Skip to main content
The Journal of Physiology logoLink to The Journal of Physiology
. 2010 Oct 11;589(Pt 4):843–853. doi: 10.1113/jphysiol.2010.197053

Adaptation of vestibular signals for self-motion perception

Rebecca J St George 1, Brian L Day 2, Richard C Fitzpatrick 1
PMCID: PMC3060364  PMID: 20937715

Abstract

A fundamental concern of the brain is to establish the spatial relationship between self and the world to allow purposeful action. Response adaptation to unvarying sensory stimuli is a common feature of neural processing, both peripherally and centrally. For the semicircular canals, peripheral adaptation of the canal-cupula system to constant angular-velocity stimuli dominates the picture and masks central adaptation. Here we ask whether galvanic vestibular stimulation circumvents peripheral adaptation and, if so, does it reveal central adaptive processes. Transmastoidal bipolar galvanic stimulation and platform rotation (20 deg s−1) were applied separately and held constant for 2 min while perceived rotation was measured by verbal report. During real rotation, the perception of turn decayed from the onset of constant velocity with a mean time constant of 15.8 s. During galvanic-evoked virtual rotation, the perception of rotation initially rose but then declined towards zero over a period of ∼100 s. For both stimuli, oppositely directed perceptions of similar amplitude were reported when stimulation ceased indicating signal adaptation at some level. From these data the time constants of three independent processes were estimated: (i) the peripheral canal-cupula adaptation with time constant 7.3 s, (ii) the central ‘velocity-storage’ process that extends the afferent signal with time constant 7.7 s, and (iii) a long-term adaptation with time constant 75.9 s. The first two agree with previous data based on constant-velocity stimuli. The third component decayed with the profile of a real constant angular acceleration stimulus, showing that the galvanic stimulus signal bypasses the peripheral transformation so that the brainstem sees the galvanic signal as angular acceleration. An adaptive process involving both peripheral and central processes is indicated. Signals evoked by most natural movements will decay peripherally before adaptation can exert an appreciable effect, making a specific vestibular behavioural role unlikely. This adaptation appears to be a general property of the internal coding of self-motion that receives information from multiple sensory sources and filters out the unvarying components regardless of their origin. In this instance of a pure vestibular sensation, it defines the afferent signal that represents the stationary or zero-rotation state.


Non-technical summary

The semicircular canals of the inner ear provide the brain with a sense of self-rotation and orientation. How they maintain a stable sense over time is not understood, as the signals from the canals have no fixed reference for rotation speed. We show here by stimulating the vestibular nerves electrically and by real rotation that a central adaptive process automatically zeros the vestibular signal to define the signal that represents zero rotation. In doing this, we also show that electrical stimulation generates a virtual signal of angular acceleration about an axis through the head. Knowing how these stimuli work and how the brain interprets them will assist in developing unique methods to investigate and manage a range of pathological conditions that produce abnormal sensations of movement and balance, and provide the understanding to develop novel virtual-reality techniques.

Introduction

If a blindfolded subject is rotated about a head-centred vertical axis, she usually has little difficulty returning to her start position. Under these circumstances the semicircular canals provide most, if not all, of the sensory information required by the spatial orientation process. Furthermore, if the semicircular canal afferents are stimulated electrically during the outward journey to create a signal of rotation about the vertical, she returns herself to a consistent but incorrect position (Day & Fitzpatrick, 2005). The direction and speed of self-generated guided movements are likewise modulated by canal afferent stimulation (Day & Reynolds, 2005; Fitzpatrick et al. 2006). Thus, vestibular signals contribute to the internal representation and perceptions of self-motion and orientation. What are the limitations of the vestibular system for this function? These limitations relate to transduction accuracy and to time-varying signal changes that occur in the vestibular and perceptual machinery. It is the latter that interest us here.

Time-varying signal changes to any constant stimulus arise through neural processes of habituation and adaptation. These can occur at multiple levels from the receptor through to central networks. Adaptation at the receptor level is well established for the semicircular canals. When the head moves at a constant angular velocity, after an initial deflection the cupula passively returns to its initial alignment, resulting in a corresponding decay in the afferent firing rate (Goldberg & Fernandez, 1971). Among mammals, typical estimates of the cupula time constant span 4–7 s (Fernandez & Goldberg, 1971; Curthoys et al. 1977; Cohen et al. 1981; Oman et al. 1987; Morita et al. 2003).

A central brainstem mechanism exists, however, that prolongs the rapid canal-cupula adaptation, causing the signals that drive both the vestibulo-ocular reflex and the perception of rotation to outlast the peripheral signal of head rotation. Referred to as the ‘velocity storage’ mechanism, it lengthens the time constant of oculomotor decay to constant head rotation to 15–20 s when no visual input is available (Waespe & Henn, 1977; Robinson, 1986; Anastasio & Robinson, 1991; Hain & Zee, 1992; Vibert et al. 1997). The perception of rotation in darkness is closely associated with the slow phase of ocular nystagmus (Guedry & Lauver, 1961; Young & Oman, 1969; Brandt et al. 1974), although the perceptual threshold may be higher than the ocular threshold (Seemungal et al. 2004).

On top of this, there is evidence for an additional adaptation not apparent behaviourally with constant angular velocity stimuli because peripheral adaptation dominates. It can, however, be revealed by maintaining cupula deflection by constant angular acceleration, during which both perceptual and vestibulo-ocular responses can show long-term declines (Guedry & Lauver, 1961; Guedry & Collins, 1968; Brown & Wolfe, 1969). The origin of this time-varying response is uncertain.

A galvanic vestibular stimulus acts beyond the canal-cupula transduction process in the hair cell and afferent (Goldberg et al. 1984; Rabbitt et al. 2005; Aw et al. 2008) to produce a vestibular response uncontaminated by somatosensory influences. As this stimulus does not invoke the rapid signal adaptation of the canal-cupula mechanism, we apply it here to reveal the longer-term processes of adaptation and habituation. We have previously identified the direction of the rotation vector evoked by bilateral galvanic stimulation (Day & Fitzpatrick, 2005; St George & Fitzpatrick, 2010). In exploring the perceptual adaptation here, we now also ask what are the equivalent units of the galvanic rotation vector, i.e. is a constant galvanic signal interpreted as angular velocity, angular acceleration or other? By comparing perceptual responses to galvanic and kinetic canal stimuli, the latter containing both the canal-cupula and central signal transformations, we tease out and model the different adaptation processes involved in the vestibular-evoked sense of self-motion and spatial orientation.

