Optokinetic stimulation for the treatment of vestibular and balance disorders: a systematic review with meta-analysis

Objectives To analyse the effectiveness of optokinetic stimulation (OKS) for improving symptoms and function in patients with vestibular and balance disorders. Methods PubMed (MEDLINE), SCOPUS, Web of Science (WOS), CINAHL Complete, and PEDro databases were searched to identify randomized controlled trials (RCTs) that included patients with vestibular and balance disorders and compared the effects of OKS versus other interventions or no intervention on subjective or objective functional outcomes. Data were analysed by the standardized mean difference (SMD) and its 95% confidence interval. Results A total of 10 studies were selected including 468 patients, 177 of whom received OKS. There were no significant differences in scores on the Dizziness Handicap Inventory (DHI) (SMD = 0.02; 95% CI − 0.18 to 0.23; p = 0.83) or the visual analogue scale (VAS) for vertigo (SMD = 0.16; 95% CI − 1.25 to 1.58; p = 0.82). However, there were statistically significant differences in the timed up and go (TUG) test, with a large effect (SMD = − 1.13; 95% CI -2 to − 0.28; p = 0.009), and in the sensory organization test (SOT), with a medium effect (SMD = − 0.7; 95% CI − 1.21 to − 0.19; p = 0.007). Subgroup analysis showed significant effects of OKS on VAS (p = 0.017), TUG (p = 0.009) and SOT (p = 0.001) only in patients with balance disorders without vestibular disease (p > 0.05). Conclusions OKS may improve dizziness intensity measured with VAS or dynamic balance measured whit TUG and SOT in patients with balance disorders not due to vestibular disease. The quality of the evidence was low or very low due to the small number of included studies. PROSPERO Registry number CRD42023445024.


Introduction
Vestibular and balance disorders have been widely examined since they affect a significant part of the population.In the USA, 35% of adults aged 40 and older have evidence of balance dysfunction [1].These disorders can cause additional problems, such as increased morbidity among older subjects with multiple comorbidities, who may experience falls and greater use of health resources [2].
Optokinetic stimulation (OKS) consists of exposure to visual large-field motion stimuli [3] that could be used to improve a patient's tolerance for instability in motion environments.This approach is commonly prescribed to patients with unilateral spatial neglect poststroke [4].Furthermore, it is increasingly being applied to patients with vestibular disorders, such as dizziness or postural instability while they are in shopping malls or airports; in such environments, many objects are moving at the same time in different directions and at different speeds, thus generating a visually demanding environment [5].
OKS generates visually evoked postural responses that can usually be suppressed by repeated exposure, which indicates learning from ocular and cervical proprioceptors [3] and adaptive neuroplastic changes in visual dependency [6].In fact, it has been observed that the use of OKS activates cortical areas related to visual motion sensitivity and ocular motor and vestibular function [6].
Some recent research has found that this approach applied through virtual reality reduces visual dependence in postural perception in healthy patients [7].It has also been suggested that the application of OKS via virtual reality leads to improvements in visual vertigo symptoms in patients with peripheral vestibular dysfunction [8].Additionally, an earlier study analysed the effects of OKS in 112 patients with unilateral and bilateral vestibular deficits, obtaining normalization of optokinetic nystagmus after 6 to 10 sessions [9].
Based on these findings and considering that, to our knowledge, no systematic review has examined the use of OKS for vestibular disorder rehabilitation, our aim is to carry out a systematic review and meta-analysis to analyse the strongest existing knowledge about the effectiveness of OKS for improving symptoms and function in patients with vestibular and balance disorders.

Study design and protocol registration
This systematic review with meta-analysis was conducted in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [10] and the Cochrane Handbook for Systematic Reviews of Interventions [11].The protocol of this review was registered in PROSPERO with the following registration number: CRD42023445024.

Literature search and databases
To carry out this systematic review with meta-analysis, the PubMed MEDLINE, SCOPUS, Web of Science (WOS), CINAHL Complete, and PEDro databases were searched up to July 2023.In addition, the reference lists of previously published articles, conference proceedings, expert manuscripts, and grey literature were searched.For the search strategy, we identified two search domains based on the PICOS principle [12]: population, patients with vestibular diseases; and intervention, optokinetic exercises.To develop the search strategy, free terms such as "optokinetic" or "vestibular disorders" were combined with MeSH terms indexed in PubMed MEDLINE.Additionally, the Boolean operators "AND" and "OR" were used to elaborate the search strategy.No filters related to language or date of publication were established.The literature search was carried out by two authors, and a third author provided support in this phase.Table 1 shows the search strategy used in each database.

