1.1 Bilateral vestibulopathy

Bilateral vestibulopathy (BV) is defined as a chronic severe loss of function of both balance organs (quantified by a significantly impaired or absent function of the vestibulo-ocular reflex), which represents a major handicap involving strong balance disturbances, higher risk of falling, oscillopsia (i.e. a symptom of blurred vision during head movements), and associated loss of autonomy and quality of life [1–4]. About three million people globally are believed to be affected by bilateral vestibulopathy, but this is most likely an underestimation due to misdiagnosis [5]. At this point, only limited rehabilitation strategies are possible, focusing on compensation, adaptation and substitution [6], instead of treatment or restoration of vestibular function.

1.2 The vestibular implant

The concept of a vestibular implant (VI) was conceptually proven in animal studies [7–9]. The VI is a neuroprosthesis that aims to restore vestibular function. Similar to a cochlear implant (CI), the dysfunctional sensory input in the vestibular end organ can be bypassed by directly stimulating the vestibular nerves [10]. Head motion is sensed by gyroscopes and accelerometers. A processor converts this head motion information into electrical signals. The implanted stimulator transmits these signals to the vestibular (ampullary) nerves [10] via electrodes implanted in the vicinity of the neural targets.

Currently, different VI prototypes are being evaluated regarding feasibility and functionality by several research groups worldwide [11, 12, 13–21]. Since hearing can be (further) compromised when the labyrinth is opened during surgery [22], a combined VI and CI prototype was developed, referred to as a vestibulocochlear implant (VCI), to restore vestibular function and hearing function concurrently. The first VCI prototypes containing one vestibular electrode contact were implanted in humans in 2007. These prototypes evolved into multichannel devices, implanted for the first time in humans in 2012. These multichannel devices contain multiple vestibular branches with on top of each branch one electrode contact. In the last decade, multiple promising results involving VI stimulation were reported, including a (partial) restoration of the vestibulo-ocular reflex (VOR) [15, 16]. This restoration resulted in an improved dynamic visual acuity, which has the potential of restoring gaze stabilization [19, 23, 24]. Additionally, natural frequency-dependence of VI stimulation was demonstrated [17] and eye movement alignment was investigated [25]. These studies show that subjects tolerate and adapt to VI stimulation and that electrically evoked VOR (partly) mimics responses also seen in the “normal” vestibular system. Moreover, postural responses were elicited and influenced by electrical stimulation [26, 27]. Improvements in postural responses and gait were described in patients receiving stimulation of the vestibular organ [28].

Until now most studies investigated the acute and intermittent use of a VI. Recently, a few research groups started to evaluate chronic effects of semicircular canal stimulation [25, 29] and otolith stimulation [30]. These groups showed stable restoration of the VOR over time and positive subjective outcomes. There were no adverse effects of prolonged continuous daily stimulation with a VI-only device [25]. However, based on these studies alone, with limited patient numbers and specific etiologies, it is not possible to generalize the results to a larger patient population or create more general statements of rehabilitation prospects in the long term. To the best of our knowledge, randomized controlled trials (RCTs) measuring VI performance in a number of strategies are currently nonexistent. Most publications focus on the efficacy of one fitting paradigm, in order to compare these settings with an OFF-set mode. To the best of our knowledge, there are no studies comparing a number of different stimulation modes based on different fitting parameters. Moreover, long periods of continuous stimulation are at this point only investigated with a VI-only device, and not yet with a VCI. Therefore, there is a need to investigate the safety and efficacy of prolonged stimulation with a VCI in a randomized controlled trial setting.

1.3 The VertiGO! Trial

To take vestibular implant research a step further, a new clinical trial will be conducted in a specific patient population suffering from both BV and severe sensorineural hearing loss. This trial primarily focuses on the overall safety and efficacy of prolonged VCI stimulation. A new refined multichannel VCI prototype will be evaluated, including concurrent multichannel VI and CI stimulation. Furthermore, this study aims to provide fundamental insights into the programming (fitting) of future VI devices by comparing to the condition without VI stimulation to three different modes of VI stimulation: (A) motion-modulated stimulation with baseline stimulation, (B) motion-modulated stimulation with reduced baseline stimulation, and (C) baseline stimulation (no modulation). A schematic representation of these three modes of stimulation is shown in Methods, paragraph 2.2.5. Baseline stimulation is defined as a continuous constant-amplitude electrical pulsatile signal, mimicking the “resting” activity of the vestibular nerve. Baseline stimulation is not simply considered as a control condition in this trial. Prolonged unmodulated stimulation might enhance the residual vestibular function (including the potential effects of stochastic resonance) [11]. Moreover, this unmodulated strategy is already applied in studies investigating the potential effect of otolith stimulation [30] and in studies investigating the effect of galvanic vestibular stimulation [31]. In these studies, the constant pulse train of baseline stimulation is believed by some researchers to mask the absence of physiological stimuli in BV patients [30].

