Atlanta Veterans Administration, Rehabilitation Research and Development, Decatur, GA; 2Department of Rehabilitation Medicine, Emory University, and 3Department of Otolaryngology-Head and Neck Surgery, Emory University, Atlanta, GA
Upright posture is inherently unstable in human beings: a heavy upper body must be balanced over a smaller lower body. The maintenance of upright balance requires that the center of mass be positioned within the base of support, either of which may be moving. Furthermore, in order to meet the demands of a constantly changing environment, body position and movement must be continuously monitored and updated with information received from the visual, somatosensory and vestibular systems. This multimodal sensory information is integrated within the central nervous system and, based on the perception of current demands on postural stability, an appropriate motor response is generated. The motor response must be accurately timed and scaled in order to prevent a fall. Failure in any one of the sensory or motor systems results in impaired ability to control posture and may result in a fall. The effect of sensory or motor system loss on maintaining balance varies with the degree of challenge to stability. For example, the balance challenge to an individual is very different when standing still compared to standing on a bus that suddenly lurches. This chapter will focus on:
1 vestibular contributions to postural stability;
2 the effect of peripheral vestibular loss on balance and postural control;
*Supported in part by Research Career Development Award C3249V awarded by the Veterans Administration.
3 the effect of eye movements on balance;
4 the role of vestibular rehabilitation (VR) in the remediation of imbalance and gaze instability.
In order to study the unique contribution of the visual and somatosensory systems during quiet stance, researchers systematically reduce or alter the input of each. Using current technology, it is a relatively trivial problem to remove or alter visual or somatosensory information used for postural control. Briefly, postural sway can be measured under conditions in which somatosensory and visual feedback is altered (see Chapter 8 of Volume II for details). As a person stands on a force platform, changes in vertical pressure produced by body sway is used to move either the support surface or visual surround in synchrony with the individual's sway (referred to as "sway-referenced"; Baloh et al., 1998). Movement of the support surface in parallel with the individual alters somatosensory cues normally used for postural stability. For example, in quiet stance we normally have a small amount of anterior and posterior (AP) sway, which results in changes in ankle angle. If the support surface moves with body sway, the change in the ankle angle is minimized. This alters the somatosensory feedback, rendering it less effective as a signal in the maintenance of upright posture. Similarly, if the visual surround is moved in parallel with postural sway, the visual feedback cues are novel and cannot be used as effectively to maintain postural stability. In reality, movement of the support surface and of the visual surround in parallel with body sway occurs only at low frequencies of sway (<0.3 Hz) because of the mechanical constraints of the equipment. At higher frequencies of body sway, support surface and visual surround movement is out of sink with the body. The result, however, is still novel sensory feedback that cannot be used as effectively for postural stability. The test is organized into a series of six conditions of increasing difficulty. The first three conditions are performed on a firm surface with eyes open, eyes closed and finally with vision sway-referenced. The final three conditions are performed with the support surface sway-referenced with eyes open, eyes closed and with vision sway-referenced.
Manipulating vestibular inputs cannot be accomplished so easily because we are bound by the earth's gravitational field (Nashner, 1982). Thus, most of our knowledge about the role of the vestibular system in postural control is derived from studies of individuals with loss of vestibular function. Interpretation of information garnered from these studies is confounded by the fact that the findings reflect both the loss of vestibular function and the compensation for that loss.
Additionally, until recently researchers and clinicians have judged vestibular loss by measuring the function of the horizontal semicircular canals. We could not measure the function of the vertical semicircular canals, the saccule and the utricle until the development of methods such as off-axis rotational testing, vestibular evoked myogenic responses and measurement of subjective visual vertical (Colebatch and Halmagyi, 1992; Halmagyi and Colebatch, 1995; Bohmer and Mast, 1999; Li et al., 1999; Shepard and Howarth, 1999; Vibert et al., 1999; Welgampola and Colebatch, 2001). Thus studies of people with bilateral vestibular loss may be muddled by remaining otolith or even posterior canal function in some but not all subjects. These caveats should be kept in mind as we explore the contribution of vestibular system to postural stability.
