Postural compensation for unilateral vestibular loss

Abstract

Postural control of upright stance was investigated in well-compensated, unilateral vestibular loss (UVL) subjects compared to age-matched control subjects. The goal was to determine how sensory weighting for postural control in UVL subjects differed from control subjects, and how sensory weighting related to UVL subjects' functional compensation, as assessed by standardized balance and dizziness questionnaires. Postural control mechanisms were identified using a model-based interpretation of medial-lateral center-of-mass body-sway evoked by support-surface rotational stimuli during eyes-closed stance. The surface-tilt stimuli consisted of continuous pseudorandom rotations presented at four different amplitudes. Parameters of a feedback control model were obtained that accounted for each subject's sway response to the surface-tilt stimuli. Sensory weighting factors quantified the relative contributions to stance control of vestibular sensory information, signaling body-sway relative to earth-vertical, and proprioceptive information, signaling body-sway relative to the surface. Results showed that UVL subjects made significantly greater use of proprioceptive, and therefore less use of vestibular, orientation information on all tests. There was relatively little overlap in the distributions of sensory weights measured in UVL and control subjects, although UVL subjects varied widely in the amount they could use their remaining vestibular function. Increased reliance on proprioceptive information by UVL subjects was associated with their balance being more disturbed by the surface-tilt perturbations than control subjects, thus indicating a deficiency of balance control even in well-compensated UVL subjects. Furthermore, there was some tendency for UVL subjects who were less able to utilize remaining vestibular information to also indicate worse functional compensation on questionnaires.

Keywords: balance; compensation; posture; unilateral; vestibular.

Figure 1

Feedback control model of sensory control of balance and posture adopted from Cenciarini and Peterka, . Stick figure shows our measures of lateral body-sway (BS) and lateral surface rotation (SS) and body movement relative to feet (BF) in addition to the internal representations of BS and BF (bs, bf) derived from vestibular and proprioceptive sensory systems. The model diagram illustrates the assumed feedback control structure whereby the corrective torque (T_c) applied to control body orientation is determined primarily by a weighted combination of vestibular and proprioceptive orientation signals (weights W_V and W_P, respectively) with additional feedback from sensory systems that detect the corrective torque applied to the body. Visual feedback is not included since experiments were performed with eyes-closed. Detailed descriptions of model components are provided in Table 2.

Figure 2

Medial–lateral (ML) body-sway evoked by support surface rotations. (A) Time course of one cycle of pseudorandom ML support surface rotation angle at four different amplitudes (left column) and mean ML body-sway angle are shown for a control subject (middle column) and a unilateral vestibular loss subject (UVL, right column). Body-sway response means with 95% confidence intervals (shaded gray) are shown. (B) Root mean square (RMS) values of ML body-sway are plotted as a function of peak-to-peak amplitude of the pseudorandom surface-tilt stimulus. The gray shaded region indicates the range where stimulus RMS values are less than the ML sway RMS values.

Figure 3

Group mean frequency response functions (FRFs expressed as gain and phase functions) and coherence functions of medial–lateral body-sway responses to four amplitudes of pseudorandom surface-tilt stimuli for control (left column) and unilateral vestibular loss subjects (right column).

Figure 4

Vestibular (W_V) and proprioceptive (W_P) weights vary as a function of support surface stimulus amplitude for control subjects and unilateral vestibular loss subjects. (A) Mean W_V and W_P values for control subjects. (B) Mean W_V and W_P values for unilateral vestibular loss subjects. (C) Comparison of W_V for controls and unilateral vestibular loss subjects. Mean W_V values in unilateral vestibular loss subjects were always less than in control subjects and never exceeded 0.5 (50% reliance on vestibular information). Means ± SE are plotted (N = 11 subjects in each group).

Figure 5

Vestibular weights (W_V) across stimulus amplitudes for individual control subjects (A) and individual unilateral vestibular loss subjects (B). Numbers labeling W_V data of unilateral loss subjects correspond to labels in Table 1.

Figure 6

Relationship between vestibular weights (W_V) and functional status assessed using the vestibular disorders activities of daily living (VDADL) scale for subjects with unilateral vestibular loss. (A) Linear regression and linear correlation coefficients relating VDADL score and W_V are shown for W_V values obtained from sway responses to support surface stimuli with 4° and 8° amplitudes. (B) Segregation of W_V data into two sets (N = 5 in each group) according to VDADL scores demonstrates that unilateral loss subjects with better function (lower VDADL) make greater use of vestibular information for stance control at larger stimulus amplitudes.