By comparing two types of prolonged motion-modulated stimulation with different levels of baseline stimulation, it is expected to further refine the potential efficacy of vestibular stimulation. The potential efficacy of every stimulation paradigm will be evaluated and compared in terms of maximum effectivity, tolerance and compliance, flexibility and sustainability. A sham-like stimulation, in the form of ‘reversed’ stimulation, will not be applied due to unwanted added burden for the participants (e.g. significant prolongation of study interventions). Secondary goals involve exploring potential interactions between simultaneous CI and VI stimulation, overall auditory performance with the device (CI performance), and long-term follow-up of the VCI-implant over a period of 5 years. This study also aims to collect data on patient experiences and perspectives as input for early health technology assessments. Furthermore, this trial will serve as a proof of concept for a VI-fitting program and a VI-specific rehabilitation program. Finally, different stimulation parameters (i.e. modulation types and transfer functions) and more fundamentally central responses to VI stimulation will be evaluated.

Four hypotheses are addressed in the VertiGO! Trial:

1. Prolonged stimulation with a VCI is safe and effective; VI stimulation will result in improved dynamic visual acuity, increased postural control, higher VOR gain values and fewer BV-related symptoms compared to no VI stimulation.
2. Motion-modulated stimulation with 50% or reduced baseline stimulation will result in improved dynamic visual acuity, increased postural control, higher VOR gain values and fewer BV-related symptoms compared to baseline stimulation alone. Furthermore, it is expected that the level of baseline stimulation influences VOR gain resulting in higher but asymmetric VOR gain values and improved dynamic visual acuity in situations with reduced baseline stimulation.
3. Responses to VCI stimulation are expected to change during the first days of stimulation, since most of the adaptation to misalignment that might be present between the actual head rotation and the electrically evoked VOR is expected to be compensated for during the first week of stimulation [32]. Alongside the potential VOR directional plasticity, it is likely that other adaptation processes would occur during the first weeks of stimulation, potentially influencing the stimulation efficiency and induced responses. This is already a known phenomenon in CI-only implementation [33]. Therefore, it is expected that outcomes will improve throughout each prolonged stimulation period. The effect of acute stimulations are expected to stay stable during the whole follow-up period.
4. VI stimulation is expected to interact with CI stimulation and vice versa. It is expected that these bidirectional effects on vestibular or hearing performance are not clinically relevant.

2.1 Design, setting and recruitment

A single-center clinical trial will be performed in the Maastricht University Medical Centre+ (MUMC+), a tertiary university medical center, in cooperation with Geneva University Hospitals. The proposed design will be a single-blind randomized crossover controlled trial. The protocol is approved by the local medical ethical committee (MUMC+, NL73492.068.20/ METC 20–087), is registered at clinicaltrials.gov (NCT04918745) and will be conducted in accordance with the Declaration of Helsinki. All subjects included in this study were informed by a medical doctor, which is part of the research team and have signed an informed consent (S1 Appendix, Supporting information, model consent form) form before participation. The trial will be monitored by the local ethical committee, independent from investigators. The research protocol will be made available through the supplemental materials published with this article. Every single modification to the protocol will be communicated to the local medical ethical committee. In case a formal amendment to the protocol is required and approved, this amendment will be included in a new version of the informed consent prior to implementation according to the local regulations. The protocol with formal amendments will be made available together with the final trial reports after finalization of the trial. Modifications that do not require a formal amendment will be documented with version control. The SPIRIT reporting guidelines are used in reporting this trial (S2 Appendix, Supporting information, SPIRIT guidelines) [34]. See S3 Appendix, Supporting information, for the World Health Organization Trial Registration Data Set.