Another method utilized to study the role of vestibular inputs is to directly stimulate the vestibular system. This has been accomplished via small, direct electrical current applied to the mastoid processes (galvanic vestibular stimulation, GVS; Nashner and Wolfson, 1974) or by direct head displacement (Horak et al., 1994). GVS has been shown to directly affect posture by increasing activation of distal ankle musculature, thereby increasing postural sway and ankle torque (Nashner and Wolfson, 1974; Magnusson et al., 1990). Finally, postural reactions have been examined using different techniques to perturb balance. The perturbations include translational and pitch perturbations of the support surface, sinusoidal rotation of the support surface at different frequencies and sudden movement or flow of the visual world. These methods both perturb balance and manipulate sensory cues.
20.3 Overview of sensory contributions to postural control
The three systems (visual, somatosensory and vestibular) that provide the main sensory inputs for postural control each contribute unique information regarding body posture and motion. This information is used to generate automatic postural responses and also contributes to voluntary postural control. No single sensory system, however, provides us with sufficient information to definitively determine body position and movement.
Visual information is used to orient the body relative to the visual world and provides information used to determine whether the person is moving or the visual world is moving. There are several different visual cues used for postural stability. One long-held concept has been that, as a person sways, even during quiet standing, retinal slip information is used to determine body movement relative to environmental movement. This concept has been challenged by the recent studies of Jahn et al. (2002) who suggest that efference copy of ocular motor signals, rather than retinal slip, is used to regulate postural sway. Other visual cues include changes in image size and retinal disparity, which would occur with fore-to-aft sway. Visual stabilization of balance appears to be primarily dependent on central vision and is related to the distance from the eyes to the visual target (Brandt et al., 1973; Paulus et al., 1984; Brandt et al., 1985). As distance to a visual target increases, relative changes in visually mediated cues used for balance (eye movement, retinal slip and retinal disparity) decrease and become less effective in maintaining postural stability. It is only when the distance decreases from 100 to 10 cm that postural stability improves. Visual signals by themselves provide ambiguous information about postural stability. We have all had the experience of quickly putting our foot on the brake when a car pulls up next to us at a stoplight. Input from either the somatosensory or vestibular system is required to resolve this sensory conflict and determine the correct locus of the movement.
The somatosensory system provides information concerning body position and movement relative to the support surface. In addition, the somatosensory system provides information regarding the position and movement of body segments relative to each other. Signals from skin, muscle and joint receptors affect postural responses in several ways. Shifts in alignment alter which pressure receptors in the feet are firing; joint receptors may send signals about the absolute position of the joint, about velocity of joint movement or about the direction of joint movement. Even in quiet stance, we normally have a small amount of AP sway, which results in changes in ankle angle and in foot pressure. This alteration in somatosensory feedback is an effective signal in the maintenance of upright posture. One of the more important roles of somatosensory cues in postural stability appears to be the prevention of falls. The majority of falls occur as a result of an external perturbation, such as a slip or trip. In order to prevent a loss of balance or fall, the automatic postural responses must be appropriately timed and scaled. Through use of sudden translations or rotations of the surface upon which the subject stands, researchers simulate the conditions under which a fall might occur. Through these studies, we have gained considerable insight into the role of the sensory cues in automatic postural responses. The weight of evidence identifies the role of somatosensory input as triggering and selecting postural responses to platform perturbations (Keshner et al., 1987; Horak et al., 1990; Allum and Honegger, 1995).
