurophysiol, vol. 82, pp. 3066-3081, 1999.
[MAH 05] MAHBOOBIN A., LOUGHLIN P.J., REDFERN M.S., SPARTO P.J., “Sensory re-weighting in human postural
control during moving-scene perturbations”, Exp Brain Res, vol. 167 (2), pp. 260-267, 2005.
[MAK 96] MAKI B.E., McILROY W.E., “Postural control in the older adult”, Clin Geriatr Med., vol. 12,
pp. 635-658, 1996.
[MAN 89] MANCHESTER D., WOOLLACOTT M., ZEDERBAUER-HYLTON N., MARIN O., “Visual, vestibular and somatosensory contributions to balance control in the older adult”, J Gerontol, vol. 44, pp. M118-M127, 1989.
[MCC 02] MCCOMAS J., SVEISTRUP H., “Virtual reality application for prevention, disability awareness
and physical therapy rehabilitation in neurology: our recent work”, Neurol Report, vol. 26, pp. 55-61,
2002.
[MER 97] MERGNER T., HUBER W., BEKER W., “Vestibular-neck interaction and transformation of sensory
coordinates”, J Vestib Res, vol. 7 (4), pp. 347-67, 1997.
[MER 05] MERGNER T., SCHWEIGART G., MAURER C., BLUME A., “Human postural responses of real and virtual
visual environments under different support base conditions”, Exp Brain Res, vol. 167, pp. 535-556,
2005.
[NAS 78] NASHNER L., BERTHOZ A., “Visual contribution to rapid motor responses during postural control”,
Brain Res, vol. 150 (2), pp. 403-407, 1978.
[PRE 04] PREUSS R., FUNG J., “A simple model to estimate force plate inertial components in a moving
surface”, J Biomechanics, vol. 37, pp. 1177-1180, 2004.
[RUN 98] RUNGE C.F., SHUPERT C.L., HORAK F.B., ZAJAC F.E., “Role of vestibular information in initiation of
rapid postural responses”, Exp Brain Res, vol. 122, pp. 403-412, 1998.
[SPA 06] SPARTO P.J., REDFERN M.S., JASKO J.G., CASSELBRANT M.L., MANDEL E.M., FURMAN J.M., “The influence of dynamic visual cues for postural control in children aged 7-12 years”, Exp Brain Res, vol. 168
(4), pp. 505-16, 2006.
[STR 06] STREEPEY J.W., KENYON R.V., KESHNER E.A., “Field of view and base of support width influence
postural responses to visual stimuli during quiet stance”, Gait Posture, vol. 25 (1), pp. 49-55, 2006.
[SVE 04] SVEISTRUP H., “Motor rehabilitation using virtual reality”, J Neuroengineering Rehabil, vol. 1
(1), p.10, 2004.
[VAN 99] VAN
DER
KOOIJ H., JACOBS R., KOOPMAN B., GROOTENBOER H., “A multisensory integration model
of human stance control”, Biol Cybern, vol. 80 (5), pp. 299-308, 1999.
[VID 82] VIDAL P.P., BERTHOZ A., MILLANVOYE M., “Difference between eye closure and visual stabilization
in the control of posture in man”, Aviat Space Environ Med, vol. 53 (2), pp. 166-170, 1982.
[WHI 02] WHITNEY S.L., SPARTP P.J., BROWN K., FURMAN J.M., JACOBSON J.L., REDFERN M.S., “The potential use
of virtual reality in vestibular rehabilitation: Preliminary findings with the BNAVE”, Neurology Report,
vol. 26, pp. 72-78, 2002.
[WIN 91] WINSTEIN C.J., “Knowledge of the results and motor learning: implications for physical therapy”,
Physical Therapy, vol. 71, pp. 140-149, 1991.
[WIN 96] WINTER D.A., “Unified theory regarding A/P and M/L balance in quiet stance”, J Neurophysiol,
vol. 75, pp. 1-10, 1996.
Schedae, 2010, prépublication n° 5 (fascicule n° 1, p. 59 - 76).
http://www.unicaen.fr/services/puc/ecrire/preprints/preprint0052010.pdf