Methods

All procedures received approval from the Human Research Ethics Committee of the University of New South Wales. All participants gave written, informed consent and the studies were conducted in accordance with the Declaration of Helsinki. Twelve healthy subjects, six of each sex, participated. They had no history of neurological or otological impairment and ranged in age from 21 to 54 years (mean 29.5, s.d. 8.7).

Vestibular stimulation

Two types of vestibular stimulation, kinetic and galvanic, were sequenced randomly between subjects.

The kinetic stimulus was a whole-body rotation imposed while subjects stood on a servo-controlled, smoothly rotating platform built into the floor of the laboratory. The stimulus had an acceleration phase of 1.25 deg s−2 for 16 s, a constant angular velocity of 20 deg s−1 (measured 95% confidence ±0.1 deg s−1) for 120 s, followed by a symmetrical deceleration (Fig. 1A). This acceleration profile was chosen because it is above reported perceptual thresholds of the semi-circular canals, which range between 0.22 and 1.18 deg s−2 (Rodenburg et al. 1981; Seemungal et al. 2004). To minimise centrifugal forces on the body and otolithic sensors, the subject was positioned so that the axis of rotation was between the feet and passed through the head.

Figure 1. Vestibular stimulation.

Figure 1

A, whole body rotation about a vertical axis. The velocity (ω) and acceleration (α) profiles of the rotation are shown. There was 16 s of constant acceleration, 2 min of constant velocity, followed by a symmetrical deceleration and then at least 60 s of no stimulation or until subjects reported no rotation. B, bipolar binaural transmastoidal galvanic stimulation. The stimulus current of either 0.75 or 1.5 mA was ramped on over 5 s, held constant for 2 min, and then ramped symmetrically down to zero where it remained for the duration of the trial.

The galvanic stimulus was delivered with a voltage-controlled current source via 3 cm2 Ag–AgCl electrodes placed over each mastoid with the side of the cathode randomised between subjects. The stimulus profile was trapezoidal with the current intensity linearly increasing from zero to 1.5 mA over 5 s and symmetrically decreasing back to zero after a plateau of 120 s (Fig. 1B). The ramp profile eliminates cutaneous sensations at the on and off steps. The peak stimulus intensity of 1.5 mA was chosen after a pilot investigation demonstrated that this level was tolerated over the testing time and elicited illusions of motion similar in magnitude to the kinetic stimulus. To investigate the relationship between stimulus intensity and response, a 0.75 mA stimulus was also used. Subjects stood on the platform, which was stationary during the galvanic stimulation, so that they were at the same position in the laboratory as during the kinetic stimulation.

Experimental protocol

Each subject performed three trials. A kinetic stimulus trial was randomly selected as clockwise or counterclockwise, and two galvanic stimulus trials at 1.5 mA and 0.75 mA were randomly selected as anode right or anode left. Trials were on different days with the order of presentation randomised between subjects.

The head was held in the same posture, facing the ground in front of the feet, for the duration of every trial, galvanic and kinetic alike. Using a spirit-level and protractor, the head was tilted forward so that Reid's plane (containing the inferior orbital margins and external auditory meatus) was approximately 18 deg to the vertical. This posture ensured that the galvanic-evoked rotation vector was vertical so all perceptions of motion were in the yaw plane (Day & Fitzpatrick, 2005; Fitzpatrick et al. 2006) and balance reflexes are minimised and negligible (Cathers et al. 2005). To remind subjects to keep the head at this angle throughout the trial, cotton straps between a head-piece and waist belt were adjusted so that they became taut at this angle. Mean maximum deviation from this desired posture over a trial was 1.2 deg. Noise-cancelling headphones played white noise to limit acoustic orientation cues and vocal contact, if needed, was always made with the experimenter positioned directly behind the subject.

Subjects stood blindfolded and stationary throughout all trials. They reported their perception of rotation direction (left, right) and rotation speed as a 0–5 score (0 = no rotation, 1 = just perceptible, 5 = very fast). They reported every few seconds and more frequently if perception changed quality. The vestibular stimulus, kinetic or galvanic, started 10 s after the subject called their first report, which was always zero.

Measurement and analysis

Response scores were recorded against time. Data were combined across stimulus directions after inverting perceptions of speed obtained from the clockwise kinetic stimulus trials and anode-left galvanic stimulus trials. Each subject's scores were normalised to the maximum score reported across all conditions. Repeated-measures ANOVA was used to compare the peaks and durations of the perception. Independent factors were stimulus (kinetic or galvanic), phase (pre-stimulus, post-stimulus) and stimulus current (0.75 mA, 1.5 mA). P < 0.05 was considered statistically significant.

A system model of rotation perception was adapted from the model of Leigh et al. (1981) describing vestibular–ocular interactions. The different time constants were estimated by non-linear, least-squares fits of the model to the empirical data using Matlab 7.4 and Simulink 6.6 software (The Mathworks Inc., Natick, MA, USA). Time constant and gain parameters of the model were constrained to positive values. Goodness of fit was determined with adjusted coefficients of determination (adjusted-R2).

Results

All subjects completed the experiments without signs of balance instability. Some initially reported a weak cutaneous sensation from the galvanic stimulus that abated over time.

Perception of rotation

The responses to kinetic and galvanic stimulation of a typical subject and the group mean are shown in Fig. 2.

Figure 2. Perceptions of rotation.

Figure 2

Perceived rotation during exposure to kinetic stimulation (A) and galvanic stimulation (B) are plotted for a typical subject and for the group in the lower plot. Units are the reporting scale: 0 = no rotation, 5 = very fast. The galvanic stimulus currents are 0.75 mA (red) and 1.5 mA (blue). The shaded regions show ±s.e.m. about the group means.

