Patients with stable unilateral vestibular deficits are a difficult group in which to achieve satisfactory outcomes with common treatment modalities. Many individuals with acute vestibular pathology experience spontaneous amelioration of their symptoms without treatment[39–42]; however, despite some degree of spontaneous compensation in the weeks and up to three months after the onset of the deficit, many still exhibit symptoms of imbalance, adopt self-imposed limits on head movement, and are likely to struggle with difficult or dynamic balance tasks.[43]
There is strong evidence that some form of vestibular rehabilitation is better than none.[11,44] However, many patients do not achieve satisfactory outcomes from home-based or supervised exercises.[16,39,40] There is growing evidence that supporting rehabilitation with technologies that provide feedback on postural sway can improve balance and reduce postural sway[45]; however, in many cases, improvements in stability revert when the feedback support is removed.[10] Another technological adjunct for promoting vestibular compensation is CDP. There has been limited study of CDP-based interventions.[46–48] To our knowledge, changes in objective posturography associated with computerized vestibular retraining have not been reported for individuals with unilateral vestibular deficits.
For this study, we enrolled participants whose symptoms were persistent for greater than six months. This was to avoid confounding by spontaneous resolution in the acute phase. We also excluded patients with pathologies associated with a high rate of spontaneous resolution or that are associated with variable symptoms. Because our eligibility criteria required participants to be symptomatic at the time of enrolment, any naturally variable condition would have a high likelihood of experiencing transient amelioration of symptoms during follow up. By enrolling only those with stable symptoms lasting longer than six months, we hope to measure improvement associated with retraining rather than spontaneous resolution or cyclical variability. The majority (9 of 13) of these individuals had received previous vestibular physiotherapy without satisfactory outcomes.
The ability to maintain equilibrium when provided with conflicting sensory cues requires weighting of reliable cues over unreliable ones. Among those with a vestibular deficit, vision plays an important role in adaptation[10]; however, a study of patients recovering from unilateral vestibular neurectomy found some individuals rely more heavily on vision and others on somatosensory inputs.[9]
The individual conditions of the SOT challenge the participant to maintain equilibrium with a full complement of somatosensory, visual, and vestibular information and then systematically removes or creates sensory conflict with the somatosensory and visual information. Ratios of these scores indicate preference for one sensory input over another.
When the platform remained fixed (SOT conditions 1 to 3), allowing for reliable somatosensory information, median values were not significantly different from published normative data for individuals with no documented vestibular deficit.[21] These scores did not improve with CDP-assisted retraining. This is consistent with reports that show static balance frequently resolves spontaneously in days or weeks, whereas dynamic balance, which involves integrating sensory cues that may be in conflict, resolves slowly or incompletely.[9,42]
In contrast to the results with a fixed platform, activation of the sway-referenced platform (SOT conditions 4 to 6), significantly impacted the participants’ ability to maintain their balance prior to retraining. VIS and VEST ratios were significantly lower and highly variable between participants. When instructed to close their eyes on the sway referenced platform (condition 5), some participants performed the same or better than with their eyes open, while others performed markedly worse, suggesting that participants may have had different compensation strategies upon entering the study.
Scores for conditions 4 to 6 improved significantly after CDP-assisted retraining (Fig. 1A). Participants displayed a significantly smaller decrement in postural control on the sway referenced support surface compared to the fixed surface than they did prior to retraining (Table 2). This is consistent with improved capacity for dynamic balance.
The visual nature of the retraining exercises might suggest that computerized vestibular retraining leads to compensation by sensory substitution towards vision; however, we observed that participants were better able to tolerate absent visual information (eyes closed) or conflicting visual information (sway referenced visual surround) after retraining, demonstrating that their improved postural control was not reliant on visual information. Indeed, after treatment the PREF ratio was equivalent to 1 (1.01 [0.88 to 1.05]) indicating that participants were able to maintain their equilibrium as well with no visual information (eyes closed) as with sway referenced visuals. After retraining, SOT scores compared well with published age-matched values for individuals with no vestibular deficit.[21] The ranges and confidence intervals between participants, which had been very wide prior to retraining, decreased significantly.
Taken together, these findings suggest that after computerized vestibular retraining, participants were weighting information from their vestibular organs – either on the unaffected side or from intact organs on the affected side – over vision. Substitution to intact organs on the contralateral side is known to be an important mechanism of compensation for patients with unilateral deficits.[10] In addition, restoration of function of injured semicircular canals has been reported after vestibular neuritis.[49] Lacour posited that restoration of function after vestibular neuritis could come about by regeneration of peripheral sensory hair cells, from new afferents in the vestibular nerve, or by increased synaptic weight of remaining vestibular inputs.[9] Another potential mechanism for restoration comes from experiments in mice that showed the regeneration of hair cells after ablation with diphtheria toxin.[50,51] These data suggest that substitution to contralateral vestibular organs or restoration of ipsilateral function could contribute to compensation in our subjects; however, it is less well understood how common restoration is for patients with unilateral deficits and how important it is for compensation.
Horak found that well-compensated patients had lower vestibulo-ocular reflex gains than poorly compensated patients and suggested that residual, possibly distorted vestibular information was worse than none at all; however, those who learned to use remaining vestibular information from the intact ear performed better than those who relied heavily on visual and somatosensory cues.[10] Our data suggests that such learning was taking place for the participants in this study.
We have published participant-reported outcomes for this cohort in a separate paper.[29] We observed that participants with mild cases, as determined by a pre-treatment dizziness handicap inventory (DHI) score ≤ 30, experienced no measurable benefit from CDP-assisted retraining, as measured by three questionnaires. In the current report, participants with DHI scores ≤ 30, likewise, demonstrated only modest changes in objective posturography performance, whereas patients with pre-treatment DHI > 30 demonstrated significant improvements in the SOT composite score, as well as for the VIS and VEST ratios.
There is conflicting data in the literature about how well subjective patient-reported measures correlate with posturography results.[12,52,53] We found that prior to retraining, SOT composite scores correlated well with all three participant-reported measures, the DHI, the ABC scale, and the FES-I. Changes in the SOT composite scores after CDP-assisted retraining correlated well with changes in FES-I and less closely with the ABC scale. SOT scores did not correlate with the DHI. This may be because, while the FES-I and ABC measure perceptions around performing specific tasks such as walking on a slippery surface or reaching for objects, the DHI includes an emotional domain, as well as functional and physical parameters. It might be expected that emotional status would correlate less well with posturography measures than the task-based questions in the FES-I and ABC.
Participants in this study demonstrated improved postural stability in a dynamic test with conflicting visual and somatosensory cues and this improvement correlated with reduced FES-I scores, which measures perceived fall risk. We have also reported that computerized vestibular retraining is associated with a larger functional stability region – the area over which the center of mass can be displaced by leaning without causing loss of balance (manuscript under review). These three findings – the improved performance in the SOT, the reduced perceived fall risk, and the larger functional stability region – are all consistent with improved postural control and reduced risk of falls for patients with unilateral vestibular deficits.