WO2025238298A1 - Electrical vestibular stimulation method - Google Patents

Abstract

A method is disclosed. The method comprises utilizing electrical vestibular stimulation (EVS) on a person; and measuring postural sway for the person. Further, A method of electrical vestibular stimulation to induce and measure postural sway in a person is disclosed. An arranged for carrying out either method is disclosed. Use of the apparatus to measure postural sway is disclosed.

Description

ELECTRICAL VESTIBULAR STIMULATION METHOD

FIELD OF THE INVENTION

The invention relates to methods and an apparatus for electrical vestibular stimulation (EVS ) to induce and measure postural sway in a person .

SUMMARY

This summary is provided to introduce a selection of concepts in a s implif ied form that are further described below in the detailed description . This summary is not intended to identify key features or essential features of the claimed subj ect matter, nor is it intended to be used to limit the scope of the claimed subj ect matter .

Example embodiments of the present disclosure enable a method utili zing electrical vestibular stimulation (EVS ) to induce postural sway in a person . This effect may be achieved by the features of the independent claims . Further example embodiments are provided in the dependent claims , the detailed description, and the drawings .

According to a first aspect , a method comprises :

• utili zing electrical vestibular stimulation (EVS ) on a person ; and

• measuring postural sway for the person .

According to an example embodiment of the first aspect , a waveform for the EVS may be of a varying frequency .

According to an example embodiment of the first , the waveform for the EVS may be of a decreasing frequency . According to a second aspect , a method of electrical vestibular stimulation to induce and measure postural sway in a person comprises : providing a stimulator ; providing at least two stimulation electrodes connected to the stimulator, wherein each stimulation electrode is non-invasive ; providing means for measuring postural sway, wherein the means for measuring postural sway measure directional acceleration or pressure ; placing the at least two stimulation electrodes near or over the left and right mastoid processes of the person, and optionally placing one stimulation electrode on the midline of the head of the person and/or the back of the neck of the person ; connecting the person to the means for measuring postural sway; applying stimulation to the at least two stimulation electrodes via the stimulator for a time period, thus inducing postural sway, wherein the stimulation is a frequency-modulated waveform, in particular a frequency-modulated sinusoidal alternating current ; and measuring, during the time period of the stimulation, a response comprising a time series measurement of the pressure or acceleration detected by the means for measuring postural sway .

According to a third aspect , an apparatus is arranged for carrying out the method of any preceding claim .

According to a fourth aspect, the apparatus of the third aspect may be used to measure postural sway .

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings , which are included to provide a further understanding of the invention and constitute a part of this specification, illustrate embodiments of the invention and together with the description help to explain the principles of the invention . In the drawings :

Figure 1 shows four diagrams showing examples of stimulation waveforms .

Figure 2 shows spectrograms of Center of Pressure signals in response to bilateral EVS .

Figure 3 shows a comparison of EVS frequency thresholds in different conditions .

Figure 4 shows electrical vestibular stimulation current waveforms .

Figure 5 shows Power spectrum of lateral Center of Pressure (CoP_x) signals from intensity sweep trials .

Figure 6 shows the creation of the computational model for the vestibular system .

Figure 7 shows spectrograms of lateral Center of Pressure (CoP_x) signals .

Figure 8 shows band bass filtered lateral Center of Pressure (CoP_x) signals as a function of stimulation current amplitude with constant frequency .

Figure 9 shows comparison of linear regression slopes between different types of electrical vestibular stimuli for intensity sweep trials .

Figure 10 show electric field model ing in a representative subj ect .

Figure 11 shows the dependence of the postural sway on the electric field and frequency in the vestibular system

Figure 12 shows the subj ect standing during a measurement .

DETAILED DESCRIPTION

Reference will now be made in detail to example embodiments , examples of which are illustrated in the accompanying drawings . The detailed description provided below in connection with the appended drawings is intended as a description of the present examples and is not intended to represent the only forms in which the present example may be constructed or utili zed .

Although the specification may refer to "an" , "one" , or "some" embodiment ( s ) in several locations , this does not necessarily mean that each such reference is to the same embodiment ( s ) , or that the feature may not apply to other embodiments . Single features of different embodiments may also be combined to provide other embodiments .

Furthermore , although the numerative terminology, such as "first" , "second" , etc . , may be used herein to describe various aspects or embodiments it should be understood that these aspects or embodiments should not be limited by this numerative terminology . In particular, the embodiments listed herein, are not to be understood as referring to only the " first" or the " second" aspect , instead may be applicable to both the "first" and the "second" aspect .

According to a first aspect , a method comprises :

• utili zing electrical vestibular stimulation (EVS ) on a person ; and

• measuring postural sway for the person .

With such a method, EVS may be used to induce postural sway in a person and measure the postural sway . For example , EVS may be utili zed via using stimulation electrodes and a stimulator . For example , measuring postural sway may be done by means of measuring postural sway .

According to an example embodiment of the first aspect , a waveform for the EVS may be of a varying frequency . This may be understood as that the stimulation used in the EVS is a waveform that may have a varying frequency or is frequency-modulated . The waveform may vary in frequency over time , such as continuously over time , such as over a time period .

According to an example embodiment of the first aspect , the waveform for the EVS may be of a decreasing frequency .

This may be understood as that the stimulation used in the EVS may have a varying frequency that may vary by decreasing . For example , the decrease is linear . For example , the frequency may vary by decreasing over time .

According to an example embodiment of the first aspect , the waveform for the EVS may at least 60 seconds of length, in particular at least 120 seconds of length .

This may be understood as that the EVS , the waveform, has a length, e . g . a time period during which stimulation is applied .

According to an example embodiment of the first aspect , two waveforms , in particular with time-varying frequencies , may be combined for the EVS .

This may be understood as that the EVS may comprise two waveforms that each may have a varying frequency that varies over time . For example , the EVS may comprise two waveforms , where one waveform has an increasing frequency, and the second waveform has a decreasing frequency .

According to an example embodiment of the first or second aspect , the method may comprise determining a frequency threshold as the highest frequency at which the EVS affects the postural sway .

This may be understood as that after utili zing EVS and measuring postural sway for the person, the highest frequency that still induces postural sway ( response ) is determined .

According to an example embodiment of the first or second aspect , the frequency threshold may be 5- 6 Hz or above , in particular 10 Hz or above .

According to an example embodiment of the first aspect , the method may comprise evaluating the effects of the EVS on the postural sway through changes in center of pressure (CoP) and/or linear acceleration .

This may be understood as that when postural sway is measured in a person, for example with means for measuring postural sway, the effects of the EVS on the postural sway are evaluated . This may comprise measuring a response from the means for measuring postural sway, which may be indicative of the effects of the EVS on the postural sway .

According to an example embodiment of the first aspect , the EVS may be performed with left- and/or right-sided unilateral stimulation .

