Although an initial report indicated that the firing rate of vestibular nucleus neurons is depressed for a prolonged period following a bilateral labyrinthectomy (Ryu and McCabe, 1976), more recent studies showed that spontaneous activity of vestibular nucleus units returns within hours following bilateral elimination of labyrinthine inputs, and is nearly identical to prelesion levels within less than a week (Waespe et al., 1992; Ris and Godaux, 1998; Miller et al., 2008). For example, Figure 2 compares the spontaneous activity and firing regularity (coefficient of variation, the SD of interval between spikes divided by mean interval between spikes) of neurons in the inferior and caudal medial vestibular nuclei of a conscious cat before and in the first week after a combined labyrinthectomy and vestibular neurectomy (Miller et al., 2008). The firing rates before and after elimination of vestibular inputs [30 ± 2 (SEM) vs. 32 ± 2 spikes/s, respectively], as well as the coefficient of variation of firing rates (0.78 ± 0.05 vs. 0.76 ± 0.05), were virtually identical in the two populations. Ris and Godaux (1998) conducted a longitudinal study of firing rates of vestibular nucleus neurons in conscious guinea pigs before and at 1 h, 1 day, and 1 week after a bilateral vestibular neurectomy. In control animals, no silent vestibular nucleus units could be detected; 53% of the cells were inactive at 1–5 h after a bilateral labyrinthectomy, and ∼35% were inactive at 1 day after lesions. By a week after elimination of labyrinthine signals, no silent neurons could be observed in the vestibular nuclei (Ris and Godaux, 1998).

Although vestibular nucleus neurons are insensitive to horizontal rotations following the removal of labyrinthine inputs (Ris and Godaux, 1998), the firing rates of some cells can be modulated by 15° tilts in vertical planes (Yates et al., 2000; Miller et al., 2008). In conscious cats, such response modulation was uncommon (7/168 neurons recorded in three animals; Miller et al., 2008); however, 18/67 neurons recorded from the vestibular nuclei of decerebrate cats that had undergone a combined bilateral labyrinthectomy and vestibular neurectomy over a month previously responded to vertical rotations (Yates et al., 2000). The response properties of vestibular nucleus neurons to vertical rotations in animals lacking labyrinthine inputs are illustrated in Figure 3. Figure 3A compares the vector orientations for responses to vertical tilts in animals lacking vestibular inputs to those observed in labyrinth-intact decerebrate and conscious animals tested using the same tilt table. In both decerebrate and conscious cats lacking vestibular inputs, the response vector orientations of most neurons were near the pitch plane (mean deviation from the pitch axis of 15°). In contrast, in labyrinth-intact cats, the response vector orientations were much nearer the roll axis; the mean vector deviation from the pitch axis was 50° in conscious animals (Miller et al., 2008) and 57° in decerebrate animals (Jian et al., 2002). The differences in response vector orientations between labyrinth-intact and labyrinthectomized animals were shown to be significantly different (p < 0.01) using a non-parametric one-way ANOVA (Kruskal–Wallis test). Figure 3B illustrates the dynamic properties of the responses of vestibular nucleus neurons to vertical tilts in labyrinthectomized animals. The response gain for units was relatively constant across stimulus frequencies, whereas the response phase was near stimulus position at low frequencies, and lagged position slightly at higher frequencies. Such properties are consistent with the responses being elicited by graviceptive inputs. In contrast, a large fraction of neurons in the vestibular nuclei of labyrinth-intact cats have responses to vertical tilts that are similar to those of semicircular canals: the response gain increases with advancing stimulus position and the response phase is near stimulus velocity (Jian et al., 2002; Miller et al., 2008). The marked differences in the responses to tilt of vestibular nucleus neurons in labyrinth-intact and labyrinthectomized animals show that the activity recorded in the latter group is not due to failure to eliminate inputs from the inner ear. Responses to vertical rotations were also recordable in bilaterally labyrinthectomized animals from additional regions of the central nervous system that receive vestibular inputs, particularly the cerebellar fastigial nucleus (Yates and Miller, 2009).