During kinetic stimulation (real whole-body rotation), subjects correctly reported the direction of turn during the acceleration phase. However, once the velocity of the platform was constant, the perception of turn decayed exponentially with an average time constant of 15.8 s, briefly reversed, and settled to zero after 26 s. As the platform decelerated to a stop, subjects perceived rotation in the opposite direction not different in peak amplitude (F1,11= 0.55, P= 0.47) and decay time (F1,11= 1.08, P= 0.32) from the perception at the onset of motion.

With galvanic stimulation (virtual rotation), subjects initially perceived rotation in yaw towards the cathodal electrode. This continued to increase for ∼10 s after the stimulus reached was at a constant current before slowly declining to zero. Peak perceived rotation was significantly faster at 1.5 mA than 0.75 mA (mean increase 83%; F1,11= 20.51, P= 0.001). These perceptions declined slowly to reach zero by 103 s and 84 s, respectively (F1,11= 7.74, P= 0.018). The perception of rotation evoked by the galvanic stimulus was maintained significantly longer (F1,11= 82.1, P < 0.001) than that of the kinetic stimulus, despite the raw peak rotations rated higher during the kinetic rotation (mean peak = 4.2) than during the galvanic rotation (mean peaks = 4.1 for 1.5 mA and 2.2 for 0.75 mA).

When the galvanic stimulus was turned off, subjects perceived rotation in the opposite direction (toward the anode) that did not differ significantly in absolute peak magnitude (F1,11= 0.48, P= 0.50) from the responses at the start of stimulation. These after-perceptions declined to zero by 96 s and 79 s for 1.5 mA and 0.75 mA, respectively and, although appearing shorter by visual inspection, these durations were not different statistically to those during the initial phase (F1,11= 3.00, P= 0.11).

System identification

The system model shown in Fig. 3 was used to characterise the signal processing that gives rise to these perceptual responses. Head velocity is transduced by the canal-cupula mechanism, which is modelled as a high-pass filter with decay time constant τc. This signal feeds the velocity-storage positive feedback loop that contains a low-pass element with time constant τv and a feedback gain kv. In parallel is a negative feedback operator with time constant τa that represents the adaptation process. This is included as a signal operator to indicate an overall signal processing rather than indicate specified physiological operators. It could operate at different points in the transfer between peripheral transduction and central perception (dashed lines) to create the same perceptual responses but different afferent traffic.

Figure 3. Analysis model of vestibular signal processing.

Figure 3

A, the peripheral vestibular system is modelled by the canal-cupula operator with time constant τc. The central vestibular pathways are modelled as a positive feedback velocity-storage operator with time constant τv and a negative feedback adaptation operator with time constant τa. The galvanic current (mA) is scaled by gain factor kg Reporting perception involves a time delay t and scaling by kr for the verbal score. s is the Laplace operator.Time constants and gain factors were estimated by least-squares fit of the experimental data. B, responses of the model for constant velocity and constant acceleration inputs (a), on entering the brainstem (b), and at the central perceptual level (c). Of note: (i) the constant-velocity responses show the effect of the velocity-storage signal extension and the adaptation after-effect, and (ii) the constant acceleration responses show the symmetrical adaptive reversal in the perceptual response and variable afference depending on relative action centrally and through the transduction process, as indicated by the dashed lines in A.

The output of this system (point c in Fig. 3) with an appropriate motor threshold would directly contribute to slow phase ocular motion in the model of Leigh et al. (1981). Here we apply a perceptual threshold of ±1 deg s−1, a conservative estimate of the vestibular angular velocity perception threshold (Mergner et al. 1993), and time delay (t) and scaling factor (kr) between perception and reporting. The galvanic stimulus input is downstream of the peripheral canal-cupula operator as it acts at the hair-cell/afferent-terminal complex. As an increasing angular velocity of the head is transduced into a mostly constant afferent signal (Highstein et al. 2005), the constant galvanic input must therefore not undergo the peripheral transformation.

The transfer function describing perceived rotation during the galvanic stimulation is:

graphic file with name tjp0589-0843-m1.jpg

where P is the rotation perception (−5 to +5 scale), the stimulus current I is in mA and s is the Laplace operator. With the known input (I: independent variable) and recorded values (P: dependent variable), non-linear least-squares error estimation yielded time constants τv= 7.70 s and τa= 75.9 s, a time delay to response t= 680 ms, and scaling constants kv= 0.52 and kr= 0.51. The galvanic scaling factor kg was 4.8 for the 0.75 mA and 3.9 for the 1.5 mA stimulus current (suggesting a power law relationship between sensation and stimulus strength). Figure 4A shows the model estimation and experimental data for the galvanic stimuli. The adjusted-R2 for both the 0.75 and 1.5 mA stimulus currents was 0.94.

Figure 4. Transfer function fitted to experimental data.

Figure 4

Outputs of the galvanic and kinetic transfer functions after least-squares parameter estimation (black curves) are plotted with the experimental data. A, galvanic stimuli and the mean responses for the 1.5 mA (blue) and 0.75 mA (red) stimuli, with the P/I transfer functions. B, kinetic rotation stimulus and the rotation perception, with the P/ω transfer function.

For the kinetic stimulus of whole-body rotation, the transfer function between angular velocity and perception is:

graphic file with name tjp0589-0843-m2.jpg

where P is the rotation perception (−5 to +5) and ω is the platform angular velocity (deg s−1). The parameter estimates of the galvanic transfer function were used to estimate the peripheral time canal-cupula time constant, which operates only in this kinetic rotation. By this least-squares estimation, τc= 7.3 s. The model fit and experimental values for this kinetic stimulus show a close association with an adjusted-R2 of 0.89 (Fig. 4B).

Discussion

For all sensory systems, the perception of a continuous or often repeated stimulus generally declines over time through habituation or adaptation at different levels in the nervous system. Peripheral habituation arises through receptor desensitisation whereas central habituation occurs when an irrelevant stimulus is ignored. Adaptation, however, brings about a fundamental change in the interpretation of the stimulus shown by an opposite after-effect when the stimulus is removed. The perceptions of rotation seen here show a strong long-term adaptation of the vestibular signal in addition to the well-known peripheral short-term adaptation with little evidence of signal habituation.