This may be understood as a method where at least three stimulation electrodes , suitable for EVS , are provided and two may be placed near or over both left and right mastoid proces ses of the person and one on the midline of the head or the back of the neck of the person .

According to an example embodiment of the first aspect , the measuring may be performed utili zing a force platform, such as a force plate , and/or one or more wearable sensors , such as linear acceleration sensors .

The force platform may be understood as one example of a means for measuring postural sway and linear acceleration sensors may ben an alternative means for measuring postural sway . For example , the force platform may comprise pressure sensors that may be utili zed to measure postural sway in a person . For example , the force platform may comprise four pressure sensors or at least four pressure sensors .

According to an example embodiment of the first aspect , sinusoidal alternating current may be used for the EVS .

This may be understood as the waveform being a sinusoidal alternating current , e . g . the stimulation being a sinusoidal alternating current , where the current is approximately sinusoidal in a time period .

According to a second aspect , a method of electrical vestibular stimulation to induce and measure postural sway in a person comprises : a . providing a stimulator ; b . providing at least two stimulation electrodes connected to the stimulator, wherein each stimulation electrode is non-invasive ; c . providing means for measuring postural sway, wherein the means for measuring postural sway measure directional acceleration or pressure ; d . placing the at least two stimulation electrodes near or over the left and right mastoid processes of the person, and optionally placing one stimulation electrode on the midline of the head of the person and/or the back of the neck of the person ; e . connecting the person to the means for measuring postural sway; f . applying stimulation to the at least two stimulation electrodes via the stimulator for a time period, thus inducing postural sway, wherein the stimulation is a frequency-modulated waveform, in particular a frequency-modulated sinusoidal alternating current ; and g . measuring, during the time period of the stimulation, a response comprising a time series measurement of the pressure or acceleration detected by the means for measuring postural sway .

With such a method, EVS may utilized to induce and measure postural sway in a person . The method comprises EVS , which comprises a stimulation that is a waveform . The waveform may be frequency-modulated or have a varying frequency . While the stimulation i s applied the stimulation may induce postural sway . During the stimulation, a response to the stimulation is measured with means for measuring postural sway .

According to an example embodiment of the first or second aspect , at least three stimulation electrodes may be provided, and two may be placed near or over both left and right mastoid proces ses of the person and one on the midline of the head of the person and/or the back of the neck of the person .

This may be understood as providing means to perform EVS with bilateral , left- and/or right-sided unilateral stimulation .

According to an example embodiment of the first or second aspect , the means of measuring postural sway may comprise a force platform, comprising at least four pressure sensors , whereby the force detected by the four pressure sensors can be measured when the person stands on the force platform .

According to an example embodiment of the first or second aspect , a foam pad may be provided and placed on the force platform, whereby the person would stand on the foam pad, which is on the force platform .

According to an example embodiment of the first or second aspect , the means of measuring postural sway may comprise one or more wearable sensors , which comprise or are acceleration sensors , which may be placed on one upper arm of the person and/or on the top of the head of the person, whereby directional acceleration can be measured.

This may be understood as that at least one wearable sensor, such as an acceleration sensor, in particular linear acceleration sensor is provided and may be placed on the person. The acceleration sensor may, for example, be placed on one upper arm, such as the left upper arm, of the person, on the top of the head of the person, or on both.

According to an example embodiment of the first or second aspect, the one or more wearable sensors may comprise or consist of two wearable sensors, which may be placed on one upper arm of the person and on the top of the head of the person.

According to an example embodiment of the first or second aspect, the time period of the stimulation may be in the range of 60 - 180s, or the time period may be 120 s, or the time period may be split into two subperiods of 0 - 60 s and 60 - 120 s.

This may be understood as the waveform for the EVS being at least 60 seconds of length, such as 120 seconds of length.

According to an example embodiment of the first or second aspect, the frequency-modulated wave-form may have a peak amplitude in the range of 0.2 mA - 1.5 mA, or 0.2 mA - 1 mA, or 0.2 mA - 0.75 mA, or 0.2 mA - 0.5 mA, or 0.5 mA - 1.5 mA, or 0.5 mA - 1 mA, or 0.5 - 0.75 mA, or the peak amplitude may be 0.2 mA, or 0.5 mA, or 0.75 mA, or 1.0 mA, or 1.5 mA.

This may be understood as the waveform for the EVS having a peak amplitude, such as in the range of 0.2 mA - 1.5 mA, or 0.2 mA, or 0.5 mA, or 0.75 mA, or 1.0 mA, or 1.5 mA.

According to an example embodiment of the first or second aspect, the frequency-modulated waveform may have a frequency in the range of 10 - 0.1 Hz, or 10 - 2 Hz, or 2 - 0.1 Hz, or 4 - 12 Hz, or 6 - 14 Hz, or 2 - 6 Hz over the time period.

This may be understood as the waveform for the EVS having a peak amplitude such as in the range of 10

- 0.1 Hz over the time-period of the stimulation.

According to an example embodiment of the first or second aspect, the frequency-modulated waveform may have a linearly decreasing frequency in the range of 10

- 0.1 Hz, or 10 - 2 Hz, or 2 - 0.1 Hz over the time period .

This may be understood as the waveform having a decreasing varying frequency.

According to an example embodiment of the first or second aspect, the frequency-modulated waveform may have a linearly increasing frequency between two frequencies selected from: 2 - 10 Hz, 4 - 12 Hz and 6 - 14 Hz over a sub-period followed by a linearly decreasing frequency between the same two frequencies 10 - 2 Hz, 12 - 4 Hz and 14 - 6 Hz respectively over a sub-period.

This may be understood as the waveform is of a varying increasing frequency for, for example, 60 seconds of length, and then the waveform is of a varying decreasing frequency for, for example, 60 seconds of length .

According to an example embodiment of the first or second aspect, the stimulation may comprise a combination waveform which comprises two frequency-modulated waveforms, in particular two frequency-modulated sinusoidal alternating currents, that have linearly increasing or decreasing frequencies that vary in frequency in the range of 2 - 6 Hz or 6 - 2 Hz respectively over two sub-periods .

This may be understood as two waveforms, in particular with time varying frequencies, are combined for the EVS whereby the waveforms are of a varying frequency, and the first is of increasing frequency and the second is of a decreasing frequency, in the range of 2 - 6 Hz or 6 - 2 Hz .

According to an example embodiment of the first or second aspect , the stimulation may further comprise 0 . 5 s rectangular DC-pulses before and after the frequency-modulated waveform .

This may be understood as the waveform for the EVS having a 0 . 5 s rectangular DC-pul se before and after the waveform .

According to an example embodiment of the first or second aspect , the threshold of the frequency or peak amplitude that induces a response in the person may be calculated .

This may be understood as determining a frequency threshold as a highest frequency at which the EVS affects the postural sway .