The trial aims to include a minimum of eight, with a maximum of 13, patients with diverse BV etiologies and different durations of disease, to represent a heterogeneous patient population. Recruitment started on 01 July 2021, will end in the first quarter of 2024, and takes place in the MUMC+. Every patient that is believed to suffer from a balance disorder or is candidate for cochlear implantation, according to their health records, undergoes a vestibular screening, as part of the standard clinical routine at the ENT department of the MUMC+. Patients who appear to be eligible for (and are interested in) participating in VI-related studies serve as the population base and will be personally informed by an MD about the VertiGO! Trial in the outpatient clinic during one of their regular follow-up visits at the ENT department. The time after diagnosis of vestibulopathy and/ or hearing loss at inclusion may vary. No advertisements will be utilized. Adequate enrolment will be promoted by actively monitoring outpatient clinics devoted to vestibular patients and potential CI candidates.

In order to be eligible to participate in this study, a subject must meet all of the following criteria, which are based on the recent opinion statement about vestibular implantation criteria [35]:

1. Be aged 18 years or older;
2. Suffer from disabling symptoms of postural imbalance and/ or oscillopsia;
3. Have reduced or absent bilateral VOR function based on at least one of the tests shown in Table 1 (meeting criteria A, with the other tests meeting criteria B of the vestibular implantation criteria statement [35]);
4. With an onset of bilateral vestibulopathy after the age of two;
5. With documented vestibular dysfunction from a peripheral origin;
6. Have severe to profound sensorineural hearing loss, meeting Dutch CI-candidacy criteria (in the ear to be implanted);
7. Documented post-lingual onset of profound hearing loss (> 4 years of age).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Vestibular inclusion criteria.

https://doi.org/10.1371/journal.pone.0301032.t001

A subject is excluded when meeting one or more of the following criteria:

1. Radiographic evidence of absent/abnormal vestibular/cochlear end-organs impeding implantation;
2. Has any sign of central vestibular/cochlear dysfunction or structural vestibular/cochlear nerve pathology;
3. Has received a cochlear implant in the contralateral ear from another brand than the VCI brand (MED-EL, Innsbrück, Austria);
4. Is unwilling to stop the use of vestibular suppressant medications;
5. Shows signs of recovery of vestibular function, oscillopsia and imbalance within 6 months from the onset of symptoms;
6. Is not a proficient speaker of the Dutch language;
7. Presents any contra-indication for VCI surgery;
8. Shows non-vestibular pathologic conditions of sufficient severity to confound vestibular function tests;
9. Presents any contra-indication for MRI or CT imaging prior to surgery.

All patients who signed an informed consent document witnessed by a medical doctor and passed all screening tests, are identified by an anonymized 2-digit patient ID (CVI-xx).

2.2 Study interventions

A complete overview of study visits, including all scheduled procedures and outcomes per visit, is presented in Fig 1. A timeline of all study-related visits per subject is shown in Fig 2. Briefly, after inclusion, patients will be scheduled for VCI surgery. After implantation, subjects will follow a regular CI rehabilitation program with the addition of one postoperative control visit, a cone beam computed tomography (CBCT) scan and four hours of acute VI testing. Once the CI rehabilitation is finished, a four-day VI fitting period, one day of baseline testing and three periods of four days (three-period crossover design) of prolonged VI stimulation will be scheduled as the main trial experiments of this study outline. During each regular CI follow-up visit, some additional VI-related follow-up tests will be conducted. In total, a five-year follow-up period after the first implant fitting will be performed. Before every day of study interventions, the participant will be personally contacted by one of the researchers to explain all upcoming experiments of the next day, the importance of specific instructions and preparations will be underlined. One week after each prolonged stimulation week, the participant will be personally contacted to ask about any positive or negative effects they are experiencing after the intervention (part of the safety outcomes). Due to frequent contact moments between participants and researchers, the participant retention will be promoted.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schedule of enrollment, interventions and assessments per study visit per subject.