Responses to kinetic stimuli

The profile of the perceptual response to kinetic, whole-body rotation in the dark increased during the acceleration phase but declined exponentially during the constant-velocity phase with a time constant of 15.8 s, reflecting the short-term adaptation of the angular velocity signal by canal-cupula transduction and prolongation of the signal in the vestibular nuclear complex by the velocity-storage mechanism. Reversal of endolymphatic pressure during deceleration elicited an almost symmetrical perception of rotation in the opposite direction.

Responses to galvanic stimuli

Constant bilateral galvanic stimulation with the head oriented in forward pitch evoked a sensation of rotation about a vertical axis towards the cathodal side. Over time this sensation peaked and decayed to zero but lasted markedly longer than the kinetic response. The perception lasted longer at the higher stimulus intensity, largely explained by perception falling below threshold earlier with the lower stimulus intensity.

Human balance studies have not reported adaptive after-effects following up to 20 s of constant DC galvanic stimulation (Day & Cole, 2002; Wardman et al. 2003). During such protocols with the head, the galvanic stimulus evokes roll motion of the head and body. The strict requirements for upright balance increase the involvement of, and conflict with, other sensory systems causing quick return to the original stable posture. Such a phenomenon is observed with a head-bobbing nystagmus-like movement during sustained galvanic stimulation (Wardman et al. 2003). In the present study with the head pitched forward, the galvanic-evoked signal indicates rotation in yaw and is therefore free to evolve without conflicting signals from other sensory channels.

Comparison of stimuli

The perceived motion decays differently for the kinetic and galvanic inputs as these stimuli have different sites of action. A range of evidence shows that galvanic stimuli act at the hair cell and afferent complex, proximal to the signal integration of the canal-cupula mechanics (Goldberg et al. 1984; Rabbitt et al. 2005; Aw et al. 2008) and, as the effect of percutaneous currents is an approximately linear modulation of afferent firing rates (Goldberg et al. 1984), the perceptual response to a constant current should look like the real head movement that would create the same pattern of afferent discharge. As the canal-cupula transfer function is a low-pass filter that approximates integration, a corresponding real head movement should approximate the time differential of the galvanic stimulus. In other words, a constant-stimulus current is perceived approximately as constant angular acceleration. This is, in fact, intuitive because if the real head movement looks like the velocity profile of the stimulus, the perception of rotation dies away rapidly, as seen in Fig. 2.

Guedry & Lauver (1961) delivered constant angular accelerations about the vestibular axis as seated subjects recorded by button press every perceived 90 deg of turn. The idiosyncratic nature of the galvanic stimulus means that amplitude scaling will vary subject-to-subject and day-to-day but, from our previous studies (Day & Fitzpatrick, 2005; Fitzpatrick et al. 2006), we estimate that 1 mA is the equivalent of a real acceleration of 0.5–1 deg s−2. Thus, in Fig. 5 the reported velocities of Guedry & Lauver (1961) for accelerations of 1.0 and 2.0 deg s−2 are plotted alongside the reported turning rates of this study during galvanic stimulation at 0.75 and 1.5 mA. The prolonged decay profiles of the two perceived rotations are strikingly similar (aligned to their peaks, R2= 0.92 between the galvanic 0.75 mA and acceleration 1.0 deg s−2 data) with both reaching threshold values at approximately the same time (80–90 s). In addition, doubling the stimulus intensity for the galvanic (1.5 mA) and acceleration (2.0 deg s−2) stimuli showed similar scaling effects on their perception. This comparison indicates that perceptual centres interpret the galvanic stimulus profile as a real angular acceleration about the head-referenced axis identified previously (Day & Fitzpatrick, 2005).

Figure 5. Constant galvanic stimulation compared with constant acceleration.

Figure 5

The continuous plots show the profiles (mean ±s.e.m.) of the reported perceptions of rotation during and after the constant galvanic stimuli with the body stationary (blue 1.5 mA, red 0.75 mA). Superimposed points are the re-plotted data of Guedry & Lauver (1961) showing the profile of perceptions to sustained whole-body angular acceleration and deceleration at 1 deg s−2 (black) and 2.0 deg s−2 (grey; truncated because of motor limit). The decay profiles are within the 95% confidence limits of the galvanic-evoked perceptions. Note, however, that the deceleration represents cupulae deflections in the reverse direction whereas the galvanic reversal effect occurs on removing rather than reversing the galvanic stimulus.

Rotational and centrifugal forces could provide additional somatosensory cues during kinetic rotation that are not present during galvanic rotation but the similar adaptation profiles indicate that either they make a small contribution to the total perception or that they undergo habituation with a similar time course. For example, slowly-adapting (SAII) mechanoreceptors have a long habituation time constant of ∼30 s (Chambers et al. 1972). Over time, one might expect habituation in the vestibular response through either a reduction in the weight given to the vestibular channel or waning subject attention. When the galvanic stimulus was turned off, the sensation of rotation in the opposite direction (towards the previous anode) had a profile and strength close to the original sensation but the difference was not statistically different, indicating that adaptation dominates the perceptual process with little signal habituation over this time scale and with a single presentation.

The site of long-term signal adaptation

The time constants of the canal-cupula and velocity-storage processes identified in this work are within the range of previously reported values. The long-term signal adaptation is less commonly studied and how it comes about is less well understood. Rabbitt et al. (2005) show that with sustained cupula deflection, firing rates of canal primary afferents decline exponentially with individual time constants ranging from tens of milliseconds to hundreds of seconds, encompassing the time domain of the perceptual habituation observed here. The response of vestibular afferents to imposed head movements follows a power law (up to a point of saturation) with an exponent of 0.73 (Rabbitt et al. 2010). Balance reflexes evoked by galvanic stimuli obey a similar power law (exponent 0.55; Day et al. 2010), and in the present study measuring perceptual responses, the galvanic stimulus scaling factor (kg) was proportionally smaller at the higher stimulus and, assuming a power law, the data yield an exponent of 0.70. These similarities indicate that this effect, which provides for amplification of small signals, is mediated peripherally and that the perceptual response reflects the ‘size’ of the vestibular afference.