Study selection: inclusion and exclusion criteria
To select the studies to be included in this systematic review with meta-analysis, all retrieved references were thoroughly reviewed by title and abstract.When a potential reference was identified for inclusion in the synthesis, two authors carefully reviewed that reference.Any discrepancy between the reviewers was resolved by consulting a third author, who is an expert in this topic.
The inclusion criteria were as follows: (1) experimental studies, RCTs or pilot RCTs; (2) the sample was made up of patients with vestibular and balance diseases; (3) studies that analysed the effectiveness of optokinetic stimulation; (4) the control group performed another type of intervention or no intervention; and (5) outcome measures related to subjective symptoms and objective balance evaluations.The exclusion criterion was a sample that was not entirely composed of patients with vestibular diseases.

Data extraction
Data from selected studies were extracted and coded into a data sheet standardized Microsoft Excel file created for this review by two authors.Discrepancies were resolved by consulting a third author.The following data were collected AB ("Optokinetic" OR "Optokinetic training" OR "Optokinetic response" OR "Optokinetic stimulation") AND ("Vestibular disease" OR "Vestibular disease" OR "Vertigo" OR "dizziness") Web Of Science TOPIC ("Optokinetic" OR "Optokinetic training" OR "Optokinetic response" OR "Optokinetic stimulation") AND ("Vestibular disease" OR "Vestibular disease" OR "Vertigo" OR "dizziness") PEDro Optokinetic AND vestibular SCOPUS TITLE-ABS-KEY (("optokinetic" OR "optokinetic training" OR "optokinetic response" OR "optokinetic stimulation") AND ("vestibular disease" OR "vestibular disorders" OR "vertigo")) from each of the included studies: (1) general characteristics (authorship, publication date, country and funding); (2) patient characteristics (total sample size, type of vestibular disease, number of participants in each group, age, and sex); (3) characteristics of the experimental group and the control group (type of intervention, number of sessions); and (4) data from variables (variable assessed, test employed and time-point assessment).The data used to perform our meta-analysis were the means and their standard deviations and/or differences between groups and p values.When a study did not provide data related to standard deviation, this was obtained from other data presented in the study, such as range, interquartile range, or standard error, as indicated by standardized statistical procedures [11,13].

Variables
The variables of interest were as follows: disability, which was measured using the Dizziness Handicap Inventory (DHI); postural stability, which was measured using the sensory organization test (SOT); perceived level of dizziness, which was measured using the visual analogue scale (VAS) for dizziness; and dynamic balance and risk of falls, which was measured using the timed up and go test (TUG).

Risk of bias and methodological quality assessment
The evaluation of the risk of bias in each included study and of the quality of the evidence of the main findings was carried out by 2 authors independently.Any doubts were resolved by a third author.Initially, the methodological quality and the risk of bias of the included studies were assessed using the PEDro scale [14].The PEDro scale is an 11-item checklist that can be scored "yes" if the criterion is met and "no" otherwise.The total score can range from 0 (very low methodological quality and high risk of bias) to 10 (excellent methodological quality and very low risk of bias), while item 1 is not used to calculate the total score because of its relationship with external validity [15].The methodological quality of a study is considered "excellent" if it reaches a score of 9-10 points; "good" when the score ranged from 6 to 8 points; "moderate" when the score ranged from 4-5 points; and "poor" when the score was 3 or below.
Second, the quality of the evidence was analysed with the GRADE scale (Grading of Recommendations, Assessment, Development and Evaluation) and with the GRADE checklist by Meader et al. [16].To express a level of evidence for a result, five parameters are assessed: risk of bias in the included studies, risk of publication bias in the results, inconsistency, imprecision, and indirectness.Both the risk of publication bias and heterogeneity will be explained in the statistical analysis section.The precision of the results can be high (> 10 studies or > 300 participants), moderate (10-5 studies and 300-100 participants) or low (< 5 studies and < 100 participants).Finally, evidence is considered to be indirect when scales that do not directly measure the variable of interest are used.The combination of these 5 parameters can make the quality of the evidence high, medium, low or very low.The level of evidence will be lowered for each point that is not met.
Both the evaluation of the methodological quality and the analysis of the quality of the evidence were carried out by two authors independently, and discrepancies were resolved by a third author.