According to an example embodiment of the first or second aspect , three stimulation electrodes may be provided, and two may be placed near or over both left and right mastoid processes of the person and one on the midline of the head or the back of the neck of the person, wherein stimulation may be applied to the two stimulation electrodes over the mastoid processes (bilateral ) , to the stimulation electrode over the left mastoid process and the stimulation electrode on the midline of the head or the back of the neck ( left-sided unilateral ) , or to the stimulation electrode over the right mastoid process and the stimulation electrode on the midline of the head or the back of the neck ( rightsided unilateral ) .

This may be understood as providing means for the EVS to be performed with left- and/or right-sided unilateral stimulation . According to an example embodiment of the second aspect , the stimulation of step f may be repeated, with an initial step f followed by one or more subsequent steps f , and measurement of step g may be repeated, with an initial step g followed by one or more subsequent steps g, optionally with a break between repetitions .

This may be understood as performing the method comprising utili zing EVS on a person and measuring postural sway for the person, after which performing the method one or more further times .

According to an example embodiment of the first or second aspect , when the stimulation of steps f is repeated, during each of the one or more subsequent steps f the stimulation may be applied to the same stimulation electrodes as during the initial step f or to stimulation electrodes independently selected from : bilateral , left-sided unilateral , or right-sided unilateral .

This may be understood as understood as providing means for the EVS to be performed with left- and/or right-sided unilateral stimulation and then performing the method with left- and/or right-sided unilateral stimulation, and then repeating the method with a further stimulation selected from : left- and/or right-sided unilateral stimulation .

According to an example embodiment of the first or second aspect , in the initial step f and each of the one or more subsequent steps f , the frequency-modulated alternating waveform may have the same or different frequency and amplitude in at least one sub-sequent step f .

This may be understood as performing the method and utili zing a waveform for the EVS , that is of a varying frequency and then repeating a method with a waveform that that is the same as the previous waveform or with a waveform with a different varying frequency .

According to an example embodiment of the first or second aspect , during the initial step f and/or at least one of the one or more subsequent steps f the stimulation may be a sham signal , whereby no frequency- modulated waveform is provided .

This may be understood as repeatedly performing the method wherein at least one instance of the method has a waveform for the EVS that doesn' t have an amplitude or a frequency .

According to a third aspect , an apparatus is arranged for carrying out the method of any preceding claim .

With such an apparatus , utili zing EVS on a person or applying stimulation to induce postural sway can be done .

According to an example embodiment of the third aspect , the apparatus may comprise an electrical vestibular stimulation system, in particular at least two stimulation electrodes and a stimulator .

This may be understood as comprising at least two stimulation electrodes , such as two or three stimulation electrodes .

According to an example embodiment of the third aspect , the apparatus may comprise a postural sway detector, in particular a force platform comprising at least four pressure sensors and/or one or more wearable acceleration and/or voltage sensors .

This may be understood as the apparatus comprising means for measuring postural sway, such as an force platform and/or one or more wearable sensors , such as acceleration or linear acceleration sensors .

According to an example embodiment of the third aspect , the one or more wearable acceleration and/or voltage sensors may comprise or consist of two wearable acceleration and/or voltage sensors .

According to a fourth aspect , the apparatus of the third aspect may be used to measure postural sway .

INVESTIGATION OF THE FREQUENCY RESPONSE OF ELECTRICAL VESTIBULAR STIMULATION WITH FREQUENCY-MODULATED CURRENTS

The sense of balance can be disturbed by applying electrical stimulation through the vestibular system . This technique is called electrical vestibular stimulation (EVS ) and has been used to study the function and sensory processing related vestibular system . However, there are gaps in the current understanding of how EVS stimulates the vestibular system as well as the relationship between intensity and frequency of the stimulation . Therefore , the purpose of the study is to measure responses to EVS through in vivo experiments and examine how frequency and current amplitude of sinusoidal alternating current relate to the magnitude of the sway response in different conditions .

Introduction

Electrical vestibular stimulation (EVS ) is a non-invasive technique that can be used to alter a person' s sense of balance and induce false sensations of movement [ 1 ] . Typically, this effect is achieved by applying electrical stimulation to the vestibular system through electrodes placed behind the ears on each s ide of the head . I f direct current is used as a stimulus , the technique is often referred to as galvanic vestibular stimulation and causes subj ects to lean in the direction of the anodal ( i . e . , positively charged) electrode until the stimulation stops [ 2 ] . When alternating current is used and the polarity of the electrodes changes over time , subj ects tend to follow and sway from side to side [ 3 ] .

In addition to full-body swaying, EVS can evoke a number of other responses , such as rapid eye movements and electromyography responses and has thus been used to study the function and sensory processing of the vestibular system [ 4 ] . I t has also recently gained interest as a therapeutic tool to improve postural stability to help patients suffering from vestibular dysfunction [ 5 , 6 ] . Even so , there are gaps in the current understanding of how EVS influences the vestibular system as its effects vary based on orientation of the head with respect to the body, whether the subj ect ' s eyes are open and the parameters of the applied stimulation .

The current study examines how postural sway responses of the electrical vestibular stimulation depend on the frequency and amplitude of sinusoidal alternating current through in vivo experiments with voluntary participants . The aim is to determine frequency thresholds for EVS , as well as to measure how frequency and current amplitude relate to the magnitude of the sway response in different conditions .

METHODS

The data of the study was gathered from two experiments that were designed to measure the effects of EVS on postural sway in terms of current amplitude and frequency of sinusoidal alternating current . The first experiment focused on frequencies below 10 Hz . However , our results showed that postural sway was affected even at 10 Hz . Therefore , the experimental setup was modified to account for frequencies up to 14 Hz . The details of both experiments are explained further below . Experimental Procedure As mentioned, the data was gathered from two sets of experiments . A single measurement session with one participant consisted of a set of two-minute trials during which stimulation was applied and participant ' s movement was recorded by a force platform (Wii Balance Board, Nintendo , Kyoto , Japan) . The subj ects were given approximately 60 s to rest after each trial to prevent fatigue and avoid possible carryover effects . During the rest period participants were additionally asked to describe their experience of the stimulation .

The first experiment had five participants ( ages between 26-40 ) . A measurement session consisted of 23 trials during which stimulation was applied and one without to measure natural postural sway during normal standing . During the trial the participant stood arms crossed, eyes closed and feet together on a foam pad placed on the force platform . Stimulation electrodes were placed over the left and right mastoid processes and on the back of the neck, forming three montages : ear to ear, left ear to neck and right ear to neck .

The second experiment had ten participants ( aged between 24- 63 ) . A measurement session included twelve trials with stimulation and two without . The measurement procedure was similar to the first experiment , except for every other trial the participants kept their eyes open and were instructed to stare at a marker on their eye level on a wall 5 m in front of them. Only one electrode montage was used with two electrodes placed over the left and right mastoid proces s ( ear to ear) .

All participants were in good general health, reported no history of vestibular dysfunction and had no contraindications to transcranial stimulation . Each subj ect gave their written informed consent , and the experimental procedure was approved by the local ethics committee. A trial or the entire session could be stopped at any point if the participant so wished.

