Galvanic Vestibular Stimulation Advancements for Spatial Disorientation Training

Vestibular-Mediated Modes of Spatial Disorientation

SD often arises from both inadequacies of the vestibular organs to transmit accurate spatial orientation information in flight and the subsequent failure of the central nervous system (CNS) to correctly interpret this traduced information.¹¹ Because GVS offers the potential to enhance SD training by augmenting the vestibular sense, we focus herein on a particular subset of SD, often referred to as “vestibular” illusions due to their onsets. To understand how GVS may be used to produce unreliable vestibular cues during training, we first examine vestibular-mediated mechanisms impacting (mis)perceptions.

Producing gyroscopic angular velocity cues which are sent to the CNS, the semicircular canals have dynamical endolymph-cupula system characteristics that result in decaying signal transduction at constant angular velocities because the canals are physically stimulated through angular accelerations.¹² This characteristic of the canals means that for a rotating motion, starting at zero and approaching a constant angular rotation rate, the canals will signal a rotation rate, carried by the afferent neuron fibers, that will eventually decay to zero during constant rotation. Upon stopping this rotation, the canals will be restimulated and produce a signal transducing angular rotation in the opposite direction of the original motion profile that eventually decays to zero. Without other, reliable sensory cues (e.g., visual, auditory, and somatosensory cues), human perception will follow the sensory dynamics of the canals. However, perception decay is delayed more than the afferent signal,¹³ characterized by a larger time lag and coined “velocity storage” (i.e., perception decays with a time constant between 14 and 35 s,¹⁴ whereas the afferent signal decays at a time constant of around 6 s¹²). Furthermore, the semicircular canals have high-pass filter characteristics, resulting in decreased stimulation from low-frequency (< ∼0.1 Hz) angular rotations.¹² These physiological factors can lead to angular rotation misperceptions, commonly referred to as somatogyral illusions.

The otolith organs transmit a signal corresponding to the net accelerations experienced in terms of the net gravito-inertial force (GIF). Due to Einstein’s equivalence principle, inertial sensors measuring acceleration cannot on their own disambiguate linear acceleration from gravity, thus the otoliths are too restricted in this manner. To determine what is gravity and what is linear acceleration (without other, reliable sensory cues), it is currently believed that the CNS relies on internal models¹⁵ that perform well when the semicircular canals are stimulated (at motion frequencies > ∼0.1 Hz). However, in conjunction with the covered inadequacies of the canals, humans can become disoriented when relying on central processing of the otolithic sensory signals. The subset of vestibular illusions resulting from a failure to disambiguate GIF cues are often called somatogravic illusions.

Beyond these mechanical and internal model sources of disorientation, sensory noise also contributes to decreased human perception precision of self-motion.¹⁶ Coined vestibular thresholds, vestibular perceptual precision defines the stimulus magnitude that can be reliably perceived by individuals. For angular rotations where the net GIF stimulus is not changing in a head-centric reference frame, rotational thresholds have been quantified by peak angular velocity for a motion duration.^17–19 Further, static tilt thresholds in roll and pitch (where the GIF stimulus is changing in a head-centric reference frame) have been quantified by the peak tilt angle of a motion.^20–22 Crucially for SD, there exist small passive motion stimuli (subthreshold) that go unperceived by humans. Considering all of these combined factors, we can begin to understand why a host of vestibular illusions occur in human operators of aircraft. We summarize relevant illusions in Table I below. In addition to these well-defined illusions, there is also potential for aviators to experience SD due to unexpected vestibular pathophysiology²³ requiring piloting through vertigo.

Table I.Overview of Vestibular Illusions Relevant for GVS-Enhanced SD Training.

View Fullscreen

The Current State of SD Training in the U.S. Military

Current SD training for U.S. military aircrew consists of regularly occurring training on topics such as human sensory systems (capabilities and limitations), environmental conditions that may induce SD, and widely accepted theories on the types of SD. However, most of this training occurs in a classroom setting and experiential training is infrequent, with the most common implementation being a static flight simulator. Inadequate at prepping pilots for vestibular illusions, static flight simulators offer no vestibular stimulation and often rely solely on visual and audible inputs when simulating a disorienting scenario, thus omitting the source of vestibular mediated SD. Additionally, there is no training requirement for aircrew to experience visual illusions in the static simulator. Instructors may elect to demonstrate a visual illusion during an instrument check in the simulator, but this would be an elective training opportunity and would be secondary in nature as the purpose of the training is to demonstrate proficiency with instrument flying, with no qualifier for disorientation of any kind. In these limited scenarios, aircrew are afforded the opportunity to demonstrate emergency procedures to recover from an unusual aircraft attitude or fly with a partial-panel configuration. In the case of the U.S. Air Force, time in a spatial disorientation trainer (GYRO IPT) is required during initial physiology training only,²⁴ and recurrent requirements consist only of encouraged spatial disorientation presentations by an aerospace physiologist or flight surgeon.²⁵