Vestibular input is used to determine head position relative to gravity and to provide information regarding linear and angular head acceleration to detect self-motion. The vestibular system can detect even the small head movements resulting from body sway during quiet stance. Motor responses to head acceleration are then mediated through the vestibu-lospinal system. Direct stimulation of the vestibular system (via GVS) results in the perception of movement with a resultant increase in leg muscle activation and body sway (Nashner and Wolfson, 1974; Magnusson et al., 1990). Direct stimulation of the vestibular system (via head displacement) results in activation of the same trunk and leg muscles as seen following surface translation (Horak et al., 1994). For example, forward displacement of either the head or body results in activation of the gastrocnemius, hamstrings and paraspinals to maintain upright posture. The amplitude of the muscle response to direct vestibular stimulation is very small (approximately one-third) compared to the response following surface translation initiated through somatosensory input. Therefore, while the vestibular system can trigger appropriate balance responses in the trunk and legs, the primary role of vestibular inputs appears to be stabilization of the head in space.
Vestibular inputs become more critical when visual or somatosensory inputs are absent, reduced or under more challenging balance conditions. When the surface is firm, somatosensory inputs from trunk and legs dominate. When surface is altered (narrow or compliant), input from vision and vestibular dominates. When somatosensory information is reduced (e.g., with sway-referenced surface) the amplitude of muscle response following head displacement is increased (Horak et al., 1994). When the surface is unstable so that the upper extremity is necessary to maintain upright balance, direct stimulation of the vestibular system (via GVS) results in activation of the upper extremity musculature (Britton et al., 1993).
Simultaneous translation of the support surface and GVS results in exaggerated responses (Inglis et al., 1995; Horak and Hlavacka, 2002). Final postural alignment is significantly altered such that subjects have much greater trunk leans than can be accounted for by either stimulus alone or in combination. It may be that the automatic postural responses were triggered by somatosensory input from the platform perturbation and that the vestibular system contributed the internal representation of vertical used to adjust final postural alignment, in this case, increased lean. Ultimately, input from the lower extremity (somatosensory) is integrated with information from the head (vision and vestibular) to provide an accurate internal representation of body position and movement (Mergner and Rosemeier, 1998).
Redundancy: not substitution
There is a certain amount of redundancy in the contributions of the different sensory systems to postural stability. Although this is important to remember, it is equally important to note that each sensory signal appears to have optimal conditions and frequencies over which it works (Dichgans et al., 1976; Berthoz et al., 1979; Diener et al., 1982; Diener et al., 1984; Tokita et al., 1984). For example, when healthy subjects are standing in a normal environment (flat, stable surface), the somatosensory and visual systems have more influence on posture than does the vestibular system (Nashner, 1982). However, as will be discussed later, vestibular information becomes more important in stabilizing head and trunk motion in space under more challenging conditions such as standing on a moving surface or walking and running (Buchanan and Horak, 2001-2002; Creath et al., 2002). Furthermore, Xerri et al. (1988) demonstrated that the control of posture at lower frequencies (<0.25 Hz) is dominated by visual input while at higher frequencies, vestibular inputs (from otololiths) dominate. The end result is that the different sensory systems cannot fully substitute for each other when there is a loss of function. Healthy subjects can maintain balance without vision or somatosensory input, however postural sway increases when either input is removed (Lee and Lishman, 1975; Diener et al., 1984). When somatosensory and visual information are present but novel, the vestibular system still provides appropriate inputs to resolve the sensory conflict and maintain upright posture but again there is an increase in sway (Nashner, 1982). Conversely, visual and somatosensory cues would only partially compensate for lost vestibular function.
In spite of this, studies show that visual cues help maintain balance in patients with bilateral vestibular hypofunction (BVH; Bles et al., 1983, 1984). Bles et al. (1984) have shown that during the course of recovery following BVH, patients change how they rely on sensory cues. Initially, they rely on visual cues as a substitute for the lost vestibular cues but over a 2-year period they increase their reliance on somatosensory cues to maintain balance. The use of proprioceptive cues would not fully compensate for vestibular function, however. When postural stability is perturbed by viewing a sinusoidal lateral tilt of the visual surround, body sway recovered to within normal limits at lower frequencies but not at higher frequencies (Bles et al., 1983). Thus neither visual nor somatosensory cues fully substitutes for the range of frequencies over which the vestibular system works.