Electrical Vestibular Stimulation

Stimulation was applied with a wireless, multichannel stimulator (Neuroelectrics Starstim 20, Barcelona, Spain) . The stimulation electrodes were circular (25 cm² surface area) and saturated in 0.9% saline solution. The stimuli waveforms were given in random order and neither the subject nor the experimenter knew which waveform was being applied at the time of measurement.

Stimulation waveforms were sinusoidal with no DC offset and lasted for 120 seconds. In the first experiment the waveform had either linearly increasing peak current amplitude (0-1.5 mA, Figure la) with constant frequency (0.1-2 Hz) or linearly decreasing frequency (10-0.1 Hz, Figure lb) with constant peak amplitude (0.5-1 mA) .

The waveforms of the second experiment all had constant current amplitude (0.75, 1.0 or 1.5 mA) with two types of varying frequency. For the first type the frequency simply increased and decreased linearly in time in the shape of a triangle (Figure 1c) between set frequencies (2-10, 4-12 and 6-14 Hz) . The second type combined two waveforms with time-varying frequencies between 2 and 6 Hz, forming two crosses as illustrated in Figure Id. The two waveforms are from here on referred to as 'Triangle' and 'Cross' types.

Waveforms of both experiments included short (0.5 s) rectangular DC-pulses before and after the actual stimulation. Sham also included DC-pulses at the start and end of measurement to make it more difficult to recognise among the other trials. Stimulation was monitored during the measurements and recorded for offline synchronisation of measurement instruments. The monitoring was done by measuring voltage from the skin in the first experiment and by measuring output current of the stimulator in the second experiment .

Figure 1 : Examples of stimulation waveforms , ( a) Current peak amplitude increases over time , (b) Frequency of the stimulation decreases over time , ( c) Stimulation frequency varies linearly between two frequencies ( 2-10 Hz , 4-12 Hz or 6-14 Hz ) . (d) Stimulation waveform is a combination of two sinusoidal waveforms with varying frequency between 2- 6 Hz .

Data Analysis

To start , the measurement data was synchronised after the measurements and cut to stimulation duration based on DC-pulses , which marked the start and end of stimulation . Then the centre of pressure (CoP) was calculated in order to analyse the effect of EVS on postural sway . The force platform had a pressure sensor in each of its corners that were used to calculate the CoP in left-right (CoP_x) and forward- backward (CoP_y) directions . Since the subj ects were often not perfectly al igned with the axi s of the platform, the sway directions were re-aligned with principal component analysis ( PCA) , so that the direction with greatest sway was on the left-right axis . Signals were then filtered around the stimulation frequency with a bandpass filter to only include movement affected by EVS . Spectrograms were calculated of modified CoP signals with Welch periodogram method with frequency resolution of 0 . 25 Hz and overlap percent of 95 % .

To determine frequency thresholds for EVS , pre- processed CoP_x signals of the triangle waveforms ( Figure 1c ) of the second experiment are compared to the original corresponding stimulation waveforms through crosscorrelation . Normalised cross-correlation between stimulation and CoP_x were calculated in a moving window of 15 seconds . The same process was applied to CoP signals of sham trials and 95th percentile value was determined for each subj ect . Lastly, binary logistic regression was used to determine frequency thresholds with the binary criterion that response correlation being either above or below the sham' s 95th percentile .

RESULTS AND DISCUSSION

The study is still ongoing, and its results shall be presented in further detail at the conference . However, examples of preliminary findings are presented here .

Figure 2 : Spectrograms of CoP_x signals in response to bilateral EVS . ( a) Experiment 1 spectrograms from trials where stimulation frequency decreased during the measurement divided into rows by applied current amplitude and columns by subj ect (denoted with ' S# ' ) . (b) Experiment 2 averaged spectrograms of ten subj ects divided into rows by type of waveform ( Figure 1 ) and whether subj ects ' eyes were open or closed . Used current amplitude and range of frequencies are listed above each image .

Figure 2 shows example spectrograms of original CoP_x signals ( i . e . , not realigned with PCA) of experiment 1 from trials , where stimulation frequency decreased over time . In most spectrograms EVS produced a line closely resembling the change in applied stimulation frequency ( Figure lb) . Thi s means stimulating the vestibular system at a specific frequency amplified natural left-right postural swaying at that frequency . Visual inspection indicated that this response was not present in CoPy signals ( forward-backward sway) and no higher-order harmonics were detected in the spectrograms , suggesting that the response was linear .

Results of experiment 1 showed that sensitivity to stimulation varied between subj ects . For one subj ect ( S 5 ) the spectrogram line was only visible below 2 Hz , while for others the highest amplified frequency reached at least to 5 Hz and even up to the highest frequency of 10 Hz . Unfortunately, the highest frequencies could not be determined aside from visual inspection due to the low signal-to-noise ratio . Thus , the measurement system was improved, and the second part of the study increased the upper range of examined frequencies to 14 Hz .

Figure 3 : Comparison of EVS frequency thresholds between eyes open/closed condition and current amplitude .

Experiment 2 included ' Triangle ' and 'Cross ' type waveforms and the average spectrograms of ten subj ects are shown in Figure 2b . These spectrograms demonstrate that increasing current amplitude amplifies swaying at stimulated frequency as in experiment 1 . Furthermore , the spectrograms of Cross trials show that the effect can also be applied to two or more distinct frequencies simultaneously .

Figure 3 compares frequency thresholds ( i . e . , the highest frequency at which EVS affects postural sway) calculated with cross-correlation and binary logistic regression in different trial conditions . As expected, increasing the intensity of the stimulation also increased the frequency threshold . With the greatest current amplitude used in this study, the frequency threshold of both eyes open and closed conditions reached 11-12 Hz .

Although EVS has been widely studied, most previous studies on humans have focused on small scale responses (e . g . , limb EMG or eye movements ) and examined the effects of stimulation frequency in regard to them . While there have been studies that examined full-body swaying through CoP, those studies used either noisy waveforms with a wide frequency range [5, 7] or sinusoidal currents at frequencies below 1 Hz [3, 8] , reasoning that postural responses diminish rapidly as frequency is increased. Yet, the results here show that EVS can affect postural sway and the response can be detected even above 10 Hz.

In conclusion, the effects of electrical vestibular stimulation were examined using frequency-modulated currents through in vivo experiments. The results indicate that EVS causes harmonic swaying at the frequency of the stimulation even above 10 Hz. The results can be used to study the function of the vestibular system and the human sense of balance.

REFERENCES

[1] Fitzpatrick, R. C. and Day, B. L. (2004) . Probing the human vestibular system with galvanic stimulation. J. Appl . Physiol., 96 ( 6) : 2301-2316.

[2] Wardman, D. L., Day, B. L., and Fitzpatrick, R. C. (2003) . Position and velocity responses to galvanic vestibular stimulation in human subjects during standing. J. Physiol., 547 ( 1 ) : 293-299.