The rapid restoration of vestibular nucleus neuronal activity following a bilateral labyrinthectomy is likely due to an increase in the relative influence of non-labyrinthine excitatory inputs to the vestibular nuclei. The injection of retrogradely transported tracers into the inferior and caudal medial vestibular nuclei showed that this area receives direct inputs from several areas of the nervous system that process non-labyrinthine inputs, including the spinal gray matter, prepositus hypoglossi, pontomedullary reticular formation, inferior olivary nucleus, lateral reticular nucleus, medullary raphe nuclei, the spinal and principal trigeminal nuclei, and the facial nucleus (Jian et al., 2005). Other anatomical studies showed that the caudal regions of the inferior and medial vestibular nuclei receive direct inputs from primary afferent fibers entering the cervical spinal cord (Bankoul et al., 1995). In addition, neurons throughout the vestibular nucleus complex receive direct and polysynaptic inputs from cerebral cortex (Wilson et al., 1999). The firing rate of a majority of neurons in the medial, inferior, and lateral vestibular nuclei is affected by stimulation of somatosensory afferents from the limbs (Fredrickson et al., 1966; Wilson et al., 1966; Rubin et al., 1977; Jian et al., 2002); most neurons were excited by limb inputs. The excitability of some cells was also altered by activation of visceral afferents (Jian et al., 2002). However, the influences of limb afferents on vestibular nucleus neuronal activity are not ubiquitous. For example, few neurons that mediate vestibulo-ocular reflexes respond to stimulation of limb nerves (Rubin et al., 1978). Vestibulo-ocular neurons are concentrated in the rostral portion of the vestibular nucleus complex, in the superior and rostral medial and lateral nuclei (Graybiel and Hartwieg, 1974; Gacek, 1977, 1979a,b). Since afferents from the spinal cord only provide heavy inputs to the caudal regions of the vestibular nuclei (Rubertone and Haines, 1982; McKelvey-Briggs et al., 1989; Bankoul et al., 1995), the paucity of somatosensory influences on vestibulo-ocular units is not surprising. Thus, the non-labyrinthine inputs that are responsible for restoration of vestibular nucleus neuronal activity following bilateral damage to the inner ear may differ along a rostral–caudal gradient. For the caudal portions of the vestibular nuclei (caudal medial, inferior, and lateral nuclei), ascending inputs from the spinal cord could play a major role in regulating neuronal excitability. In contrast, descending inputs from cerebral cortex and thalamus could play a larger role in regulating the excitability of neurons in the rostral vestibular nuclei (superior and rostral medial nuclei).

As noted in Section “Eye Movements and Oscillopsia,” body rotations in the sagittal plane modulate the activity of some vestibular nucleus neurons in animals lacking labyrinthine inputs. Caudal vestibular nucleus neurons become more sensitive to somatosensory and visceral stimulation subsequent to a bilateral labyrinthectomy, suggesting that their postural-related responses after removal of vestibular inputs could be due to these inputs (Jian et al., 2002). This notion is supported by the observation that the responses were abolished by spinalization (Cotter et al., 2004). These findings suggest that non-labyrinthine sensory inputs generated by body movement, such as those related to stretch of muscles, brushing of the skin, or movement of the viscera, can elicit responses of vestibular nucleus neurons that reflect body position in space. As such, these findings raise the prospect that recovery of behavioral and physiological responses after BVH is due to substitution of non-labyrinthine for labyrinthine inputs in the central vestibular system.