The reversal of the perceived rotation in the study of (Guedry & Lauver, 1961) can be explained by the reversal of cupulae deflections when changing from acceleration to deceleration. A continuous mechanical stimulus results in a slow decline of the firing rate of afferent vestibular neurons brought about through different processes within the receptor hair cells (Fernandez & Goldberg, 1976; Brichta et al. 2002; Gillespie & Cyr, 2004; Rabbitt et al. 2005). These peripheral signal transduction processes could contribute to the slow decline of the perceptual response to the kinetic stimuli but, as both phases present a constant non-zero stimulus (i.e. +1.0 deg s−2 and −1 deg s−2), the decline in the responses is evidence for signal habituation but the reversal of the signal with deceleration does not necessarily show adaptation.

In the present study, the galvanic stimulus returns to zero and not the state of opposite polarity that occurs with the reversal of the real acceleration vector with deceleration. It is not clear what effect this has on the afferent firing rates after terminating a prolonged constant-stimulus current. However, following brief periods of galvanic stimulation that increases afferent firing, an after-effect of reduced spontaneous firing that would be consistent with adaptation rather than peripheral habituation is not observed in the records of Goldberg et al. (1984). The same is seen with stimulation of efferent vestibular neurons to modulate afferent firing rates (Boyle & Highstein, 1990; Brichta & Goldberg, 2000; Rabbitt et al. 2010). After periods of increased discharge, albeit brief (up to 20 s), firing rates return to pre-stimulus levels without a reversal that indicates a peripherally generated adaptive rebound. These observations indicate that peripheral signal transduction processes can explain the decline in afferent rate or habituation of the response to a sustained stimulus, but the adaptation that creates the perceptual reversal is mediated centrally.

The similar habituation of perceptual and ocular responses to constant angular acceleration (Guedry & Lauver, 1961) and between the decaying perceptual responses recorded here and the slow-phase velocity of ocular nystagmus to constant rotation (Hain & Zee, 1992) suggests that common processes, with brainstem involvement, influence the vestibulo-ocular system and more rostral perceptual centres. Responses to galvanic vestibular stimulation do not habituate, even when anticipated and self-administered (Guerraz & Day, 2005), supporting the view that central processing involves subcortical networks. The decay of the perception of rotation during the kinetic step-velocity stimuli here was followed by a brief overshoot in which subjects perceived rotation in the opposite direction (Fig. 2). This overshoot appears to be associated with feedback action of a long-term adaptation process somewhere between perception and the site afferent activation as no overshoot is predicted when this operator is removed from the model. A similar overshoot is seen in the reversal of direction of the slow-phase vestibulo-ocular nystagmus following exponential decay to a prolonged rotation stimulus (Jeannerod et al. 1975; Sills et al. 1978; Furman et al. 2000). This implies that the long-term adaptive process is not exclusive to perceptual centres but also shapes brainstem projections to ocular motoneurons, supporting the hypothesis that the vestibulo-ocular reflex operates in the self-motion perceptual space (Bloomberg et al. 1991a,b;).

The evidence above indicates that the long-term adaptation involves changes within the brainstem and the afferent system. Vestibular efferent neurons synapse with hair cells and afferent neurons, and can have both excitatory and inhibitory effects on resting discharge and motion sensitivity (Furman et al. 1989; Boyle et al. 2009; Sadeghi et al. 2009). This gain modulation appears to be a key function of efferent control and it is not the means by which afference related to imposed head movement is separated from self-generated movement (Jamali et al. 2009). The observations above lead us to propose that the processes of adaptation and habituation involve centrally generated efferent pre- or postsynaptic innervations of the end organ. This is a means by which adaptation would be seen both peripherally and centrally with similar effects on both vestibulo-ocular and perceptual centres.

The function of long-term adaptation

A general principle of neural function is a drive towards homeostasis, whereby prolonged input signals are progressively accepted as the norm, adapting towards zero over time. It could be that the central adaptation is not specific to the vestibular system as such, but rather a general function of sensorimotor and perceptual processes to adapt sustained signals of self-motion. Adaptation of the constant-level vestibular input would allow greater sensitivity to detect changes in acceleration within the new framework that sets a ‘zero-rotation’ set point according to the long-term constancy of the bilateral afferent discharge. By this process, these frequency-encoded non-referenced signals can provide a signal of no rotation.

It should be considered that a signal of sustained angular acceleration as delivered by galvanic stimulation in this study or by kinetic stimulation in the study of Guedry & Lauver (1961) is probably never encountered in normal life. The length of the time constant of adaptation (76 s) could be considered appropriate then as it is well outside of the durations of accelerations experienced in most activities. There is some correspondence of this adaptation time constant in other sensory systems. For example, the perception of self-rotation evoked from viewing visual rotation of the surround shows signs of adaptation after 1–3 min exposure, as evidenced by a reversal in the direction of self-motion perception once in darkness (Brandt et al. 1974). Similarly, the sense of rotation generated from the somatosensory-motor pattern during locomotion adapts when conditioned with stepping in place on a rotating surface (Gordon et al. 1995). A robust locomotor after-rotation (podokinetic after-rotation) is observed with 7.5 min of conditioning (Weber et al. 1998) and current research from our laboratory suggests the time constant of podokinetic adaptation is as low as 2–3 min.

For linear acceleration, which is detected by the otolith system, we are continuously exposed to gravitational acceleration on which our movements are superimposed. Long-term adaptation can ‘centre’ this signal so that we do not experience a constant sensation of acceleration. Another natural cause of a persisting vestibular signal is an imbalance in the tonic levels of activity due to peripheral labyrinthine disturbances. Adaptive mechanisms could help to nullify sensations of acceleration and vertigo. Such mechanisms, however, may only operate effectively for small degrees of imbalance. After total unilateral loss of labyrinthine function, it takes weeks for vestibular tone to be brought back into balance through longer-term compensatory processes, and central adaptation could be an initial event in that process.