[3] Coats, A. (1972) . The sinusoidal galvanic body-sway response. Acta Otolaryngol., 74 (1- 6) : 155-162.

[4] Dlugaiczyk, J., Gensberger, K. D., and Straka, H. (2019) . Galvanic vestibular stimulation: from basic concepts to clinical applications. J. Neurophysiol., 121 ( 6) : 2237-2255.

[5] Iwasaki, S., Yamamoto, Y., Togo, F., Kinoshita, M., Yoshifuji, Y., Fujimoto, C., and Yama- soba, T. (2014) . Noisy vestibular stimulation improves body balance in bilateral vestibulopathy. Neurology, 82 (11) : 969-975. [6] Wuehr, M., Decker, J., and Schniepp, R. (2017) . Noisy galvanic vestibular stimulation: an emerging treatment option for bilateral vestibulopathy. J. Neurol., 264:81-86.

[7] Hannan, K. B., Todd, M. K. , Pearson, N. J., Forbes, P. A., and Dakin, C. J. (2021) . Vestibular attenuation to random-waveform galvanic vestibular stimulation during standing and treadmill walking. Sci. Rep. , 11 (1) : 8127.

[8] Latt, L., Sparto, P., Furman, J., and Redfern, M. (2003) . The steady-state postural response to continuous sinusoidal galvanic vestibular stimulation. Gait & posture, 18 (2) :64-72.

In vivo and dosimetric investigation on electrical vestibular stimulation with frequency- and amplitude-modulated currents

Objective: Normal function of the vestibular system can be disturbed using a noninvasive technique called Electrical Vestibular Stimulation (EVS) , which alters a person's sense of balance and causes false sensations of movement. EVS has been widely used to study the function of the vestibular system, and it has recently gained interest as a therapeutic tool to improve postural stability and help those suffering from vestibular dysfunction. Yet, understanding of how EVS stimulates the vestibular system, the current in- tensity needed to produce an effect and the frequencies at which it occurs have remained unclear.

Methods: The effect of EVS on postural sway was examined in five participants using sinusoidal alternating current with time-varying amplitude from 0 to 1.5 mA and frequency from 0.1 to 10 Hz for three electrode con- figurations. Dosimetry of the current flow inside the head was conducted using anatomically realistic computational models created individually for each subj ect based on magnetic resonance imaging data . An estimate for the minimal field strength capable of affecting the vestibular system was calculated with the finite element method .

Results : Bipolar EVS at frequencies up to 10 H caused harmonic full-body swaying, and the frequency of the sway was the same as that of the stimulation current . The si ze of the sway was amplified by increasing the current intensity . Dosimetry modeling indicated that , for 0 . 2 mA current , the average electric field strength in the vestibular system was approximately 10 -30 mV/m, depending on the electrode montage . The si ze of the measured postural sway was proportional to the montagespecific electric field strength in the vestibular system .

Significance : The results provide insight to EVS ' s working mechanisms and improve its potential as a tool to study the sense of balance .

Humans maintain balance via constant , small movements based on the input of dif ferent sensory sys tems working together with the most notable ones being visual , somatosensory and vestibular systems . The vestibular system is located inside the inner ear and is composed of three semicircular canals , which detect angular movements of the head, and two otolith organs , which sense linear accelerations . When this system functions normally and the information it provides of the body ' s spatial orientation matches that of the other systems , there should be no i ssue for a healthy person to remain standing upright .

However, the normal function of the vestibular system can be disturbed by stimulating it with electrical currents to create a false sensation of movement and alter the sense of balance (Fitzpatrick and Day, 2004) . This technique is referred to as Galvanic Vestibular Stimulation (GVS) or as Electrical Vestibular Stimulation (EVS) , where the former more often denotes the use of direct current. The stimulation is most commonly applied bilaterally with two electrodes placed on the skin near the mastoid processes on both sides of the head, in order to stimulate the left and right vestibular systems simultaneously (Dlugaiczyk et al., 2019) . Another option is to target only one side using unilateral stimulation with one electrode placed on the mastoid and the other on the midline of the head, for example at the back of the neck. If direct current is used, blindfolded subjects tend to lean toward the side of the anodal electrode until the stimulation stops (Wardman et al., 2003a, b) . This occurs because the current passing through the vestibular system modulates the spontaneous firing rate of vestibular nerve afferents. With bilateral stimulation, the firing rate is increased on the cathodal side and decreased on the anodal side (Goldberg et al., 1984) . When alternating current is used, the electrode polarity changes over time, causing subjects to follow and sway from side to side (Coats, 1972) . A forward-backward oriented sway can be induced with mono- polar electrode configuration (Cauquil et al., 1998; Scinicariello et al., 2003) , where electrodes of the same polarity are placed over the mastoids and a reference electrode is placed further away. Despite the common notion that EVS only acts on the nerve afferents, recent evidence suggests that it also directly activates vestibular hair cells, thereby contributing to the elicited response (Aw et al., 2008; Gensberger et al., 2016) .

EVS can evoke a number of responses, ranging from aforementioned fullbody swaying to involuntary eye movements and electromyography (EMG) responses, leading to its extensive use as a research tool to study the function and sensory processing of the vestibular system (Dlugaiczyk et al., 2019; Fitzpatrick and Day, 2004) . Generally, vestibular stimulation in the 1 mA range can reliably alter balance but direct current as low as 0.3 mA can cause a sideways lean (Wardman et al., 2003a) while eye torsion can be produced at 0.1 mA (Severac Cauquil et al., 2003) and at least up to 20 Hz with sinusoidal stimulus (Mackenzie and Reynolds, 2018) . Additionally, due to the great effect EVS has on balance, many recent studies have focused on exploring its potential as a method to improve postural stability. Noisy or stochastic vestibular stimulation, where a wide bandwidth of frequencies (generally between 0-30 Hz; Dakin et al. (2010) ; Dalton et al. (2014) ) is applied, has been shown to improve stability while standing and walking (Iwasaki et al., 2014; Wuehr et al., 2017) . As such, noninvasive EVS could be used to help patients suffering from conditions that cause persistent postural imbalance such as bilateral vestibulopathy (Wuehr et al., 2017; Dlugaiczyk et al., 2019) . However, EVS without noise, either with direct current or alternating current changing at a specific rate, only serves to destabilize the subject. In regard to alternating current at discrete frequencies, the destabilizing effect can easily be seen by sight alone at frequencies under 1 Hz and has been examined up to 4 Hz (Petersen et al., 1994; Latt et al., 2003; Coats, 1972) .