20.4 What vestibular hypofunction tells us about the role of the vestibular system in postural stability
Ankle musculature responses to support surface perturbation
It is interesting to note that bilateral vestibular loss has no effect on the latency, sequence and timing of ankle musculature in response to perturbation of the support surface. That is, individuals with BVH respond with normal muscle onset latencies of the ankle musculature, muscle activation patterns, rate of change of force applied to the surface and symmetry of response between legs (Allum et al., 1988, 1994; Herdman et al., 1994). However, vestibular inputs do appear to modulate the amplitude of muscle activation. Individuals with BVH exhibit reduced stabilization at the ankle as measured by decreased amplitude of ankle muscle activity (to 30% of normal) and a concomitant reduction in ankle torque (Keshner et al., 1987; Allum et al., 1988, 1994). An additional problem that occurs with loss of vestibular function is that ankle muscle torque develops more slowly in patients with bilateral vestibular loss than in healthy subjects. Keshner et al. (1987) found that for random pitch perturbations of the support surface, the slope of ankle muscle torque (i.e., rate of change) was less steep for BVH than healthy controls. This suggests that patients with BVH are at risk of developing too little torque too slowly, which may result in an increased risk for falling.
The effect of vestibular loss on the amplitude of muscle response at the ankle is related to several factors, the most important being the degree of deficit. For example, subjects with unilateral vestibular loss have similar muscle onset latencies and amplitudes to translational perturbations as do normal controls (Nashner et al., 1982; Black et al., 1983). However, if you look at patients with acute unilateral vestibular hypofunction (UVH) separate from those with chronic UVH, there is a gradation of muscle response amplitude and resultant ankle muscle torque developed to pitch rotations of the support surface (Allum et al., 1988). Individuals with chronic UVH have the largest amplitude of muscle response, followed by those with acute UVH and individuals with BVH had the smallest amplitude of responses.
Trunk musculature responses to support surface perturbations
While responses in lower extremity muscles are reduced in BVH subjects, the responses of trunk muscles, particularly paraspinal muscles, are enhanced (Allum et al., 1994; Allum and Honegger, 1995). Thus, subjects with BVH exhibited increased trunk angular velocity during balance recovery. This increase trunk velocity is evidence of greater instability. As individuals with BVH do not generate adequate ankle torque following a balance perturbation, they may have difficulty maintaining upright posture following an external perturbation.
Balance strategies following support surface perturbations
Based on early work, Horak et al. (1990) proposed that individuals with BVH are unable to perform a hip strategy. Individuals with BVH exhibited normal balance responses except when standing on a narrow beam (Black et al., 1988; Horak et al., 1990). Individuals with chronic BVH did not alter muscle activation or shear forces under shortened surface condition. Controls, on the other hand, significantly increased early abdominal and quadricep activation on the shortened surface as well as shear forces (indicative of a hip strategy). More recent work by Runge et al. (1998) demonstrated that individuals with chronic BVH can use a hip strategy (defined by early hip torques) in response to larger perturbations on a normal support surface. These results highlight the importance of vestibular information in balance responses in the presence of reduced somatosensory inputs.
Modulation of response to support surface perturbations
Healthy individuals quickly learn to maintain their balance and are able to decrease the amplitude of response to repeated, identical perturbations of the support surface. Loss of balance is unusual except during the initial trial in younger subjects (Wolfson et al., 1992). Several studies have suggested that the vestibular system is not part of this modulation of motor responses. Herdman et al. (1994) demonstrated that subjects with BVH modulate the response to a series of identical pitch perturbations of the support surface as well as do healthy subjects. Although more subjects with BVH had an inappropriate response on the initial trial than did healthy subjects of the same age, the patients with BVH quickly reduced the amplitude of response on subsequent trials as do healthy subjects. Keshner et al. (1987)
studied the role of the vestibular and visual systems in adaptation. They compared muscle amplitude of trials 1-3 with trials 4-10 and found that both healthy controls and BVH exhibited a decrease (or adaptation)
in response amplitude of ankle and neck muscles in the second set of trials with and without vision. Thus, adaptation is not dependent on either an intact vestibular or visual system.
Was this article helpful?