Application of galvanic stimulation

After a long history of exploratory experimental application (Fitzpatrick & Day, 2004), galvanic vestibular stimulation is now used widely in physiological, imaging, electrophysiological and psychophysical research into vestibular function and its role in balance, oculomotor and orientation control. Interest is increasing in its clinical diagnostic and rehabilitation potential (e.g. Rorsman et al. 1999; MacDougall et al. 2005; Marsden et al. 2005; Rosengren et al. 2009) as it provides a unique means of evoking a controlled pure vestibular signal, uncontaminated by somatosensory influences. Previous studies have identified the axis of the angular rotation signal evoked by galvanic stimulation (Cathers et al. 2005; Day & Fitzpatrick, 2005). The results of this study show that beyond the canal-cupula mechanical transduction, the galvanic-evoked afferent signal is treated as if it was a real angular acceleration of the head. Any intended application of galvanic stimulation should therefore consider the stimulus profile in addition to its direction. The commonly used constant stimulus currents (e.g. Zink et al. 1998; Bent et al. 2000; Day & Fitzpatrick, 2005) are treated by the nervous system as constant angular acceleration. Behavioural responses that deviate from the acceleration prediction indicate adaptation in the processing of the vestibular signal or the influences of other sensory systems. If the purpose of the galvanic stimulus is to evoke responses of natural movements, then the angular acceleration movement profile should be transformed to the equivalent galvanic stimulus with the canal-cupula transfer function (e.g. see Goldberg & Fernandez, 1971; Shkel & Zeng, 2006) to evoke an appropriate central signal. If the aim is to evoke responses to sustained angular acceleration, a constant galvanic current provides an easy way to elicit those responses without stimulating other sensory channels.

Conclusions

This study demonstrates that transmastoidal galvanic current, bypassing the peripheral vestibular integration, is processed by the nervous system as if its current profile was a real angular acceleration about a specific head-referenced axis. By investigating perceptions of rotation to both kinetic and galvanic stimuli, three time-varying signal transformations of vestibular processing pathways are revealed. Using the galvanic stimulation technique with the head aligned so that the virtual rotation axis is vertical and does not threaten stability, the long-term adaptation process which is normally not apparent during natural motion was shown to have a time course over a couple of minutes. This adaptation should improve the sensitivity of the system and establish the zero-rotation reference point about which functional actions can be organised.

Acknowledgments

This work was supported by the National Health and Medical Research Council of Australia. We thank Mr Hilary Carter for his valuable assistance with the construction of the equipment.

Author contributions

The experiments were conducted in Sydney at Neuroscience Research Australia. All authors were involved with conception and design. R.StG. collected and analysed data and wrote the first draft. All authors contributed to revision and approved the final submission.