Some deficits produced by bilateral damage to the inner ear in animals, such as loss of reflexive eye movements during head rotations (Baarsma and Collewijn, 1974; Barmack et al., 1980; Waespe and Wolfensberger, 1985; Waespe et al., 1992) and reduced navigational abilities (Baek et al., 2010), do not diminish over time. However, rapid recovery occurs for other consequences of bilateral loss of vestibular inputs, particularly impaired postural stability (Thomson et al., 1991) and inability to maintain stable blood pressure during postural alterations (Jian et al., 1999). It thus is useful to consider the similarities and differences between these responses to gain insights into the process of compensation. As illustrated in Figure 4, one difference between the deficits that dissipate and those that do not is the region of the vestibular nucleus complex where they are mediated. Vestibulo-ocular reflexes are mainly elicited by neurons in the rostral portion of the vestibular nucleus complex (Graybiel and Hartwieg, 1974; Gacek, 1977, 1979a,b), as are cognitive responses that are dependent on labyrinthine inputs (Brown et al., 2005; Shinder and Taube, 2010). In contrast, many of the neurons that control balance (Nyberg-Hansen and Mascitti, 1964; Petras, 1967; Peterson et al., 1978; Carleton and Carpenter, 1983; Carpenter, 1988) and influence blood pressure (Uchino et al., 1970; Yates et al., 1993; Kerman and Yates, 1998) are located caudally in the vestibular nucleus complex. In addition, the major components of vestibulo-ocular reflexes are dependent on inputs from semicircular canals (Money and Scott, 1962; Suzuki and Cohen, 1964, 1966; Baker et al., 1982; Hess et al., 2000; Sadeghi et al., 2009; Yakushin et al., 2011). Although data are limited, at least some cognitive responses related to vestibular inputs also appear to require inputs from semicircular canals (Muir et al., 2009). In contrast, while the properties of vestibulo-spinal reflexes are altered by canal plugging, which inactivates the semicircular canals but not the otolith organs, postural stability mainly requires otolithic inputs (Money and Scott, 1962; Watt, 1976; Schor and Miller, 1981). Similarly, vestibular system influences on the sympathetic nervous system are mainly related to signals from the otolith organs (Yates and Miller, 1994). Thus, following a bilateral labyrinthectomy, compensation occurs for responses that are elicited by otolith organ inputs processed by the caudal portion of the vestibular nucleus complex, but not for responses elicited predominantly by semicircular canal inputs that are processed by the superior and rostral medial vestibular nuclei. This dichotomy is at least partly related to the fact that otolithic inputs tend to target more caudal regions of the vestibular nuclear complex (Dickman and Angelaki, 2002; Newlands and Perachio, 2003; Newlands et al., 2003).

Since a variety of sensory systems provide graviceptive signals to the central nervous system (Mittelstaedt, 1992, 1995, 1996; Mittelstaedt and Mittelstaedt, 1996; Balaban and Yates, 2004), it is not surprising that these inputs can be substituted at least partially for those from the otolith organs. Some of these graviceptive inputs are conveyed to the vestibular nuclei (see “Cognitive deficits”), and are likely responsible for the modulation of vestibular nucleus neuronal activity that can be detected during body rotations in the pitch plane conducted following a bilateral vestibular neurectomy (Yates et al., 2000; Miller et al., 2008). Bilateral lesions placed in the caudal region of the vestibular nuclei result in a permanent deficit in cardiovascular system responses during postural alterations (Mori et al., 2005), as well as deficits in postural control that are more severe and long lasting than those resulting from a bilateral labyrinthectomy (unpublished observations). These findings suggest that the vestibular nuclei participate in the “sensory substation” that facilitates recovery of postural stability and movement-related cardiovascular adjustments following the loss of labyrinthine inputs. However, the specific role that the central vestibular system plays in the compensation process is yet to be determined.

Humans with BVH are able to maintain normal stance in the light, although when the support surface they stand upon moves, falls are much more likely to occur (Nashner et al., 1982; Mergner et al., 2009). When the support surface is systematically tilted in the eyes closed condition, BVH subjects either sway with the platform movement (at lower peak tilt angular velocity) or fall (at higher peak tilt angular velocity; Maurer et al., 2006). Tandem Romberg stance (one foot in front of the other) is a challenging posture to maintain: control subjects without assistance are able to maintain the posture in light or dark, but subjects with BVH fall in both conditions (Lackner et al., 1999). The elevated fall risk presented by unstable support surfaces and challenging postural positions translates to increased risk of fall and injury in the daily lives of BVH patients.