aECAP (auditory electrically evoked compound action potential), ART (audio response telemetry), BVQ (bilateral vestibulopathy questionnaire), CAREN (computer assisted rehabilitation environment system), CBCT (cone beam computer tomography), CI (cochlear implant), CNC (consonant-nucleus-consonant), CT (computer tomography), cVEMP (cervical vestibular evoked myogenic potentials), DHI (dizziness handicap inventory), DiN (Digits-in-Noise), DQ (daily questionnaire), DVA (dynamic visual acuity), EQ-5D-5L (euroqol five-dimensional questionnaire), FES-I (falls efficacy scale international), fHIT (functional head impulse test), HADS (Hospital anxiety and depression scale), HUI-3 (health utility index mark 3), ICECAP-A (ICEpop capability measure for adults questionnaire), mini-BESTest (mini-balance evaluation systems test), MRI (magnetic resonance imaging), OSQ (oscillopsia severity questionnaire), oVEMP (ocular vestibular evoked myogenic potentials), PSFS (patient-specific functional scale), SSQ-12 (speech spatial and qualities of hearing scale), TQ (tinnitus questionnaire), VBR (vestibular brainstem response). VCI (vestibulocochlear implant), vECAP (vestibular electrically evoked compound action potential), vHIT (video head impulse test), VI (vestibular implant). *If tinnitus complaints are higher at post-operative measurement, follow-up measurements will be performed. ⁱ Part of regular CI work-up. ^# Measured in two conditions: static and dynamic.

https://doi.org/10.1371/journal.pone.0301032.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Timeline per subject of the VertiGO! trial.

Timeline in months (m). VCI (vestibulocochlear implant), CI (cochlear implant), VI (vestibular implant).

https://doi.org/10.1371/journal.pone.0301032.g002

A classical RCT with parallel groups multiplies the number of participants which is not feasible in this specific intervention. Moreover, it increases the risks of suboptimal treatment and outcomes in one of the groups. Therefore, we propose an alternative crossover study design that is feasible, takes a within-subject perspective, and in which the participants serve as their own control. For each subject, the three stimulation modes will be randomly assigned to one of three prolonged stimulation periods by using a block randomization-based design [36].

Since there are three stimulation modes which will be randomized over three time periods, there are six distinct algorithm sequences. To equally distribute all the patients over six algorithm sequences a block size of six is used. A custom “urn” will be created in the following way: Each of the six possible algorithm sequences is added to the urn once. Additional algorithm sequences are chosen at random without replacement from the total of six algorithm sequences and also added to the urn. This is done until the amount of algorithm sequences in the urn is equal to the amount of patients which will be included. The urn now contains as many allocation possibilities as there are patients, with each allocation possibility featuring at least once and no more than twice. Randomization will be performed in the period after VI fitting and before the prolonged stimulation periods. Previous experience with vestibular stimulation illustrated that turning on stimulation is clearly noticeable by both the patients and the examiners [12]. To decrease patient-induced bias during the prolonged stimulation periods, the patients will be blinded as to which stimulation mode is used in which week. Each stimulation mode is coded with a letter, code breaks to the patient should only occur in circumstances when knowledge of the actual stimulation mode is essential for further management of the patient. Although these circumstances seem highly unlikely, participants and/ or researchers who believe that unblinding is necessary, are encouraged to discuss this with the independent physician of this trial. In all other circumstances, researchers are encouraged to maintain blinding totally. All code breaks will be reported to the principal investigator. Unblinding will not be a reason for study discontinuation.

2.2.5 The prolonged stimulation period.

One prolonged stimulation period consists of four days of stimulation, eight hours a day. This schedule will be performed three times (Fig 3). The only parameter that will randomly change between each period is the used stimulation paradigm. The three stimulation paradigms are schematically shown in Fig 4. The parameters (e.g. level of baseline stimulation and modulation) of each paradigm will be determined after analyzing the results of VI fitting. The DR will be determined on the first day of every prolonged stimulation period, in order to verify the stability of results found earlier during acute stimulation and VI fitting. Most outcome measures will be measured at least twice per week, as described in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Example schedule of a prolonged stimulation period.

^A Steady state habituation and short fitting check. ^B Impedance, UCL and VAS-based list of questions. It should be noted that actual timeslots may differ between subjects. CAREN (computer assisted rehabilitation environment system), CNC (consonant-nucleus-consonant), DiN (Digits-in-Noise), DVA (dynamic visual acuity), mini-BESTest (mini-balance evaluation systems test), PSFS (patient-specific functional scale), UCL (upper comfortable level), VAS (visual analog scale).

https://doi.org/10.1371/journal.pone.0301032.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Three modes of stimulation by the vestibular implant.

(A) Motion-modulated stimulation with baseline stimulation. (B) Motion-modulated stimulation with reduced baseline stimulation. (C) Baseline stimulation (no modulation). Baseline stimulation is given as a constant-amplitude electrical pustule signal in all three stimulation modes. Excitatory modulation is shown in light gray and inhibitory modulation is shown in dark gray.

https://doi.org/10.1371/journal.pone.0301032.g004

Apart from the wide range of experimental procedures, two sessions of a 60-minutes physiotherapeutic supervised exercise program will be scheduled. A personalized mixture of gaze stabilization exercises, habituation exercises and balance and gait training will be carried out.

2.3 Hardware and software

The VCI implant is a modification of a MED-EL Synchrony cochlear implant (PIN Mi1200 variant) with 12 electrode contacts (MED-EL, Innsbruck, Austria). In the VCI configuration, the three most basal electrodes are separated from the main cochlear electrode array into three individual branches (Fig 5A). These three individual branches are designed to be inserted into the semicircular canals, to allow for stimulation of the vestibular system via the ampullary nerves.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5.

(A) The multichannel VCI prototype. (B) The Audio-Motion Processor (AMP) with coil. Reprinted from [43] under a CC BY license, with permission from the SWISS Foundation for Innovation and Training in Surgery and MED-EL, original copyright 2023.

https://doi.org/10.1371/journal.pone.0301032.g005

Participants will be provided with a Sonnet 2 speech processor, used during CI rehabilitation and daily outside the hospital. During VCI experiments, participants will exchange their clinical Sonnet 2 speech processor for a research VCI processor: the audio-motion processor (AMP; Fig 5B). The AMP contains an audio processor unit, and two inertial measurement units, which are housed in the coil and in the audio-processor connector. The AMP includes a base-unit, which has been designed as a research platform to provide a user interface to select various experimental setups, and to connect to the research software (AmpFit). The software packages and all equipment are developed and provided by MED-EL (Innsbruck, Austria).

2.4 Outcome measures

Extensive outcome testing scheduled over the course of the study period, listed in Fig 1, is categorized into main primary outcomes, additional primary outcomes, and secondary outcomes, matching the wide range of objectives of the VertiGO! trial. By measuring outcomes with no vestibular stimulation at baseline (vestibular electrodes turned off, cochlear electrodes turned on) and during all three prolonged stimulation periods (vestibular and cochlear electrodes turned on), outcome values will be compared within four program versions. Moreover, by using a test–retest scheme within each prolonged stimulation period, this trial aims to investigate the potential effect of prolonged stimulation within one stimulation paradigm.

2.5 Sample size calculation

Sample size calculation is based on the primary study outcome: DVA score (see 2.4.1). Visual acuity (VA) can be expressed in several different units. However, for data analysis it is important to use the logarithm of the minimum angle of resolution (logMAR), since this unit scales linearly with the geometric progression of the visual acuity chart [73]. Data from previous measurements on the influence of vestibular stimulation on DVA was used for the sample size calculation [19]. Although this data was obtained with a different version of VCI (which did not yet allow simultaneous VI and CI stimulation), it is the closest analog available to approximate the results of this trial. To account for the between-patient difference in VA, the data from the previous experiment, expressed in logMAR, were normalized to the static VA measured for each patient. This yields mean logMAR values of 0.065 (SD 0.061) and 0.205 (SD 0.066) for VI active and VI inactive, respectively. Since the size of the dataset is too small to determine whether or not the data is normally distributed, the sample size was calculated for a two-sided Wilcoxon signed-rank test. Significance level (alpha) was set at 0.05 and power was set at 0.80 [74, 75]. Both effect size and minimal sample size were calculated with G*Power 3.1.9.7 [76]. Effect size was determined to be 2.20, which resulted in a minimum sample size of five. Given the available resources (e.g. prototype hardware and software, budget limitations due to high (surgical) costs of the procedures) and to provide this statistical background with a safety margin, this trial will aim to include a minimum of eight, with a maximum of 13, patients. Although this is still a relatively small number, both safety and efficacy of the new prototype VCI is believed to be investigated thoroughly with this sample size due to the high frequency of measurements and a total follow-up duration of 5 years.

2.6 Statistical methods, data reporting and analysis

The drop-out rate is expected to be very limited in this trial, since patients are extensively counseled before inclusion, closely followed during their standard CI rehabilitation trajectory and contacted before and after each VI related study visit. Apart from the yearly follow-up visits, all visits will be scheduled shortly after each other. Therefore, the drop-out rate is estimated to be zero. However, frequency and reasons for drop-out will be recorded.

All anonymized data will be stored and secured in the digitalized data management software, provided by the local ethical committee (i.e. MACRO). From this software, datasheets will be exported before data cleaning and initial analysis is performed using R [77]. An interim analysis of the data will be performed after a minimum of five VI fitting periods using R. Final statistical analysis will be performed using SPSS 29.0.

Due to the small size, it is expected that the assumptions for parametric tests will not be met for most datasets. To evaluate the effect of VI stimulation in the main and additional primary outcomes (i.e. DVA, fHIT, 3D HIT, Rotatory chair, self-motion perception, Mini-BESTest and CAREN-testing), repeated measures will be collected with multiple within-factors being present. Per subset of data, the median, first and third quartiles will be presented. It is expected that the Friedman test will be the most suitable test to further evaluate statistical significant differences in the datasets. Wilcoxon signed-rank test is expected to be the most suitable test for post hoc analysis. The statistical analysis of the auditory outcomes will be similar. Analysis will be performed in all randomized participants, as-randomized In case of missing data, the reason for missing will be reported for each stimulation mode and compared. No imputation methods will be used. Formal adjustment will be performed in case of multiplicity issues.

Questionnaire outcomes will be analyzed similarly to the primary outcomes. The semi-structured interview will coded and analyzed thematically, following the criteria of Braun & Clarke.

2.7 Ethics

Health risks specifically associated with study participation are limited to risks related to the surgery, which are similar to conventional CI surgery, and exposure to ionizing radiation from perioperative fluoroscopy and CT-like imaging and postoperative CBCT. The maximum calculated dose was chosen in consideration of surgical and investigational yield, not exceeding risk category IIa of the International Committee of Radiological Protection (between 0.1 and 1 mSv). Participants will be required to devote time, effort and attention to the study. Participants agree and understand that vestibular stimulation is only applied within the hospital and is not of direct added value in their home situation. Written informed consent was obtained, witnessed by a medical doctor (BLV or BV), using an informed consent form (S1 Appendix, Supporting information, model consent form). Participants will be compensated for their additional traveling expenses and traveling-related burden will be limited by offering a hotel room close to the hospital, during fitting and testing weeks. Post-trial care will be provided, if applicable, conform local ethical committee standards.

3.1 Development of a clinically usable VI fitting strategy

The approach proposed in this trial is expected to obtain essential knowledge about VI fitting. This approach aims to develop a fitting method, which can be translated into clinical practice (e.g. timing of sessions, setups). At this point, no data is present to build a clinical fitting scheme. By investigating the influence of different variables on DR, such as (level of) baseline stimulation, signal waveform (i.e. square or sine), pulse characteristics (i.e. phase duration, rate) and time (i.e. development over time after surgery), fitting will be further clarified. After initial determination of the DR, different modulation settings can be used. By varying the encoding of head movements (input) in modulations of the baseline signal (output), the effect of modulation settings can be investigated throughout the DR. Although this trial aims for a standardized fitting protocol draft, it might be possible or even essential to develop personalized fitting strategies. By comparing individual DR values structurally in these different paradigms, interindividual differences will be investigated further.

3.2 Potential CI—VI interactions and the effects of interleaved stimulation

Since vestibular structures are position close to the cochlear structures [78], it is possible that the currents delivered by the CI electrodes spread to the surrounding vestibular structures and that the same applies for currents delivered by the VI electrodes to the auditory structures. Secondly, in order to enable both cochlear and vestibular stimulation, specific auditory gaps are introduced in the classical auditory stream [79]. These auditory gaps, which are relatively large compared to the pulse duration of regular CI stimulation, are inevitable when sequential stimulation of both vestibular and cochlear electrodes is addressed within one stimulation frame. Without these gaps, it would not be possible to stimulate both vestibular and cochlear electrodes concurrently [79]. The spread of excitation and the gaps in the auditory stream might influence CI and VI performance bidirectionally. The interaction effect of brief trains of combined stimulation of the cochlear and vestibular arrays of the VCI was investigated previously [80]. Although this study only showed fundamental results from a small sample size, it suggests that interleaved stimulation might elicit potentially clinically relevant changes in electrically evoked eye velocity, loudness percept and pitch percept [80]. However, introducing specific gaps and distributions in the auditory stream for a cochlear implant did not show deterioration of speech perception [79].