References

  1. Anastasio TJ, Robinson DA. Failure of the oculomotor neural integrator from a discrete midline lesion between the abducens nuclei in the monkey. Neurosci Lett. 1991;127:82–86. doi: 10.1016/0304-3940(91)90900-e. [DOI] [PubMed] [Google Scholar]
  2. Aw ST, Todd MJ, Aw GE, Weber KP, Halmagyi GM. Gentamicin vestibulotoxicit impairs human electrically evoked vestibulo-ocular reflex. Neurology. 2008;71:1776–1782. doi: 10.1212/01.wnl.0000335971.43443.d9. [DOI] [PubMed] [Google Scholar]
  3. Bent LR, McFadyen BJ, Merkley VF, Kennedy PM, Inglis JT. Magnitude effects of galvanic vestibular stimulation on the trajectory of human gait. Neurosci Lett. 2000;279:157–160. doi: 10.1016/s0304-3940(99)00989-1. [DOI] [PubMed] [Google Scholar]
  4. Bloomberg J, Melvill Jones GM, Segal B. Adaptive modification of vestibularly perceived rotation. Exp Brain Res. 1991a;84:47–56. doi: 10.1007/BF00231761. [DOI] [PubMed] [Google Scholar]
  5. Bloomberg J, Melvill Jones GM, Segal B. Adaptive plasticity in the gaze stabilizing synergy of slow and saccadic eye movements. Exp Brain Res. 1991b;84:35–46. doi: 10.1007/BF00231760. [DOI] [PubMed] [Google Scholar]
  6. Boyle RD, Highstein SM. Efferent vestibular system in the toadfish: action upon horizontal semicircular canal afferents. J Neurosci. 1990;10:1570–1582. doi: 10.1523/JNEUROSCI.10-05-01570.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Boyle RD, Rabbitt RD, Highstein SM. Efferent control of hair cell and afferent responses in the semicircular canals. J Neurophysiol. 2009;102:1513–1525. doi: 10.1152/jn.91367.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brandt T, Dichgans J, Buchle W. Motion habituation: inverted self-motion perception and optokinetic after-nystagmus. Exp Brain Res. 1974;21:337–352. doi: 10.1007/BF00237897. [DOI] [PubMed] [Google Scholar]
  9. Brichta AM, Aubert A, Eatock R, Goldberg JM. Regional analysis of whole cell currents from hair cells of the turtle posterior crista. J Neurophysiol. 2002;88:3259–3278. doi: 10.1152/jn.00770.2001. [DOI] [PubMed] [Google Scholar]
  10. Brichta AM, Goldberg JM. Responses to efferent activation and excitatory response-intensity relations of turtle posterior-crista afferents. J Neurophysiol. 2000;83:1224–1242. doi: 10.1152/jn.2000.83.3.1224. [DOI] [PubMed] [Google Scholar]
  11. Brown JH, Wolfe JW. Adaptation to prolonged constant angular acceleration. Acta Otolaryngol. 1969;67:389–398. doi: 10.3109/00016486909125465. [DOI] [PubMed] [Google Scholar]
  12. Cathers I, Day BL, Fitzpatrick RC. Otolith and canal reflexes in human standing. J Physiol. 2005;563:229–234. doi: 10.1113/jphysiol.2004.079525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chambers MR, Andres KH, von Duering M, Iggo A. The structure and function of the slowly adapting type II mechanoreceptor in hairy skin. Q J Exp Physiol Cogn Med Sci. 1972;57:417–445. doi: 10.1113/expphysiol.1972.sp002177. [DOI] [PubMed] [Google Scholar]
  14. Cohen B, Henn V, Raphan T, Dennett D. Velocity storage, nystagmus, and visual-vestibular interactions in humans. Ann N Y Acad Sci. 1981;374:421–433. doi: 10.1111/j.1749-6632.1981.tb30888.x. [DOI] [PubMed] [Google Scholar]
  15. Curthoys IS, Blanks RH, Markham CH. Semicircular canal functional anatomy in cat, guinea pig and man. Acta Otolaryngol. 1977;83:258–265. doi: 10.3109/00016487709128843. [DOI] [PubMed] [Google Scholar]
  16. Day BL, Reynolds R. Vestibular reafference shapes voluntary movement. Curr Biol. 2005;15:1390–1394. doi: 10.1016/j.cub.2005.06.036. [DOI] [PubMed] [Google Scholar]
  17. Day BL, Cole J. Vestibular-evoked postural responses in the absence of somatosensory information. Brain. 2002;125:2081–2088. doi: 10.1093/brain/awf212. [DOI] [PubMed] [Google Scholar]
  18. Day BL, Fitzpatrick RC. Virtual head rotation reveals a process of route reconstruction from human vestibular signals. J Physiol. 2005;567:591–597. doi: 10.1113/jphysiol.2005.092544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Day BL, Marsden JF, Ramsay E, Mian OS, Fitzpatrick RC. Non-linear vector summation of left and right vestibular signals for human balance. J Physiol. 2010;588:671–682. doi: 10.1113/jphysiol.2009.181768. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Fernandez C, Goldberg JM. Physiology of peripheral neurons innervating semicircular canals of the squirrel monkey. II. Response to sinusoidal stimulation and dynamics of peripheral vestibular system. J Neurophysiol. 1971;34:661–675. doi: 10.1152/jn.1971.34.4.661. [DOI] [PubMed] [Google Scholar]
  21. Fernandez C, Goldberg JM. Physiology of peripheral neurons innervating otolith organs of the squirrel monkey. I. Response to static tilts and to long-duration centrifugal force. J Neurophysiol. 1976;39:970–984. doi: 10.1152/jn.1976.39.5.970. [DOI] [PubMed] [Google Scholar]
  22. Fitzpatrick R, Butler J, Day B. Resolving head rotation for human bipedalism. Curr Biol. 2006;16:1509–1514. doi: 10.1016/j.cub.2006.05.063. [DOI] [PubMed] [Google Scholar]
  23. Fitzpatrick RC, Day BL. Probing the human vestibular system with galvanic stimulation. J App Physiol. 2004;96:2301–2316. doi: 10.1152/japplphysiol.00008.2004. [DOI] [PubMed] [Google Scholar]
  24. Furman JM, Hain TC, Paige GD. Central adaptation models of the vestibulo-ocular and optokinetic systems. Biol Cybern. 1989;61:255–264. doi: 10.1007/BF00203172. [DOI] [PubMed] [Google Scholar]
  25. Furman JM, Koizuka IH, Schor R. Characteristics of secondary phase post-rotatory nystagmus following off-vertical axis rotation in humans. J Vestib Res. 2000;10:143–150. [PubMed] [Google Scholar]
  26. Gillespie PG, Cyr JL. Myosin-1c, the hair cell's adaptation motor. Ann Rev Physiol. 2004;66:521–545. doi: 10.1146/annurev.physiol.66.032102.112842. [DOI] [PubMed] [Google Scholar]
  27. Goldberg JM, Fernandez C. Physiology of peripheral neurons innervating semicircular canals of the squirrel monkey. I. Resting discharge and response to constant angular accelerations. J Neurophysiol. 1971;34:635–660. doi: 10.1152/jn.1971.34.4.635. [DOI] [PubMed] [Google Scholar]
  28. Goldberg JM, Smith CE, Fernandez C. Relation between discharge regularity and responses to externally applied galvanic currents in vestibular nerve afferents of the squirrel monkey. J Neurophysiol. 1984;51:1236–1256. doi: 10.1152/jn.1984.51.6.1236. [DOI] [PubMed] [Google Scholar]
  29. Gordon CR, Fletcher WA, Melvill Jones G, Block EW. Adaptive plasticity in the control of locomotor trajectory. Exp Brain Res. 1995;102:540–545. doi: 10.1007/BF00230658. [DOI] [PubMed] [Google Scholar]