Galvanic Vestibular Stimulation

GVS encompasses methodically augmenting the vestibular sense through transcutaneous electrical stimulation. This form of electrical stimulation has been proven to be safe at low current levels (<5 mA) which are orders of magnitude lower than safety limits defined for transcranial direct current stimulation in general, which result in no long-term risks.²⁶ Acute side effects do exist for GVS and are mild in their severity, described as tingling and discomfort around the electrode sites.²⁷ Electrode gel between the skin and electrodes helps reduce these irritations and has been used for decades in GVS studies.²⁸ However, electrode gel may pose a risk to individuals with paraben (i.e., a class of chemical often used as preservatives) allergies. Furthermore, the GVS stimulus intensity (commonly described by current amplitude) can be titrated based on individual comfort, alleviating concerns of irritation and discomfort. Less commonly noted in the literature, metallic tastes and light flashes have been reported by subjects during GVS,²⁹ and motion sickness symptoms have been observed when using higher peak amplitudes (above 3.5 mA) following 20 min of continuous exposure to an unpredictable GVS waveform profile intended to disrupt subjects.³⁰ Similarly, transient sensations such as light flashes have been observed during transcranial direct current stimulation, dependent on electrode placements and current profiles.³¹

Further encouraging the adoption of GVS as a well-studied means of providing vestibular stimulation, the specific physiological response of GVS has been confirmed to target the vestibular afferent neurons in invasive nonhuman primate studies.^32,33 Congruent with most GVS research, these studies used GVS in the binaural bipolar configuration; a single electrode is placed behind each mastoid process (a total of two electrodes), providing equal and opposite current. In this configuration, GVS affects the firing rates of the afferent vestibular neurons within cranial nerve VIII, the vestibulocochlear nerve.³³ Directly downstream of this nerve, the bilateral vestibular nuclei process signaled motion and direct these cues to spinal, cerebellar, ocular, and cortical pathways in the brain.³⁴ As a result, illusory sensations of self-motion are elicited by GVS stimuli, originating by stimuli at the peripheral organ afferent innervations. The influence of GVS on the CNS is depicted in Fig. 2. Quantifying the effect of direct current GVS in stationary, upright seated humans, reversal of current polarity has been found to elicit a directionally opposite perception of tilt.³⁵ In addition, temporally dynamic GVS waveforms have been found to elicit Earth-vertical rotation perceptions when humans are pitched at varying angles: pitched eyes down at around 70°³⁶ and pitched eyes up at 90°³⁷.

View Figure

• Download as Powerpoint
• Download Figure

Moreover, Kwan et al.³³ and Forbes et al.³² established the dynamic relationships between electrical stimulus (in terms of current; mA) and afferent neuron firing rates (of both irregular and regular afferent neurons). We now have a more complete understanding and quantification of how the afferents near the cathode terminal (negative) and anode terminal (positive) differ in depolarization (excitement) and hyperpolarization (inhibition), respectively, for both the semicircular canal and otolith organs, and the gain and phase responses have been established through transfer functions.

GVS as an Impairment Paradigm

For nearly 20 yr, GVS has been used as a tool for providing a nonspecific mode of vestibular-dominated sensorimotor impairment in various research studies. In the binaural bipolar configuration, a ‘pseudorandom’ (sum-of-sines) waveform has been shown to induce acute vestibular impairment, resulting in impaired postural control, examined via a computerized dynamic posturography sensory organization task protocol.^30,38 Furthermore, simulated shuttle operator proficiency was worsened in the presence of this GVS waveform.³⁹ In these studies, the waveform profile consisted of low frequencies (<1 Hz) with a peak current magnitude of up to 5 mA. Because the bipolar binaural electrode montage was used, perceptual and postural disturbances resulting from this waveform are thought to evoke sensations and perturbations of mostly roll.⁴⁰ This effect is largely thought to be a result of the net virtual stimulus evoked by GVS, particularly the CNS interpretation of the signals transmitted by the semicircular canal afferents.

While this mode of vestibular stimulation is limited in that it does not reproduce illusions that may be experienced in flight, simulator training with this form of GVS may prove to be a means of decreasing pilot reliance on unreliable vestibular-mediated perceptions. Recently, persistent unreliable vestibular information during spaceflight has been found to trigger compensatory sensory reweighting, with a higher activation of visual and somatosensory cortical regions during postflight vestibular stimulation.⁴¹ We propose using this form of GVS as a basic enhancement of SD training for military aviators in a static environment. Providing unreliable, nonspecific vestibular sensory stimulation for brief periods should be studied as a potential solution for deceasing SD mishap occurrences in flight as a supplement to current SD training. Furthermore, this application of GVS can be achieved with a low-cost, minimal-equipment setup.

GVS for Providing Specific Vestibular-Mediated Perceptions

Beyond the nonspecific vestibular disturbances supplied by pseudorandom, sum-of-sines GVS waveforms comprised of low frequencies, GVS may be used in a more specific manner, providing replicate vestibular illusions in brief intervals. Here we explore how GVS can be used to produce higher fidelity flight simulation not possible in a static environment by considering what is known about how GVS affects the semicircular canal afferents. By providing physically congruent rotational cues to the semicircular canals via GVS, a more immersive SD training environment may be achieved for practicing SD illusion recovery in a nonoperational setting. The current state of GVS promises the capability of imitating specific vestibular illusions via GVS.

First, we consider the well-studied binaural bipolar electrode montage. In this electrode montage, GVS nonpreferentially stimulates all three semicircular canal afferents (i.e., the horizontal, posterior, anterior canal afferents). While the extent of depolarization and hyperpolarization likely differs between sets (left and right) of organs in humans as it does in nonhuman primates,³² it is believed that the afferents innervating the anterior, posterior, and lateral canals are all excited/inhibited by GVS the same for a given side, as has been found in a subset of cranial nerve VIII afferents empirically evaluated in nonhuman primates.³³ Thus, the net rotation rate stimulus direction produced by the canals can be calculated, dependent only on the anatomical orientations of the canals, which have been recorded in cadavers.⁴² This net stimulus mechanism is qualitatively depicted in Fig. 3.

View Figure

• Download as Powerpoint
• Download Figure

Because canals operate in opposite pairs during physical motion (e.g., positive yaw results in depolarization of the left lateral canal and hyperpolarization of the right lateral canal), depolarizing and hyperpolarizing opposite sets of canals via GVS results in a neuronal response that is congruent with physical motion. Unlike the canals, the otolith organs do not exhibit this same duality, and the net stimulus of linear acceleration produced by each utricle and saccule individually is hypothesized to be near zero, regardless of the GVS current amplitude or electrode montage.⁴⁰ This hypothesized net-zero effect has been estimated from the arrangement of hair cells innervating the otolithic maculae about the striola. During physical linear accelerations, subregions of the hair cells innervating individual otoliths produce hyperpolarization and other subregions produce depolarization (dependent on regional stereocilia orientation to the linear acceleration stimulus). Alternatively, the entire set of afferents innervating a single otolith are estimated to be roughly equally stimulated during GVS, which may result in a near zero change in signaled GIF.⁴⁰ This estimation is independent of the empirically observed substantial activations of individual otolith afferents noted in monkey models.^32,33

Given that binaural bipolar GVS evokes a sensation mostly just of rolling, and likely does little to signal virtual GIF changes, the current state of the art in GVS using this electrode may appear limited for evoking relevant in-flight illusions, as it is only able to evoke illusions that result from the canals alone (such as the graveyard spin, Gillingham postroll, and the Coriolis illusion; see Table I). However, this electrode montage should be capable of providing additional in-flight vestibular illusion perceptions relating to misperceptions of tilt as well, additionally associated with the otolithic origins (such as the leans, false roll angle, G-excess, and the graveyard spiral). This is because the misperceptions resulting from the illusions (see Table I) are desired, not merely the physiological afferent responses. The goal of producing these misperceptions, through both physical and GVS stimuli, is demonstrated in Fig. 4.

View Figure

• Download as Powerpoint
• Download Figure

Computational predictions of spatial orientation perception can be made by using the well-validated “observer model” of spatial orientation perception^43–45 in combination with the afferent responses evoked by GVS and the anatomical orientations of the semicircular canals. While many unknowns still exist that preclude the validated use of such a model (such as how the regular and irregular afferent responses evoked by GVS are combined, how physical motion afferents are combined with GVS afferents, and how nonhuman primate models are related to humans), we can roughly predict how binaural bipolar GVS waveforms result in (mis)perceptions or self-orientation and self-motion, also dependent on the operator’s static sitting orientation in respect to gravity. The following assumptions are first made. We assume that the semicircular canal response is fully dependent on GVS with no contribution from physical motion, the irregular (or regular) afferents only need be modeled, and the otolith response is unaffected by GVS, dependent on actual gravity.

With these assumptions, direct current GVS in the binaural bipolar configuration should result in an illusory, persistent static tilt perception toward the cathode side when sitting upright in respect to gravity (as captured empirically by Niehof et al.³⁵). This effect may be used to replicate the leans, false roll angle, G-excess, and graveyard spiral illusions (Table I) in a static simulator environment. Further, if a subject is seated so that the net canal GVS stimulus vector is parallel to gravity (roughly a 17–19° head-down tilt from Reid’s base line⁴⁰), the observer model predicts that this electrode setup should be additionally capable of consistently producing roll angular rotation perceptions. Prior experiments have found that this is indeed the case.^36,37 Such a stimulus may be useful for replicating the Gillingham postroll illusion in a static simulator. Furthermore, virtual training environments without horizon cues, such as interior cockpit views, can be used in simulator training to further enhance the training environment. In this case, computational self-orientation predictions can be made using an observer model if additional visual sensory pathways are included.^43,46

Beyond the sensations of roll tilt and rolling produced via the binaural bipolar electrode montage, efforts have emerged to elicit rotational sensations about the yaw and pitch axes through other electrode montages. Notable are the Cevette et al.⁴⁷ and Aoyama et al.⁴⁸ montages, consisting of four electrodes each. A recent effort using the Cevette montage demonstrated the capability of providing persistent pitch tilt illusions in a static simulator environment through the additional pitch capability, and also used the multi-axis stimulation capability to provide disorienting multi-axis rotation perceptions, mimicking both the false pitch angle and Coriolis illusions, respectively.⁴⁹

Additional Uses of GVS for Preventing Spatial Disorientation

Beyond use as an SD training device, GVS has the promise of providing in-flight mitigation strategies during human operation of aerospace vehicles. While not the main focus of this review, these additional applications are noteworthy for their potential at reducing the occurrence of SD events in flight.

Limitations

Despite the promise of using GVS for enhancing SD training, there are a number of notable limitations to the current state of the art. If donned daily, GVS may eventually result in operator habituation, as studied by Dilda et al.,³⁰ when examining postural balance in the presence of a pseudorandom GVS waveform. Therefore, the specific training regimen frequency using GVS may need to be titrated; however, this limitation may also be indicative of a learning to be less affected by misleading vestibular cues, a key goal in preventing vestibular-mediated SD mishaps.

Another key limitation involves the uncertainty surrounding the four-pole electrode montages, such as those used by Cevette et al.⁴⁷ and Aoyama et al.⁴⁸ Unlike the binaural bipolar montage, the dynamical response of vestibular afferents in these electrode montages have yet to be studied, and the illusory sensory signals evoked by current when using these configurations are still mostly theoretical. Despite this limitation, a number of human studies (examining perceptions^47,49,56 and sway⁴⁸) have been conducted with these montages, which may be used for guidance. Additionally, it is unclear how Cevette, Pradhan, and colleagues are achieving specific perceptions from applied current, and future research studies should explore modeling their proposed approaches, including quantifying potential unwanted side effects across different GVS amplitudes and frequencies. Currently, no comprehensive, validated model of spatial orientation perception with GVS as an input exists, and future work should explore building and validating such a model. Once complete, specific perceptions resulting from GVS may be intelligently designed for use in SD training.

Conclusions

Current SD training for military aircrew consists of regularly occurring didactic training on topics such as human sensory systems (capabilities and limitations), environmental conditions that may induce SD, and widely accepted theory on the types of SD. However, experiential training is infrequent, with the most common device being a static flight simulator. Static flight simulators offer no vestibular stimulation to the brain of the flight crew, often relying solely on visual and audible inputs only when simulating a disorienting scenario, omitting the source of vestibular mediated SD. Given persistent SD mishaps across military aviation, we explore how GVS may be used to potentially reduce these occurrences by enhancing static flight simulator training for SD.

GVS offers an opportunity to create a perception of disorientation by providing supra-threshold bilateral bipolar electrical current to the vestibular system in a safe, reliable, low-cost manner. GVS has been used on humans for decades in the laboratory environment to research topics such as balance, neurological functions of healthy and diseased populations, detrained vestibular systems in spaceflight crew, and most recently in virtual reality scenarios to simulate a variety of commonly experienced somatogravic flight illusions such as Coriolis cross-coupling. We discuss using both nonspecific (disruptive) and specific (mimicking illusions) GVS waveforms to improve SD training, resulting in reliable training profiles that could be worn by military aircrew in a static flight simulator.