It is known that CI stimulation possibly impacts vestibular test results [81]. Recently, this phenomenon was used to investigate possible vestibular improvement in children. A significant reduction of the number of falls and a better postural control were measured in conditions where bilateral head-referenced CI feedback was provided [82, 83]. However, at this point it is unclear whether co-stimulation and/or restored auditory cues contributed most to the improved balance function.

These results underline the necessity to evaluate the potential interactions between VI and CI more extensively on a clinically relevant level. By including an auditory test set during the prolonged stimulation periods and during follow-up visits, the CI performance can be compared between the regular CI processor versus the VCI research processor with and without VI stimulation turned on.

3.3 Limitations

An eminent problem when implementing a crossover design in order to evaluate VCI performance is the occurrence of a first-order carry-over effect [84]. However, previous studies with repeated acute stimulations showed stable results over time with as little as one day of wash-out in between [16–18, 39, 85]. Moreover, the wash-out between each stimulation period is chosen to be at least three days to further decrease the change of carry-over effects.

Previously, a decrease in electrically elicited eye movement responses was measured after prolonged stimulation in both animal and human studies [86–88]. This might be due to a suppression at vestibular neuron synapses, resulting in a decrease in neuronal sensitivity [86]. At this point, the timing and degree of this suppression are not clear. Therefore, it is possible that changing degrees of suppression of eye movements are present during different stages within this trial. Due to the limited duration of 4 days of prolonged stimulation per stimulation mode, it is possible that a steady state phase is not yet reached within one period of prolonged stimulation.

Due to the small sample size, coupled with the range of etiologies and degrees of residual function seen in patients, the outcomes of this study may not serve as proof for population-wide claims. This trial focuses on proof related to individual efficacy. In this way, a variety of patient characteristics (e.g. etiology, duration of vestibular function) will be investigated. However, due to the use of extensive and strict inclusion criteria, it is challenging to enroll participants out of the full spectrum of BV etiologies. Although the sample size is small, the crossover design and multiple follow-up visits facilitate extensive data collection from the limited number of participants. Also, this trial aims to apply insights at an individual level to a stimulation strategy on group level. More specifically: individual fitting data will be analyzed, fittings will be optimized, and implemented before prolonged stimulation of the whole sample is scheduled.

Another limitation in external validation of the results is the hospital setting of our experiments. Although the possibilities within the hospital are limited, an out-of-hospital activation pattern will be mimicked as much as possible. This will be realized by including exercise programs and by encouraging participants to move and behave ‘as normally as possible’. The exercise program included in every prolonged stimulation period aims to activate participants and prevent any movement avoidance during stimulation. By using sets of progressive exercises, based on the updated clinical practice guideline [89], a first experience in the possibilities of vestibular rehabilitation in a VCI-patient population will be gained. Multicenter trials with increased sample sizes and stimulation in an out-of-hospital context will be interesting follow-up projects, in order to gain more data about the functional impact on daily life.

3.4 Early health technology assessment and future product development

To inform all stakeholders about the potential value of the implant and steering the innovation of cochleovestibular stimulation in an early phase, an early health technology assessment is required [90]. The most relevant outcome measures to evaluate the spectrum of expected performance of the VCI, will be identified in this trial. Changes in participant-reported disability and quality of life have already been reported after prolonged stimulation with a VI-only device [28]. Based on these findings and other previously reported perspectives in a BV patient population [44], it is expected that anticipated value by patients plays a crucial role in the early assessment of the usability of the device. Further assessment of the impact of the new VCI technology on the individual patient in their home-situation, their environment and even society as a whole will be planned during follow-up studies. These studies might help to define candidacy criteria in terms of etiologies causing vestibulopathy and/ or hearing loss.