  30. Guedry FE, Collins WE. Duration of angular acceleration and ocular nystagmus from cat and man II. Responses from the lateral canals to varied stimulus durations. Acta Otolaryngol. 1968;65:257–269. doi: 10.3109/00016486809120967. [DOI] [PubMed] [Google Scholar]
  31. Guedry FE, Lauver LS. Vestibular reactions during prolonged constant angular acceleration. J App Physiol. 1961;16:215–220. [Google Scholar]
  32. Guerraz M, Day BL. Expectation and the vestibular control of balance. J Cogn Neurosci. 2005;17:463–469. doi: 10.1162/0898929053279540. [DOI] [PubMed] [Google Scholar]
  33. Hain TC, Zee DS. Velocity storage in labyrinthine disorders. Ann N Y Acad Sci. 1992;656:297–304. doi: 10.1111/j.1749-6632.1992.tb25216.x. [DOI] [PubMed] [Google Scholar]
  34. Highstein SM, Rabbitt RD, Holstein GR, Boyle RD. Determinants of spatial and temporal coding by semicircular canal afferents. J Neurophysiol. 2005;93:2359. doi: 10.1152/jn.00533.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Jamali M, Sadeghi SG, Cullen KE. Response of vestibular nerve afferents innervating utricle and saccule during passive and active translations. J Neurophysiol. 2009;101:141–149. doi: 10.1152/jn.91066.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Jeannerod M, Magnin M, Schmid R, Stefanelli M. Vestibular habituation to angular velocity steps in the cat. Biol Cybern. 1975;22:39–48. doi: 10.1007/BF00340231. [DOI] [PubMed] [Google Scholar]
  37. Leigh RJ, Robinson DA, Zee DS. A hypothetical explanation for periodic alternating nystagmus: instability in the optokinetic-vestibular system. Ann N Y Acad Sci. 1981;374:619–635. doi: 10.1111/j.1749-6632.1981.tb30906.x. [DOI] [PubMed] [Google Scholar]
  38. MacDougall H, Brizuela A, Burgess A, Curthoys I, Halmagyi G. Patient and normal three-dimensional eye-movement responses to maintained (DC) surface galvanic vestibular stimulation. Otol Neurotol. 2005;26:500–511. doi: 10.1097/01.mao.0000169766.08421.ef. [DOI] [PubMed] [Google Scholar]
  39. Marsden JF, Playford DE, Day BL. The vestibular control of balance after stroke. Br Med J. 2005;76:670–679. doi: 10.1136/jnnp.2004.046565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mergner T, Hlavacka F, Schweigart G. Interaction of vestibular and proprioceptive inputs. J Vestib Res. 1993;3:41–57. [PubMed] [Google Scholar]
  41. Morita M, Imai T, Kazunori S, Takeda N, Koizuka I, Uno A, Kitahara T, Kubo T. A new rotational test for vertical semicircular canal function. Auris Nasus Larynx. 2003;30:233–237. doi: 10.1016/s0385-8146(03)00098-1. [DOI] [PubMed] [Google Scholar]
  42. Oman CM, Marcus EN, Curthoys IS. The influence of semicircular canal morphology on endolymph flow dynamics. An anatomically descriptive mathematical model. Acta Otolaryngol. 1987;103:1–13. doi: 10.3109/00016488709134691. [DOI] [PubMed] [Google Scholar]
  43. Rabbitt RD, Boyle R, Highstein SM. Mechanical amplification by hair cells in the semicircular canals. Proc Natl Acad Sci U S A. 2010;107:3864–3869. doi: 10.1073/pnas.0906765107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Rabbitt RD, Boyle RD, Holstein GR, Highstein SM. Hair-cell versus afferent adaptation in the semicircular canals. J Neurophysiol. 2005;93:424–436. doi: 10.1152/jn.00426.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Robinson DA. The systems approach to the oculomotor system. Vision Res. 1986;26:91–99. doi: 10.1016/0042-6989(86)90073-8. [DOI] [PubMed] [Google Scholar]
  46. Rodenburg M, Stassen HP, Maas AJ. The threshold of perception of angular acceleration as a function of duration. Biol Cybern. 1981;39:223–226. doi: 10.1007/BF00342774. [DOI] [PubMed] [Google Scholar]
  47. Rorsman I, Magnusson M, Johansson BB. Reduction of visuo-spatial neglect with vestibular galvanic stimulation. J Rehab Med. 1999;31:117–124. doi: 10.1080/003655099444632. [DOI] [PubMed] [Google Scholar]
  48. Rosengren SM, Jombik P, Halmagyi GM, Colebatch JG. Galvanic ocular vestibular evoked myogenic potentials provide new insight into vestibulo-ocular reflexes and unilateral vestibular loss. Clin Neurophysiol. 2009;120:569–580. doi: 10.1016/j.clinph.2008.12.001. [DOI] [PubMed] [Google Scholar]
  49. Sadeghi SG, Goldberg JM, Minor LB, Cullen KE. Efferent-mediated responses in vestibular nerve afferents of the alert macaque. J Neurophysiol. 2009;101:988–1001. doi: 10.1152/jn.91112.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. St George RJ, Fitzpatrick RC. The sense of self-motion, orientation and balance explored by vestibular stimulation. J Physiol. 2011;589:807–813. doi: 10.1113/jphysiol.2010.197665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Seemungal BM, Gunaratne IA, Fleming IO, Gresty MA, Bronstein AM. Perceptual and nystagmic thresholds of vestibular function in yaw. J Vestib Res. 2004;14:461–466. [PubMed] [Google Scholar]
  52. Shkel AM, Zeng FG. An electronic prosthesis mimicking the dynamic vestibular function. Audiol Neurotol. 2006;11:113–122. doi: 10.1159/000090684. [DOI] [PubMed] [Google Scholar]
  53. Sills AW, Honrubia V, Baloh RW. Is the adaptation model a valid description of the vestibulo-ocular reflex? Biol Cybern. 1978;30:209–220. doi: 10.1007/BF00361042. [DOI] [PubMed] [Google Scholar]
  54. Vibert N, De Waele C, Serafin M, Babalian A, Muhlethaler M, Vidal PP. The vestibular system as a model of sensorimotor transformations. A combined in vivo and in vitro approach to study the cellular mechanisms of gaze and posture stabilization in mammals. Prog Neurobiol. 1997;51:243–286. doi: 10.1016/s0301-0082(96)00057-3. [DOI] [PubMed] [Google Scholar]
  55. Waespe W, Henn V. Neuronal activity in the vestibular nuclei of the alert monkey during vestibular and optokinetic stimulation. Exp Brain Res. 1977;27:523–538. doi: 10.1007/BF00239041. [DOI] [PubMed] [Google Scholar]
  56. Wardman DL, Day BL, Fitzpatrick RC. Position and velocity responses to galvanic vestibular stimulation in human subjects during standing. J Physiol. 2003;547:293–299. doi: 10.1113/jphysiol.2002.030767. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Weber KD, Fletcher WA, Gordon CR, Jones GM, Block EW. Motor learning in the “podokinetic” system and its role in spatial orientation during locomotion. Exp Brain Res. 1998;120:377–385. doi: 10.1007/s002210050411. [DOI] [PubMed] [Google Scholar]
  58. Young LR, Oman CM. Model for vestibular adaptation to horizontal rotation. Aerospace Med. 1969;40:1076–1080. [PubMed] [Google Scholar]
  59. Zink R, Bucher SF, Weiss A, Brandt T, Dieterich M. Effects of galvanic vestibular stimulation on otolithic and semicircular canal eye movements and perceived vertical. Electroencephalogr Clin Neurophysiol. 1998;107:200–205. doi: 10.1016/s0013-4694(98)00056-x. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES