Control of equilibrium requires the integration of sensory information from visual, somatosensory, and vestibular sources. Although visual and somatosensory information have received considerable attention when investigating bal-ance and postural control during locomotor tasks, vestibular contributions are less understood because of the challenge of isolating vestibular inputs for manipulation.
Many investigations into vestibular involvement during locomotion have examined vestibular-deficient groups (12,15). The interpretation of results in this area is limited, however, by the ability to compensate for vestibular deficits through the weighting of other sensory systems. In this context, weighting refers to the contributions of a sensory system to movement or postural control relative to those from other sensory inputs. In the absence of vestibular inputs, for example, visual and somatosensory information are weighted more heavily to compensate for the lack of vestibular contribution. Weighting may also refer to the importance that the central nervous system puts on a particular sensory system in a given situation. Research has shown that visual information differs in contribution across the phases of a locomotor task (8). Weighting of sensory information in the context of an activity is referred to as upregulation or downregulation. Task- and phase-specific regulation of sensory information has been shown to play a large role in the execution of successful movement.
This brief review highlights the importance of vestibular information during dynamic tasks such as a forward step and during walking. Research has shown a severe deficit in movement capability in humans when vestibular information is distorted during platform perturbations (9), voluntary lateral tilt (14), and most recently during walking (2–4,7,10).
Successful locomotion requires appropriate motor commands for task progression and equilibrium control. During locomotion, the base of support (BoS) is continually in transition; therefore, control of the balance and progression of the center of mass (CoM) is largely maintained through placement of the feet. The weighting of vestibular information, and subsequently its importance in task completion can therefore be assessed through changes in foot placement during a galvanic vestibular stimulation (GVS) intervention.
Movement of the upper body in the coronal plane is believed to contribute to equilibrium control. Alignment of the body segments in a gravitational reference frame facilitates an internal assessment of the body within the environment (13) and promotes appropriate changes to placement of the lower limbs. The contribution of vestibular input to the regulation of vertical orientation can therefore be assessed through changes in the angular roll of upper body segments during a GVS perturbation.
Both upper body roll and foot placement have been the focus of our investigations into vestibular influences to maintain equilibrium while completing locomotor tasks. Vestibular weighting across different phases of a task and between the upper and lower body are discussed in relation to the potential roles that vestibular contributions have in the maintenance of equilibrium.
Vestibular Contributions during Movement: What Do We Know?
Vestibular deficient patients are capable of standing and even walking forward with their eyes closed; however, they exhibit compensatory alterations during walking in their step width, gait speed, and duration of double support (15). Peruch et al. (12) reported that patients within 1 month of a unilateral vestibular neurotomy were able to make accurate rotations to targets around a room by using other sensory cues. However, it appeared that patients never fully recovered a proper internal representation of their body in space, leading to deficits in their ability to perform complex navigational tasks. These sensory compensations for vestibular deficits, although ubiquitous, appear unable to replace a fully functional vestibular system, especially during skilled dynamic movements.
Recently we have used the technique of GVS to learn more about the specific roles of vestibular inputs during dynamic activities. GVS has a number of advantages over previously used techniques and presents a predictable and well-documented postural response. GVS involves the application of electrodes over the mastoid processes, where current is delivered to the 8th cranial nerve. The binaural, bipolar configuration of electrodes is known to cause an increase (cathode) or decrease (anode) in the firing of the peripheral vestibular afferents (9). GVS has the advantage of being a temporary perturbation, which can be adjusted to each individual’s threshold and used over select periods of time during a task (3). Therefore, unlike patient populations, GVS can provide controlled and reversible changes to information arising from the vestibular system over a single testing session. Work using GVS has established a predictable response to the stimulation in which subjects tilt toward the anode electrode when standing (14) and follow a curved trajectory toward the anode during locomotion (3,7,10).
GVS has been used to study vestibular contributions during tasks that involve acceleration of the body. These investigations have shown that GVS effects vary according to perturbation velocity, the voluntary movement speed, and the timing of the GVS relative to the movement phase (9,14). In particular, this work suggests that vestibular information likely plays a greater role in tasks in which the relationship between the CoM and BoS is dynamic. An increase in vestibular weighting during these relatively stationary tasks (i.e., stationary BoS) led us to predict large vestibular contributions during locomotor tasks where the equilibrium states are truly dynamic (changing BoS). The specific goal of our work has been to determine the nature of vestibular contributions for the execution of locomotor tasks.
Vestibular Contributions during the Transition into a Dynamic Task
We know that when vestibular information is perturbed, locomotor behavior is affected in proportion to the size of the perturbation (3,7). How this vestibular information influences locomotor behavior remains unclear, however. We began to examine this by looking at vestibular contributions during the initiation of a step from stationary standing (1,5). This task incorporates both stationary and dynamic components, allowing us to assess changes in the vestibular contribution during the transition between these two states.
To assess the weighting of vestibular inputs across all phases of the task, GVS was delivered before subjects began the step, and was continued for the duration of the trial.
All individuals demonstrated the predicted lean and bend of upper body segments toward the anode electrode during this stationary phase, as previously reported (14). Despite the ongoing vestibular stimulation, we saw no change in vestibular influence as subjects moved from stationary standing into step initiation (from onset of mediolateral center of pressure (CoP) displacement to toe-off [TO] of the swing limb). Subjects began exactly the same planned step and ended up placing the foot in the same place. Why were no differences observed in this early phase? It appears that subjects used vestibular information to align body segments during the stationary phase in preparation for step initiation. The feed-forward nature of step initiation (8) then precluded further involvement of vestibular input until a later, more dynamic, phase of the step (i.e., from first heel contact [HC] to the stable end position) when vestibular inputs were upregulated to control body movement and dynamic equilibrium, through upper body roll. Significant changes only appeared at the second TO, very late in the step task (Fig. 1A). Increased vestibular weighting was also demonstrated by CoP changes during the termination of the step when it was necessary to establish a stable end position. Thus, these data indicate that vestibular inputs are regulated differently across a forward step task, with no changes during the initiation of the step, followed by increased vestibular contributions late in the step after the transition into the more dynamic phase.
Differences became evident between the upper and lower body responses during the step task, especially when comparing trials with and without vision. The relative weighting of these sensory sources changed across the step, with vision providing the dominant source of input during the dynamic execution phase. During the stationary phase, before the step, vision did not attenuate the upper body roll despite a reduction in the mediolateral CoP deviation. This suggests a dominant vestibular role in regulating alignment of the body segments in standing.
Changes in upper body roll were also seen during gait initiation during the transition into a dynamic state (4). When we delivered GVS with the onset coinciding with specific events (anticipatory postural adjustment (APA), TO, and HC) across gait initiation, we saw a monotonic increase in body roll in line with the increasing dynamic nature of the task (Fig. 1B). We confirmed that these changes were related to the increasing dynamic component when no differences were found in the degree of roll between the three conditions during steady-state gait, long after the transition into the dynamic task was complete (2).
Although vestibular information is upregulated during transitional dynamic tasks, such as during a forward step, or in the transition from standing to walking, there does not appear to be a further increase in vestibular weighting during running (10). Jahn et al. (10) believe that vestibulospinal influences are reduced during running so as not to adversely affect the highly automated running process. Unpublished observations in our laboratory support these findings of no significant deviation toward the anode electrode during running. Interestingly, however, when we asked subjects to stop, we observed substantial corrective movements toward the anode, despite a lack of response during the running task. These observations support the idea that there is an enhanced vestibular influence in the transition between states, particularly the transition into a stable end position, similar to that observed during step termination.
Our step initiation studies clearly show that there are differences in the way vestibular information is used and regulated across different phases of a step initiation task. The next question we asked was whether there is evidence for phase-dependent modulation of vestibular inputs during walking when there are repeated transitions between stable positions from one step to the next.
Phase-Dependent Weighting of Vestibular Information
Investigations into phase-dependent modulation of vestibular inputs were tested by probing vestibular contributions at distinct times during gait initiation (APA, TO, HC) and steady-state gait (HC, midstance, TO) using GVS. The objective was to determine whether use of vestibular information was differentially modulated across separate phases of the task. This was indeed shown to be the case.
During gait initiation (4) and during steady-state gait (2), the largest and earliest response in foot placement was observed after stimulation at HC (Fig. 2). We concluded from this that for lower limb control, vestibular information is weighted more heavily during double support than at any other time of the gait cycle. This may be a manifestation of the basic role of vestibular projections in antigravity activity (11), but two other lines of evidence suggest other reasons why this particular time of the gait cycle is opportune for vestibular upregulation. First, HC has been indicated as a critical time for planning of both the current and subsequent steps during locomotion (8). Hollands and Marple-Horvat (8) demonstrated that the programming for planned stepping to a visual target occurs in the last 100 ms of the stance, or double support, phase of gait. As a result, the program for the next foot placement is complete by the time the foot leaves the ground at TO. We propose that the double support period would therefore present an opportune time for vestibular information to contribute to changes in the planned foot trajectory. Second, there will be an increase in the available meaningful somatosensory information during double support because both feet are on the ground. This information needs to be integrated with vestibular information at this time to generate an accurate internal representation of the body in space (13). The integration of these sensory sources would indicate whether a planned movement of the body relative to its base of support would result in the desired end position during both perturbed and unperturbed walking. Alterations could then be made to foot placement to prevent potential disequilibrium.
Tucker et al. (15) demonstrated that when patients with bilateral vestibular hypofunction are asked to walk at a self-selected pace, they chose to walk significantly slower than the control subjects in the study. They propose that the slower gait enables them to take advantage of spinal reflex loops that may help to facilitate or instigate postural responses to external perturbations. Research has shown an activation of spinal motoneuron pools by descending vestibular inputs (11). Such vestibular projections may help to set a level of excitability in healthy individuals during standing and at specific critical times during the gait cycle (11). Evidence of phase-dependent vestibular regulation for lower limb control may support a proactive role for vestibular input in planning the next step, as well as being a contributor to foot placement changes to prevent a fall in the event of disequilibrium. Reduced or absent vestibular inputs may therefore compromise the ability to use feed-forward strategies during locomotion.
Despite evidence of phase-dependent vestibular modulation for lower limb control during locomotion, changes in the magnitude of upper body roll did not appear to be influenced by when a perturbation arrived during the gait cycle (2). A similar upper versus lower body dissociation has been shown during standing (5). We considered next the possibility of separate roles for vestibular inputs in controlling the upper and lower body.
Upper versus Lower Body Control: Implications for the Roles of Vestibular Inputs
When we examined visual–vestibular interactions, particularly across a voluntary forward step, it appeared to us that there were possibly independent vestibular controls on the upper versus the lower body (5). In the stationary phase before the onset of the step, the presence of vision did not attenuate the vestibular-related upper body roll toward the anode electrode. In contrast, the weighting of visual inputs was greater than vestibular contributions for the lower body, and this resulted in a significantly reduced CoP shift that was related to lower limb control (Fig. 3A). We propose that vestibular information during quiet stance serves at least two separate functions. The shift in CoP corresponds in direction and temporal onset to the GVS-evoked medium latency response that is reported to have a postural function (6). As a result, this lower body response is likely to represent a vestibular contribution to postural stability. The second role we propose is the use of vestibular information to align or stack the segments of the body to provide a balanced, upright position from which to initiate movement. These findings agree with recent reports of distinct otolith and canal-related short and medium latency postural responses, respectively (6). The upper body roll response may relate to the gravitational otolith information, which cannot take advantage of visual inputs to facilitate the stacking response. Alternatively, the dynamic response of the lower limbs occurs at the time of the medium latency postural response, which Cathers et al. (6) suggest is related to the canal inputs and is susceptible to visual modulation.
The division of vestibular input for upper and lower body control is supported by evidence from our more recent studies. GVS delivered at specific events over the initiation of gait (4) and during steady-state locomotion (2) demonstrated a distinct separation in the vestibular regulation of upper body roll and foot placement. Foot placement in both studies was shown to support phase-dependent modulation of vestibular inputs whereas upper body roll did not follow a phase-dependent pattern of modulation (Fig. 3B). During the initiation of gait, roll of the upper body increased in a continuous fashion across the three events tested, when GVS was delivered at either APA, TO, or HC. If the weighting of vestibular information was phase-dependent and therefore related to the presence and weighting of other sensory information, we would expect similar upper body roll during the APA and HC trials, because somatosensory contributions from the feet, with both on the ground, are similar. However, the roll magnitude was smallest during APA and largest in the HC trials, indicating a modulation based on the transition into a dynamic state rather than a phase-dependent modulation that has been demonstrated for other sensory sources in various locomotor tasks (8). Upper body roll during steady-state locomotion did not differ between the three conditions, suggesting that vestibular weighting was homogenous, and therefore important to the task, across all the events tested.
Why would it be important to upregulate vestibular inputs for upper body control equally across all phase of the gait cycle? Head stabilization during locomotion and during complex balance activities requires strong attenuation of head movement in space (13). Two functional benefits of this head stabilization are apparent. First, the head houses both the visual and vestibular systems. Head stabilization ensures that the sensitivity of these two systems to change is optimized. Control of the head also facilitates the transition from the head-centered reference frame to the exocentric reference frame for postural control (13). The role of vestibular input for head alignment is therefore speculated to have an important contribution to equilibrium control.
How Does Vestibular Regulation of the Lower Limbs Contribute to the Locomotor Task?
Given the increased weighting of vestibular information for the control of limb movement in gait initiation and steady-state gait, what can we say about the role of vestibular information during these two different tasks? Mediolateral foot placement is a primary means of regulating balance in the frontal plane during stepping. This is because control of the CoM movement is largely influenced by the position of the CoP. The different placement of the feet in our gait initiation study (4) compared with our steady-state gait study (2) suggests that each task could use vestibular information differently to control the CoM. During gait initiation, the onset of GVS at individual events (APA, TO, HC) resulted in different magnitudes of foot placement by step one. By step three, there were no significant differences between the three conditions (Fig. 2A). In contrast, changes to foot placement during steady-state gait (GVS delivered at HC, midstance, or TO), did not become significantly different until step two and remained different over the period tested (Fig. 2B). Increased vestibular weighting during the transition into the dynamic gait-initiation task is important for establishing whether the body progression is in the appropriate direction and with the appropriate speed to enable a safe and accurate initial step. We speculate that such an assessment requires vestibular input to be integrated with somatosensory information (13). The different vestibular weighting at the three events would indicate different degrees of disequilibrium and would therefore require different magnitudes of foot placement compensation. Once a forward trajectory is attained and stability is reestablished, we see a similar direction of progression between the three conditions.
During walking in both healthy (7,10) and vestibular-deficient populations (12,15), vestibular inputs have been implicated in path navigation. Fitzpatrick et al. (7) suggest that when performing a planned walking trajectory, we compare incoming vestibular information with the expected input for correct completion of the task. Our studies show that there is a large upregulation of vestibular information at heel contact compared with the midstance and TO phases. We therefore propose that this comparison and generation of an error signal occurs intermittently during walking at the time of heel contact.
In our studies presented here, we have sought to examine the nature of vestibular contributions to specific tasks of locomotion.
Although the importance of vestibular information is often overlooked in the context of other sensory information, existing research shows that vestibular information contributes strongly to maintaining balance, especially during dynamic tasks. We have shown that the magnitude of the response to a vestibular perturbation increases during the transition into a dynamic task, is graded with the degree of vestibular perturbation, and is modulated in a phase-dependent manner (Fig. 4). Our work also provides strong evidence that vestibular control of the upper body and vestibular control of the lower body are separate entities during such dynamic activities.
Understanding how and when vestibular information controls balance has important implications for the development of appropriate and effective rehabilitation programs. In particular, future rehabilitation strategies geared toward individuals with vestibulopathy should focus on compensating for deficient vestibular inputs at critical points in the gait cycle. Vestibular information features prominently during quiet standing for upper body control. Although significantly modulated by vision during step execution and locomotion, it is also important to note the substantial role of vestibular information in whole-body equilibrium, through lower limb compensations. Incorporating these different vestibular roles during rehabilitation using locomotor activities could be valuable to retrain individuals to achieve dynamic goals.
To conclude, we have shown that vestibular sensory information is not only used to correct balance and monitor postural corrections after the body has become unstable but also has an important contribution to feed-forward strategies that prevent unstable situations from developing. The phase-dependent modulation of vestibular information suggests that the planning for locomotor progression uses vestibular information to ensure appropriate foot placement, direct forward movement, and prevent disruptions to balance. Separate control of the upper and lower body further demonstrates the complexity and the integral role for vestibular contributions in regulating successful locomotion.
1. Bent, L.R., J.T. Inglis, and B.J. McFadyen. Vestibular Contributions across the execution of a voluntary forward step. Exp. Brain Res
. 143:100–105, 2002.
2. Bent, L.R., J.T. Inglis, and B.J. McFadyen. When is vestibular information important during walking
? J. Neurophysiol
. 92:1269–1275, 2004.
3. Bent, L.R., B.J. McFadyen, V.F. Merkley, P.M. Kennedy, and J.T Inglis. Magnitude effects of galvanic vestibular stimulation
on the trajectory of human walking
. Neurosci. Lett
. 279:157–160, 2000.
4. Bent, L.R., B.J. McFadyen, and J.T. Inglis. Is the use of vestibular information weighted differently across the initiation of walking
? Exp. Brain Res
. 157:407–416, 2004.
5. Bent, L.R., B.J. McFadyen, and J.T. Inglis. Visual-vestibular interactions in postural control
during the execution of a dynamic task. Exp. Brain Res
. 146:490500, 2002.
6. Cathers, I., B.L. Day, and R.C. Fitzpatrick. Otolith and canal reflexes in human standing. J. Physiol
. 563(1):229–234, 2004.
7. Fitzpatrick, R.C, D.L. Wardman, and J.L. Taylor. Effects of galvanic vestibular stimulation
during human walking
. J. Physiol. (Lond.)
8. Hollands, M.A., and D.E. Marple-Horvat. Visually guided stepping under conditions of step cycle-related denial of visual information. Exp Brain Res
9. Inglis, J.T., C.L. Shupert, F. Hlavacka, and F.B. Horak. Effect of galvanic vestibular stimulation
on human postural responses during support surface translations. J. Neurophysiol
. 73:896–901, 1995.
10. Jahn, K., M.S. Strupp, E. Schneider, M. Dieterich, and T. Brandt. Differential effects of vestibular stimulation on walking
and running. Neuroreport
11. Matsuyama, K., and T. Drew. Vestibulospinal and reticulospinal neuronal activity during locomotion in the intact cat. I. Walking
on a level surface. J. Neurophysiol
. 84:2237–2256, 2000.
12. Peruch, P., L. Borel, F. Gaunet, G. Thinus-Blanc, J. Magnan, and M. Lacour. Spatial performance of unilateral vestibular defective patients in non-visual versus visual navigation
. J. Vestib Res
. 9:37–47, 1999.
13. Pozzo, T., Y. Levik, and A. Berthoz. Head and trunk movements in the frontal plane during complex dynamic equilibrium tasks in humans
. Exp. Brain Res
. 106:327–338, 1995.
14. Séverac Cauquil, A., and B.L. Day. Galvanic vestibular stimulation
modulates voluntary movement of the human upper body. J. Physiol
. 513:611–619, 1998.
15. Tucker, C.A., J. Ramirez, D.E. Krebs, and P.O. Riley. Center of gravity dynamic stability in normal and vestibulopathic gait. Gait Posture