Statistical analysis
Statistical analysis was performed by 2 authors using version 4 of the Comprehensive Meta-Analysis [17].Meta-analysis was only performed if at least two studies provided data for this meta-analysis.Following the recommendations of Cooper et al. [18], we applied a DerSimonian and Laird random-effects model to estimate the pooled effect [19].When heterogeneity in a fixed effects model was > 50%, the random-effects model was applied.The pooled effect was calculated using Cohen's standardized mean difference (SMD) and its 95% confidence interval (95% CI).The effect can be null (SMD 0), low (SMD 0.2-0.4),medium (SMD 0.4-0.7),and large (SMD > 0.8) [20].The findings from each meta-analysis are shown in forest plots [21].The risk of publication bias was assessed by analysing the symmetry of the funnel plot (if the funnel plot was asymmetric, it indicates the possibility of risk of publication bias) and the p value for Egger's test (if p < 0, 1 increases the risk of publication bias) [22,23].In addition, we used trim-and-fill estimation to calculate the adjusted event rate, taking into account a possible risk of publication bias to know whether the original findings are underestimates or overestimates [24,25].When the difference between the original and adjusted pooled effect was > 10%, the level of evidence was lowered by one level [26].Heterogeneity was analysed by calculating the p value for the Higgins Q test (p < 0.1 indicates the presence of heterogeneity) and the degree of inconsistency (I 2 ), which classifies heterogeneity as low (< 25%), medium (25%-50%), or large (> 50%) [27,28].
Additionally, a subgroup analysis was planned taking into account the origin of the pathology of the included patients.The pathological subgroups were: (1) vestibular disease and (2) balance disorders without vestibular alteration.

Assessment of the methodological quality and risk of bias of the included studies
The methodological quality of the studies included in the review was moderate according to the PEDro scale -the mean score was 4.8 points out of 10.Two studies presented good quality [32,38], and the remaining eight had moderate quality [29][30][31][33][34][35][36][37].None of the studies could be blinded to the participants or the therapists, so all of them presented a risk of performance bias.Other possible identified biases were selection, detection and reporting bias.Table 3 shows the PEDro score for each study.

Timed up and go (TUG) test
In this meta-analysis, 2 studies [30,33] with 4 independent comparisons that provided data from 225 patients with balance disorders without vestibular disease (57.3 per study) were included.Our findings showed a large effect (SMD = − 1.13; 95% CI − 2 to − 0.28; p = 0.009) favouring optokinetic therapy (Table 4, Fig. 4).Trim-and-fill estimation showed a possible risk of publication bias and reported an adjusted effect (adjusted SMD = − 1.5) major than the original effect, taking into account this risk of publication bias.This evidence indicates that the original effect can be underestimated by publication bias.However, the level of heterogeneity in a random-effects model was large (I 2 = 61.4%;Q = 7.8; df = 3; p = 0.05).

Qualitative synthesis
Additionally, two studies provided data that could not be integrated with that obtained by the rest of the studies.Rossi-Izquierdo 2018 et al. [31], found a significant reduction in the average number of falls at 12 months of follow-up in the OKN group, which went from 17.07 to 4.43 (p = 0.011) compared to the control group that showed a reduction from 3.36 to 2.61, which was statistically non-significant (p = 0.166).In the study by Bunn 2015 et al. [32], the OKN group showed a reduction in the Sway of the Centre of Pressure that was greater than the control group in moving visual scenery (MVS) situations.These results provide evidence of an improvement in balance in dynamic conditions and a reduction in the number of falls in subjects undergoing therapy with OKN.

Discussion
Our review aimed to locate and analyse the best evidence on the effectiveness of OKS for the improvement of symptoms and function in subjects with vestibular and balance disorders.Our findings show limited and generally low-quality evidence indicating that optokinetics had a large effect on improving dynamic balance (as measured by the TUG and CDP) and no effect on subjective measures such as the DHI and VAS for vertigo.Our review found low-quality evidence that optokinetics have no effect on improving disability due to dizziness (as measured by the DHI) when compared to other interventions, contrary to the investigators' initial hypothesis.Some previous studies have shown that vestibular rehabilitation therapy and corticosteroids can be a good combination for improving DHI in vestibular neuritis [39], propranolol can be effective in patients with vestibular migraine [40], and vestibular rehabilitation can be effective in patients with acute vestibular disorder [41] and in Meniere's disease [42].Therefore, and in the absence of more robust evidence, vestibular rehabilitation and different pharmacological measures could be better measures to improve subjective symptoms.Virtual reality-assisted therapy has been shown to have additional benefits in patients with vestibular disorders compared with conventional vestibular physical therapy for the improvement of the DHI total scores and its subscales [43].It is important to note that most of the comparisons in our meta-analysis consisted of active therapy that has been   shown to be effective in improving subjective symptoms, so our findings can be interpreted as meaning that OKS may have a similar but not superior effect to these types of interventions, such as Cawthorne-Cooksey exercises or rehabilitation with CDP.
Additionally, in relation to subjective symptoms, another variable used in the trials included in the review was the VAS of vertigo.With very low quality of evidence, our findings showed that there was no effect of OKS when compared with other interventions or no intervention.In patients with acute vertigo, the use of benzodiazepines and antihistamines has shown immediate beneficial effects on the VAS for vertigo [44].In subjects with peripheral vestibular dysfunction, the use of virtual reality as a vestibular rehabilitative intervention was able to improve VAS scores [45].
Regarding objective measurements, very low-quality evidence from four comparisons found a large effect of OKS for the improvement of dynamic balance (measured with the TUG test).The evidence on improvement of this variable in patients with vestibular and balance disorders is poorer and derives from isolated clinical trials, where improvement was observed in all arms of the trial.An improvement in the TUG test has been found with home rehabilitation programs in patients with chronic vertigo due to peripheral vestibular impairments [46] through the use of vestibular rehabilitation in older people with chronic dizziness [47], although other treatments, such as the use of rehabilitation with CDP, seem to have no effect on improving stability in older patients with instability [48].
Regarding SOT, few interventions have been successfully tested to improve composite measurement.Vestibular rehabilitation seems to improve SOT in patients with multiple sclerosis [49], but the results of the treatment of peripheral or central vestibular disorders using virtual reality-assisted therapy do not seem to provide a significant effect [43].In our study, the effect of OKS on improving SOT was moderate when compared with other interventions but with a very low quality of evidence.
This article has several limitations.First, the evidence found is scarce, with few randomized clinical trials that adopted a very heterogeneous methodology, which makes it difficult to establish specific recommendations.Therefore, the quality of the evidence is low or very low in the metaanalyses, and partially supported recommendations for practice can be made.Additionally, recommendations for new research can also be made.Despite the above limitations, to the best of our knowledge, this is the first review that has been carried out on the topic.These findings can improve the general approach and be a starting point for new and more oriented research.

Conclusions
In view of our results, it can be concluded that there is low and very low evidence that OKS is not better than other interventions or no intervention for the improvement of disability due to dizziness or the subjective perception of dizziness, respectively, in subjects with vestibular and balance disorders.
There is very low evidence indicating that OKS has a large beneficial effect on dynamic balance (measured with the TUG test) in subjects with balance disorders when compared with other interventions or no intervention.We also found very low-quality evidence indicating that OKS has a medium-sized beneficial effect on balance (as measured by the SOT) in subjects with vestibular and balance disorders when compared with other interventions or no intervention.
In reference to the effect on different pathological conditions, the OKS showed a significant effect for the decrease in the intensity of dizziness and for the improvement of dynamic balance measured with TUG and SOT in patients with balance disorders of diverse origins, but no significant effect was found in subjects with vestibular disease for none of the variables analysed.
OKS offers promising results in improving objective measurements of balance, although the methodological heterogeneity of the trials included in the review does not allow more precise conclusions to be drawn.Therefore, new research with better standardized methodology is needed to analyse the real effect of optokinetic stimulation in different groups of patients.
Funding Funding for open access publishing: Universidad de Jaén/ CBUA.

Fig. 2 Fig. 3
Fig. 2 Forest plot showing the pooled effects of Optokinetic Stimulation compared with other interventions on the Dizziness Handicap Inventory scores

Fig. 4
Fig. 4 Forest plot showing the pooled effects of Optokinetic Stimulation compared to other interventions on the Timed Up and Go Test

Fig. 5
Fig. 5 Forest plot showing the pooled effects of Optokinetic Stimulation compared to other interventions on the Sensory Organization Test (SOT)

Table 1
Search strategy

Table 2
Main characteristics of the studies included in the review

Table 4
Main findings in meta-analyses and GRADE assessment