Patterning of postural muscular responses following perturbations is altered in BVH patients, but the origin and organization of the muscular responses that occur in this patient population is debatable. It has been proposed that balance-correcting responses are elicited by somatosensory signals from the trunk or about the hip joint, which are then modulated by inputs from the ankles, knees, and the vestibular system (Horstmann and Dietz, 1990; Allum and Honegger, 1998). Other investigators have proposed that somatosensory signals originating about the ankle may be the primary drivers of postural responses (Nashner et al., 1982; Horak and Nashner, 1986; Horak et al., 1990). To add to the complexity of deciphering muscular control of posture, direction of perturbation is an important factor in maintenance of posture. Humans are more unstable when subjected to perturbations in the roll plane then when they are tilted in the pitch plane (Carpenter et al., 2001). Hip and trunk muscles are activated during roll motion, while pitch perturbations mainly involve an activation of leg muscles (Henry et al., 1998; Gruneberg et al., 2005; Allum et al., 2008). BVH subjects have been shown to demonstrate increased paraspinal muscular activity in response to toes up pitch rotations (Allum et al., 2001). The normal lower body response to roll tilt of a support surface also involves uphill leg flexion and downhill leg extension; BVL subjects have a reduction of this compensatory leg motion (Allum et al., 2008). Quasi-static roll perturbations, which are believed to be detected mainly via proprioceptors rather than otolithic graviceptors, are perceived at low thresholds in both normal subjects and BVH patients suggesting that proprioceptive inputs are satisfactory for maintenance up upright posture in relatively adynamic conditions (Teasdale et al., 1999; Bringoux et al., 2002). The relative role such proprioceptive inputs play in more dynamic conditions, such as during ambulation, remains undefined.

Dependence on other sensory systems, namely proprioceptive and visual, is widely regarded as the means that permits BVH patients to maintain upright posture. Some data suggest that a transition occurs months to years following the onset of BVH from a visually dominant postural control paradigm to a more proprioceptive guided paradigm, although longitudinal studies within individuals are lacking (Bles et al., 1983). Modeling of human posture indicates that normal subjects are able to reweigh sensory data from visual, proprioceptive, and vestibular inputs in response to changes in stimulus amplitude (i.e., non-linear changes in gain and phase occur in response to changes in amplitude of stimuli; Peterka, 2002). Subjects with BVH are not able to reweigh sensory information in this fashion. Some BVH subjects utilize a balance strategy dependent upon an increase in stiffness, which results in decreased sensory error, but necessitates a higher energy expenditure to maintain (Peterka, 2002). Subjective visual vertical (SVV) estimates during rotation of the visual field or placement of the body in the lateral position are more deviated in BVH patients compared with normal subjects, suggesting that vestibular inputs normally serve to influence or supersede less accurate proprioceptive or visual postural information (Bronstein et al., 1996). Alternatively, exaggeration of error in SVV when BVH subjects are tilted onto their sides could suggest that signals from proprioceptive pathways are enhanced following vestibular loss. As balance control strategies are not nearly as critical for preventing falls in the supine position, errors in estimated verticality in this position are less clinically relevant than if such errors were to occur with upright stance. Enhancement or disinhibition of proprioceptive pathways may be important to maintenance of upright posture in BVH. For example, the enhanced pattern of paraspinal muscular activity in toes up pitch rotations is opposite to that seen in patients with proprioceptive loss, suggesting that proprioceptive responses may be enhanced in BVH (Allum et al., 2001). Alternative sensory input in the form of putative trunk graviceptors may be another important factor in postural control following loss of vestibular input (Mittelstaedt, 1998).

Humans with BVH exhibit eye movement abnormalities. BVH results in markedly diminished or absent vestibulo-ocular reflexes (VOR) bilaterally, such that measurement of the VOR during clinical vestibular testing is used to diagnose BVH (Brown et al., 2001; Goebel et al., 2009; Jen, 2009). Responses to head thrust testing, which assesses the high-frequency VOR, are also abnormal in BVH (Halmagyi and Curthoys, 1988; Jorns-Haderli et al., 2007; Zingler et al., 2008). Over time, there does not appear to be recovery of VOR function regardless of the etiology of BVH (Baloh et al., 2001; Zingler et al., 2008). Optokinetic after nystagmus is also severely reduced or abolished in BVH (Zee et al., 1976; Ireland and Jell, 1982; Bles et al., 1984; Hain and Zee, 1991). Most groups have found smooth pursuit to be undisturbed in BVH (Kasai and Zee, 1978; Leigh et al., 1987; Waterston et al., 1992), although a slight potentiation of smooth pursuit in BVH has been reported (Bockisch et al., 2004).

Since the VOR is believed to stabilize vision during head motion, its absence in BVH requires compensatory mechanisms to optimize visual input. Mechanisms postulated to participate in this compensatory process include enhancement of the cervico-ocular reflex, central preprogramming of eye movements, and increase in the optokinetic response (Kasai and Zee, 1978; Bronstein and Hood, 1986; Huygen et al., 1989, 1991; Waterston et al., 1992). The cervico-ocular reflex, which is triggered by inputs from neck muscles, is normally of small magnitude in humans, but its gain increases following bilateral vestibular loss particularly when the head is moved at low frequency (∼0.1 Hz; Bronstein and Hood, 1986; Huygen et al., 1991). Central preprogramming of eye movements necessitates that the subject is aware of where the next target will be located or what head movement will take place; in this situation, subjects with BVH more accurately identify the visual target (Herdman et al., 2001). However, many head movements and choices for gaze direction in daily life are not predictable and this strategy would not be useful in many circumstances. Efference copy of motor commands may also facilitate accurate targeting during gaze shifts (Maurer et al., 1998), but it is unclear whether this mechanism provides for compensation in BVH patients. In some low frequency head movement circumstances, optokinetic responses can also contribute to compensatory eye movements in BVH patients (Leigh et al., 1992).

It is known that the generation of compensatory eye movements in patients with BVH requires a functioning cerebellum (Bronstein et al., 1991; Waterston et al., 1992). In the converse situation, individuals with cerebellar deficiency and intact vestibular systems have an elevated VOR gain, show an inappropriate direction of the slow phase of eye movements orthogonal to the axis of rotation, and display abnormalities during cervical vestibulo-ocular myogenic potential testing thought to be analogous to an increase in vestibulocollic reflex gain (Shaikh et al., 2011). Therefore, the cerebellum may be responsible for potentiating a number of pathways for compensation in BVH.

Oscillopsia is the perception of blurring or movement of the environment when the head is placed in motion during activities such as turning the head to the side or ambulating (Schubert et al., 2004). Oscillopsia is a common symptom in BVH, but it is not ubiquitous (McGath et al., 1989; Sargent et al., 1997). Oscillopsia is thought to occur in BVH because slip of visual targets across the retina occurs in the absence of a functioning corrective VOR. Dynamic visual acuity during ambulation is reduced in BVH, which is likely related to retinal slip (Lambert et al., 2010). Tolerance to retinal slip, rather than corrective eye movements, appears to be important for adaptation to oscillopsia. Subjective reporting of the degree of oscillopsia is negatively correlated with the amount of retinal slip (Grunfeld et al., 2000). Similarly, visual motion detection thresholds are reduced in BVH patients compared with controls (Shallo-Hoffmann and Bronstein, 2003). It is likely that adaptive compensatory mechanisms are responsible for the reduced sensitivity to visual field motion, although the pathways involved remain to be elucidated.

Patients with vestibular loss are often noted to suffer cognitive deficits such as difficulty concentrating or being in a “brain fog,” and patients with BVH are likely not an exception (Hanes and McCollum, 2006). Although cognitive deficits of this nature are sometimes difficult to study, several important cognitive effects have been demonstrated in humans with BVH, chiefly impaired spatial learning and memory and, less directly, impairment in dual tasking.

Humans with bilateral vestibular loss have impaired spatial learning and spatial memory deficits. BVH subjects had poor performance on the virtual Morris water task, which tests spatial learning and memory in the absence of somatosensory or vestibular cues (Schautzer et al., 2003). Anatomic and physiologic data support these findings and indicate that the hippocampus, a brain region known to be involved in spatial learning and memory, is responsible for the deficiencies (Brandt et al., 2005; Jahn et al., 2009; Smith et al., 2010b; Viard et al., 2011). MRI testing showed that BVH subjects have a 17% reduction in hippocampal volume (Brandt et al., 2005), and functional MRI studies demonstrated that these patients also have decreased activity within the anterior hippocampus (Jahn et al., 2009). Memory and navigation deficits in BVH patients were limited to spatial problems; general memory was found to be intact (Brandt et al., 2005). Although there is some suggestion that the right cerebral cortex is dominant for spatial orientation, right unilateral vestibular loss subjects only displayed subtle differences in task performance compared with subjects with left unilateral loss, and there is no hippocampal atrophy in unilateral vestibular loss subjects (Hufner et al., 2007; Jahn et al., 2009). Taken together, these data suggest that spatial learning and memory are dependent upon the presence of function of at least one labyrinth; without any functional labyrinthine input, deficits of spatial learning and memory arise.

Patients with vestibular loss have also been shown to have difficulty with dual mental tasks, such as performing arithmetic along with orienting or balance tasks (Yardley et al., 2001, 2002). These dual task studies mostly included patients with unilateral vestibular loss; a few patients with BVH were also studied but were not segregated, such that drawing specific conclusions on dual tasking in BVH is not currently possible. As such, further study of multitask performance is warranted in BVH patients.

Spontaneous recovery of vestibular function rarely occurs in BVH patients (Brandt et al., 2010); treatment strategies therefore necessarily rely upon coping with existing vestibular function (in the case of partial loss) or providing alternative means to present information normally sensed by the vestibular system to the brain (Minor, 1998). The mainstay of treatment for BVH is vestibular physical therapy. Treatment in an active vestibular rehabilitation program was shown in a double blind placebo-controlled study to improve dynamic gait stability considerably among patients with BVH (Krebs et al., 1993). Similarly, dynamic visual acuity has been shown to improve with vestibular physical therapy in BVH (Herdman et al., 2007). Such improvements should translate readily to improvements in quality of life amongst treatment responders.

Improvement with vestibular physical therapy has been shown to occur in approximately 50% of patients with BVH; however, this leaves a substantial population that does not benefit (Shepard and Telian, 1995; Gillespie and Minor, 1999; Brown et al., 2001). Risk factors for poor response to vestibular physical therapy include multiple medical comorbidities, slowly progressive bilateral vestibular loss, and increased severity of BVH (Gillespie and Minor, 1999). A further understanding of the compensatory process in BVH may lead to explanations about the variability of recovery in BVH and suggest new treatment strategies.

Investigators continue to search for new ways to refine balance function in patients suffering from BVH. Sensory substitution, converting information normally encoded by one system (in this case, the vestibular system) into information that can be detected with another sensory modality (such as visual cues or sensory tactile information), has been one approach taken to improve stance and gait in BVH patients. Sensory substitution in BVH is a pragmatic approach because evidence suggests that patients with BVH already may use a form of it – visual sensory substitution. While undergoing optokinetic stimulation, fMRI data from BVH patients demonstrates increased activity within visual and oculomotor areas in comparison with normal subjects, suggesting that BVH individuals use visual flow to assist with balance (Dieterich et al., 2007). Furthermore, BVH subjects have been shown to have better postural stability when permitted to use visual cues (Buchanan and Horak, 2002; Horak, 2010). Non-supportive light touch can also substitute as an earth vertical reference point and stabilize posture (Creath et al., 2002, 2008). Auditory biofeedback has also been used as a form of vestibular sensory substitution by notifying patients about the degree of postural sway through auditory cues (Dozza et al., 2007, 2011).

Several devices are under investigation to assist in sensory substitution for BVH. The most studied sensory substitution device delivers body tilt information through a vibrotactile somatosensory feedback system attached to the torso that has been shown to reduce postural sway and decrease falls (Kentala et al., 2003; Wall and Kentala, 2005). Another sensory substitution system involves wearing a head mounted vibrotactile device. This system has undergone phase 1 clinical testing and was shown to significantly reduce instability in BVH patients as measured by dynamic posturography (Goebel et al., 2009). Another device provides electrotactile feedback to the tongue based upon head position relative to gravity (Barros et al., 2010).

An alternative approach to sensory substitution based prosthetics involves direct electrical stimulation of vestibular sensory nerves with an implanted vestibular prosthetic device. Single channel and multichannel head mounted devices are currently under study in animals (Della Santina et al., 2005, 2007; Merfeld et al., 2007; Lewis et al., 2010; Dai et al., 2011). The foundation of modern vestibular prosthesis work comes from experiments in which precise eye movements were elicited through electrical stimulation of branches of the vestibular nerve (Suzuki et al., 1964, 1969). Current vestibular prostheses typically include electrodes within the semicircular canals that deliver electrical stimulation to individual ampullary nerves. The electrical stimuli are controlled by measurements from accelerometers that detect head rotation. An underlying premise for this therapy is that the animal or human with the implant will adapt to the firing rate of ampullary nerves induced by the device, and use this information to elicit appropriate compensatory eye movements. In guinea pigs, this acclimation process was shown to occur rapidly. Importantly, acclimation to chronic baseline stimulation did not interfere with generation of eye movements when electrical stimulation was used to modify the activity of particular vestibular nerves (Merfeld et al., 2006). Longer-term adaptation studies in the squirrel monkey showed that an electrically evoked VOR is synchronized with head motion after 3 months of chronic stimulation (Merfeld et al., 2007). One important concern regarding electrical stimulation is the problem of current spread from the intended ampullary nerve to other nearby ampullary nerves, resulting in cross stimulation. VOR adaptation studies have shown that over the first week following prosthesis activation in chinchillas, erroneous VOR eye movements caused by current spread rapidly diminish, such that eye movements more closely approximate those expected based upon head motion (Dai et al., 2011). To date, one human has been implanted with an electrically based vestibular prosthesis. The surgery involved placing a single electrode from a modified cochlear implant adjacent to the posterior ampullary nerve in a patient with BVH and bilateral deafness. Experiments using the device demonstrated that nystagmus diminished over time, and that eye movements were elicited in accordance with the amplitude and frequency of stimulation (Guyot et al., 2011).

Since we are at the forefront of investigation of vestibular prostheses in humans, many questions remain. It is unclear whether such prostheses will stabilize oscillopsia by accurately encoding head motion in space. Similarly, it remains to be reported whether electrical stimulation of ampullary nerves improves balance control in patients or animal subjects with BVH. Galvanic or caloric stimulation of the vestibular system of normal humans has been shown to improve memory (spatial, verbal, and facial recall; Bachtold et al., 2001; Wilkinson et al., 2008; Smith et al., 2010a). Whether electrical vestibular stimulation will affect cognitive processing in BVH patients remains unknown. It is also unclear whether vestibular prostheses will alter the sensory substitution that ordinarily occurs in BVH patients, and whether sensory substitution and electrical stimulation devices can be effectively used in combination in particular subjects.