Although EVS is a relatively simple method, its effects vary based on orientation of the head with respect to the body (Fitzpatrick and Day, 2004) and the various parameters of the stimulation. Such parameters include the amplitude and frequency of the applied current and whether the stimulation is applied bi- or unilaterally ( Thomas et al . , 2020 ) . The vestibular system is clearly sensitive to electrical stimulation and relatively small currents are enough to affect postural stability, yet the effects of sinusoidal alternating current EVS on postural sway have not been studied on discrete frequencies higher than 4 Hz . Moreover, the minimal current ampl itude needed to alter the sense of balance and the strength of the electric f ield induced inside the vestibular system to influence its function are still unclear .

Therefore , this study was designed to evaluate how the minimal current amplitude or electric field strength required to produce a measurable response depend on the stimulation frequency through in vivo experiments . The goal was to see whether there exists a threshold point at which stimulation starts to affect the vestibular system, and, if the threshold exists , to measure it across different frequencies to obtain a strength-frequency threshold curve . For this reason, the focus was on examining low current amplitudes ( s ; 1 . 5 mA) at 0 . 1 - 10 Hz . The effect of EVS on postural sway was measured with f ive volunteer subj ects and the current flow inside their heads was simulated computationally using high-resolution personali zed computational models to estimate the in situ electric field in the vestibular system .

Methods

Subject information

Five healthy volunteers took part in the study ( 2 females , aged between 26- 40 , mean age 30 ) . The subj ects reported no history of vestibular dysfunction, and had no contraindications to transcranial stimulation or magnetic resonance imaging . Each subj ect gave their written informed consent and the experimental procedure was approved by Aalto University Research Ethics Committee (approval ID: D/1219/03.04/2022) .

Measurement procedure

A measurement session consisted of 22 two-minute trials during which stimuli was applied and the subject's movement was recorded. A single trial had the subject standing with their eyes closed, arms crossed, and feet together on a forceplate (sampling rate 63 Hz, Wii Balance Board (WBB) , Nintendo, Kyoto, Japan) . A 1.5 cm thick exercise pad made of closed cell foam plastic, was placed between the subjects' feet and the forceplate to disturb somatosensory feedback and to increase sway magnitude (Dlugaiczyk et al., 2019) . The subjects were instructed to keep their heads facing forward in a natural position during the measurement. Two wearable sensors (Movesense HR2, Movesense, Vantaa, Finland) were attached to the left-upper arm and on top of the head, over the vertex. The sensors recorded linear acceleration (sampling rate 52 Hz) of the respective body parts they were attached to. The same sensors also had the capability to measure voltage from the skin. Therefore, one sensor was used to monitor the stimulation by recording volt- age (sampling rate 125 Hz) from the neck via cables connected to two surface electrodes placed roughly 3 cm below the stimulation electrodes.

The stimuli were given in a random order and neither the subject nor the experimenter knew which waveform was being applied at the time of measurement. One sham measurement (no current) was also included to measure natural postural sway during quiet standing. The subjects were given approximately 60 s to rest between trials to prevent fatigue and avoid possible carryover effects of the stimulation (Bresciani et al., 2002) . During the rest period, subjects were encouraged to move freely and asked to describe their experience of the stimulation. They were also asked to guess whether the stimulation they experienced was the sham trial. A trial could be stopped at any time if a subject wanted it for any reason.

Electrical vestibular stimulation

Stimulation was applied using a wireless, multi-channel stimulator (Neuroelectrics Starstim 20, Barcelona, Spain) through circular surface electrodes (surface area 25 cm²) saturated in 0.9% saline solution. Two electrodes were placed behind the ears, near the mastoid processes, and one on the back of the neck, forming three electrode montages: bilateral (ear to ear) , left-sided unilateral (left ear to neck) and right-sided unilateral (right ear to neck) . Electrode locations were recorded with a handheld 3D scanner (Artec Leo, Artec 3D, Luxembourg) for later dosimetric analysis .

The applied current waveforms were sinusoidal with no offset and lasted 120 s. Two types of waveforms were used: intensity and frequency weeps. Intensity sweeps (Figure 4a) had linearly increasing peak current amplitude (peak 0-0.5 mA for the first 60 s and 0.5- 1.5 mA for 60-120 s) and constant frequency (0.1, 0.5, 1 or 2 Hz) . Frequency sweeps (Figure 4b) had linearly decreasing frequency (10- 2 Hz for 0- 60 s and 2- 0.1 Hz for 60-120 s) and constant intensity (0.5, 0.75, or 1 mA) . The four intensity sweeps and three frequency sweeps were tested with three montages, which led to 21 trials with stimulation and one sham. As the original purpose of the experiment was to measure the relationship between EVS current amplitude and frequency, the waveform types were designed to form a grid over the parameter space, as illustrated in Figure 4c. Moreover, current ampl itudes were limited below 1 . 5 mA to determine the existence of EVS thresholds , or lack t hereof . For the same reason, the examined frequencies were also set below 10 Hz since the higher the frequency the greater current amplitude is required to produce a measurable response . Using intensity and frequency sweeps also shortened the length of the experiment session, since we did not need to conduct separate measurements for every discrete frequency and amplitude combination .

All stimulation waveforms and the sham trial included a 0 . 5 s , 0 . 5 mA rectangular DC-pulse 0 . 25 s before stimulation as shown in Figure 4d . This DC-pulse was used for offline data synchroni zation between measurement instruments . The frequency sweeps had a 5 s DC ramp-up before actual stimulation to avoid discomfort when the stimulus was switched on . The sham trial ended in another identical DC-pulse , which also made it more difficult for the subj ects to distinguish the sham condition among the other trials .

Data analysis

The effects of stimulation on postural sway were evaluated through changes in the Center of Pressure (CoP) . The forceplate subj ects were standing on had four pressure sensors , one in each corner, which were used to calculate the CoP in the left-right (CoP_x) and forward-backward (CoPy) directions . After the measurement , the CoP signals and the linear acceleration data recorded by mot ion/ volt age sensors were synchroni zed offline according to time labels and cut to stimulation duration based on DC-pulses that were visible in the recorded voltage signal and marked the start and end of stimulation .

The power spectrum in Figure 5 shows that over the entire trial duration sinusoidal stimulation with a constant frequency produces a spike in the power spectral density of CoP x signal at that frequency. To examine the frequency spectrum of postural sway as a function of time, spectrograms of CoP_x signals were calculated with Welch periodogram method (MATLAB Signal Processing Toolbox, The MathWorks, Inc.) with 0.25 Hz frequency resolution and 95% overlap percent.

The relationship between stimulation amplitude and magnitude of postural sway was evaluated with linear regression. The data from intensity sweep trials (i.e. constant frequency and increasing amplitude) was filtered around stimulated frequency with a band-pass filter (band ± 0.05 Hz, steepness 0.85, stop -band attenuation 60 dB) to only include movement affected by stimulation. Then to evaluate the effect of the stimulation amplitude on postural sway, a first order linear regression model was fitted to the absolute values of the peaks in the filtered CoP x signal. The regression coefficient (i.e. slope) of bilateral and unilateral stimulation was compared to each other and to that of sham condition with Student's t-test.

Figure 4 - Electrical vestibular stimulation current waveforms, (a) Intensity sweep trials: current peak amplitude increases over the stimulation duration, (b) Frequency sweep trials: current frequency decreases over the stimulation duration, (c) Visualization of the grid the two types of waveforms form over frequency-amplitude parameter space, (d) Examples of the first 10 s for a frequency sweep (blue) , a intensity sweep (red) and the sham waveform (black) . All stimulation and sham waveforms start with a rectangular DC-pulse.

Figure 5 - Power spectrum of lateral Center of Pressure (CoP x) signals from intensity sweep trials of different stimulation frequencies with light gray lines marking filtering bandwidth. Computational models of the head and vestibular system

Subject-specific computational head models were created based on magnetic resonance (MR) images of the same five subjects using a process described previously (Soldati and Laakso, 2020) . The models consisted of a segmentation of the head with 0.5 mm resolution into separate tissue classes, which were: skin, fat, muscle, eyes, intervertebral disk, cortical and cancellous bone, dura mater, cerebrospinal fluid, blood, gray matter, white matter, and inner ear.

The model of the inner ear was first segmented as a single structure with thresholding based on the MR images. Then, with smoothed and expanded templates obtained originally from the MIDA model (lacono et al., 2015) , it was divided into three separate parts: the vestibular system (i.e. semicircular canals, utricle and saccule) , cochlea and the vestibulocochlear nerve. Major steps of the process are shown in Figure 6. Tissue conductivities were set as described in Nissi and Laakso (2022) , except for the vestibular system and cochlea. Because the inner ear is composed of bony and membranous labyrinths containing fluids with conductivities similar to cerebrospinal fluid (CSF; Parazzini et al. (2006) ) , the conductivity of cochlea and vestibular system was set to 0.0314 S / m, calculated using the Maxwell- Garnett formula as a mixture of 50% CSF inclusions in cortical bone. Stimulation electrodes were modeled according to the description provided in Stimulation-section (i.e. circular sponge-type electrodes with a surface area of 25 cm²) and placed on the surface of the head according to the 3D scan data.

Electric field modeling and analysis The electric field inside the head was approximated by solving the electric scalar potential equation under the quasi-static assumption and homogeneous Neumann boundary condition. This was done separately for all models and all electrode montages. First, the models were discretized using a uniform grid of 0.5 mm cubical first-order elements with an algorithm based on the finite element method (Laakso and Hirata, 2012) , and the resulting system of linear equations was iteratively solved until the relative residual norm reached a value below 10^-6. The electric field could then be calculated from the gradient of the scalar potential.

Since the inner ear and vestibular system included very thin structures, i.e. the semicircular canals, the aforementioned process was repeated with a higher resolution. The left and right inner ear, and tissues within 2.5 mm of them, were extracted from the head and re-discretized with 0.1 mm resolution cubical elements. Then, the current density obtained with the 0.5 mm resolution was used as the boundary condition to approximate the electric field in the target structures with 0.1 mm resolution.

The strength of the electric field in the target tissues (i.e. vestibular system and nerve) was evaluated in respect to the mean and 99th percentile values of the electric field strength for each subject and over the whole group. This was done by first calculating the field strength in each voxel of the target structure in the left and right vestibular systems, then determining the statistical values for the entire data set separately in each of the five head models, and lastly, using the subject specific results to calculate the average values for the group of subjects. The process was repeated for all examined montages. Figure 6. Creation of the computational model for the vestibular system, (a) Example T2-weighted magnetic resonance (MR) images of a subject from side and above with marked location of the vestibular system, (b) An example of a computational, tissue-segmented model of the head, where the inner ear structures are highlighted in red. (c) Original inner ear model obtained through MR image thresholding, (d) Templates of the vestibular system and cochlea fitted to the original model, (e) Final inner ear model composed of the vestibular system, cochlea and the vestibulocochlear nerve, (f) Inner ear model discretized to 0.5 mm cubical voxels, (g) Enhanced discretization in 0.1 mm voxels.

Due to anatomical differences, the electric field strength varied between subjects. Linear mixed- effect analysis in MATLAB was used to examine the influence of the frequency (f) and mean field strength over the left and right vestibular systems (E_mean) on the subjects' sensitivity to stimulation (i.e. the steepness of the linear regression slope) . The model included the Emean, f, and their interaction effect (E_mean x f) as fixed effects, and subject-specific intercepts as well as slopes for Emean, f, and E_mean x f as random effects. As the full model failed to converge, random correlations were dropped from the model. The model formula was slope- 1+ Emean* f + (1 + E_mean* f I Subject) . (1)

To determine the significance of each fixed effect term, likelihood ratio tests were used to compare the model to an alternative model that did not include the term in question. The level of statistical significance was chosen as P = 0.05.

After establishing the effect of mean electric field with the full model (1) , a simpler model slope- 1 + E_mean + ( 1 + E_mean I subj ect ) ( 2 ) was used to examine the fixed and random effects of E_mean separately at each tested frequency with the theoretical likelihood-ratio test in the same manner as with the full model .

Additionally, to evaluate whether the magnitude of the sway response depended on the simultaneous activation of both left and right vestibular systems or the maximum electric field strength of only on one side , the full model ( 1 ) was ref itted with the maximum electric field value across both sides (E_max) used in place of E_mean - The Akaike Information Criteria (AIC) of the two models were then compared to determine whether E_max or E_mean was a better predictor of the response .

Results

Subjective sensations and side effects

After each trial , the subj ects were asked to describe their experience of the stimulation . All subj ects reported feelings of stinging on the skin or muscle twitching directly under one or both of the electrodes . Other effects included sensations of twitching of the earlobe , slight di z ziness , and feeling their body shaking or swaying to one side . Two subj ects , who were familiar with the appearance of phosphenes , also reported seeing them during frequency sweep trials . The sensations disappeared quickly after stimulation was turned off , and when asked 24 hours after the experiment , no participant reported any side effects persisting after the experiment or appearing later on . None of the subj ects identified the sham trial correctly from the other trials . Stimulation with changing frequency and fixed current

Figure 7 shows spectrograms of CoP x signals obtained from frequency sweep trials for bilateral stimulation as well as left- and right-sided unilateral stimulation with all tested peak currents ( 0 . 5 , 0 . 75 and 1 mA) . The images are organi zed into columns by subj ect (denoted with ' S ' and a number) and rows by the type and intensity of applied stimulus . The last column on the right shows the average spectrogram of the first 90 s .

In most of the spectrograms , EVS produced a line that closely followed the change in simulation frequency ( Figure 4b) , meaning that stimulation at specific frequency amplified the left- right postural sway at that frequency . Visual inspection of the CoPy spectrograms indicated that this response was not present in the forward-backward directed sway signa 1 . Furthermore , no higher-order harmonics were detected in the spectrograms , meaning that the response was linear .

As can be seen from Figure 7 , the subj ects ' sensitivity to stimulation appeared to vary . For one subj ect ( S5 ) the characteristic line was only faintly visible in the spectrogram below 2 Hz , while for the others the highest observable amplified frequency depended on current amplitude and the type of stimulation, extending even to the highest used frequency of 10 Hz . How- ever, because of noise in the s ignal , the highest frequency could not be defined clearly aside from visual inspection . A part of the noise was inherent to the sensors of the force plate and the rest was caused by unconscious shifts in standing posture , referred to as natural postural sway . This type of natural swaying i s mostly contained at frequencies below 2 Hz , as visible in the mean spectrograms . For some subj ects , postural sway could also be seen in spectrograms calculated from the left-right linear accelerations of the head, and partially on the arm . However, these recordings had poorer signal-to-noise ratio ( i . e . the ampl itude of sway was smaller compared to natural sway) in comparison to the CoP and could thus not be used for analysis . Nevertheless , these results indicate that the stimulation caused a full-body postural sway .

Bilateral stimulation (where both left and right vestibular systems were stimulated simultaneously) produced the clearest response in all subj ects and was vis ible even at the lowest stimulus intensity, 0 . 5 mA, for three subj ects and faintly in the mean spectrogram . The response was weaker with unilateral stimulation and the spectrogram line appeared at lower frequencies than with bilateral stimulation of the same current amplitude .

Stimulation with changing intensity at a fixed frequency

Bandwidth filtered CoP x signals from intensity sweep trials with bilateral stimulation are presented in Figure 8 as a function of the applied peak current amplitude . The red line denotes a linear regres sion model fitted to the peaks of CoP signal , as described in the data analysis section, along with its corresponding R2 -value .

At tested frequencies starting from 0 . 5 Hz , natural swaying at a given frequency increased with the stimulation current amplitude . The same effect could also be seen with unilateral stimulation, though the si ze of the maximum CoP deviation was smaller than with bilateral stimulation, which is similar to the results of the previous section . At 0 . 1 Hz , the effect could not be seen as clearly, and the regression fit was poor because both natural and stimulation-amplified swaying movements at that frequency were slower and bigger . Naturally, no effects on postural sway could be detected from CoP signals recorded during sham measurements at any frequency . Along the Y-axis (posterior-anterior) , a similar effect could be observed but the slopes were only 20 % of those in the lateral direction , and mostly because the X- and Y-axes of the subj ects were not perfectly aligned with the axes of the forceplate .

Figure 9 compares the CoP x slopes of the fitted regression models be- tween sham, bilateral and unilateral EVS at each tested frequency . At 0 . 5 Hz and above , slopes of both bilateral and unilateral EVS were significantly greater than those of sham ( Student ' s t- test , all P<0 . 05 ) , indicating that stimulation did amplify the rate at which magnitude of sway increased . Bi lateral EVS appeared to cause a greater slope than unilateral EVS on average , but the difference was not statistically significant ( all P>0 . 05 ) because of the small number of samples . Same was true for comparison between left- and right-sided unilateral EVS .

Electric field strength in the vestibular system

The reason for increasing the stimulation current slowly over time during intensity sweep trials was to see whether or not there existed a threshold at which stimulation began to interfere with the normal function of the vestibular system . However, based on the results of the previous section, such a threshold did not exist and instead, the stimulation could be said to amplify swaying almost instantly . For the purpose of estimating the lowest in situ electric fields capable of causing postural responses , a peak current ampl itude of 0 . 2 mA was chosen as the input stimulus intensity, since notable swaying could be seen at that amplitude across all tested frequencies and subjects. The value also corresponds with the results of earlier GVS studies with a focus on postural balance (Wardman et al., 2003a; Yang et al., 2015; Coats , 1973) .

The results are listed in Table 1 in terms of the mean and 99th percentile electric field averaged over all five subjects in the vestibular system and the vestibulocochlear nerve. Figure 10 shows a streamline visualization of the current flow through the entire head and an example of the electric field distribution at 0.5 mm depth inside the vestibular system of one subject. As expected, bilateral EVS produced fields of similar strength in both left and right vestibular systems, while asymmetric unilateral EVS induced stronger fields on the targeted side and weaker fields (~ 30- 40 %) on the opposite side of the head. The field strength was also higher in the targeted organs with bilateral than with unilateral stimulation. Across all montages using 0.2 mA stimulus, the average mean in the vestibular system and the nerve was approximately 20-30 mV/m and the average 99th percentile was 40-60 mV/ m. In other words, the field strength in the nerve and the vestibular system were nearly the same in regards to the mean and 99th percentile.

The hair cells of semicircular canals are located inside the crista within the ampulla at the end of each semicircular canal, while the otolith organs, saccule and utricle, are located in the middle section of the vestibular model. Looking at the field distribution in Figure 10, the maximum was found in the intersection of semicircular canals, near the ampulla, regardless of the subject, though the exact location of the maximum varied slightly. Similarly, around the approximate location of the otolith organs in the middle section, the field strength was close to its mean value over the whole vestibular system.

Linear mixed-effect models (Figure 11) were used to study the impact of the mean strength of the vestibular electric field and frequency on the size of sway. However, data from 0.1 Hz trials was excluded from the full linear mixed-effect analysis, since the many of the slopes were close to zero (Figure 8) and differed from the results of other frequencies in a non-linear manner. At frequencies above 0.5 Hz, the fixed effect of the electric field was found statistically significant (likelihood ratio tests, x² (1) ⁼ 7-50, P = 0.006) , indicating that the differences in the strength of the vestibular electric fields caused by differences between electrode montages had an effect on the sub jects' sensitivity to EVS, montages producing larger electric fields leading to a larger sway. As indicated by the contour map on the left-s ide of Figure 8, the sway size was affected by the applied frequency, higher frequencies leading to a smaller sway, and the fixed effect of frequency was found statistically significant (x² (1) = 6.25, P = 0.012) . The interaction effect between the electric field and frequency was also significant (x² (1) = 6.16, P = 0.013) . This means that the effect of electric field depended on the applied frequency, and the higher the frequency of the stimulus, the greater the field was needed to evoke the same response (8) .

Simpler linear mixed-effect models, visualized on the right side of Figure 11, were used to examine the random effect of the mean electric field individually on each tested frequency. The effect was found to differ significantly between subjects at all frequencies (likelihood ratio tests, PdO.038) , i.e., the sway amplitude for the same electric field strength varied between sub- j ects . Lastly, the model with mean electric field over left and right vestibular systems was compared to a model with maximum field strength over both systems through AIC model select ion . The model with mean electric field was retained for ana lysis as it had lower AIC score (AIC difference : 6 . 48 ) . This means the mean electric field was a better predictor of the sway amplitude and that the larger sway of bilateral EVS was more likely caused by the in situ electric field affecting both vestibular systems simultaneously instead of only the vestibular system with higher maximum field strength being affected by EVS .

Table 1 - Mean and 99th percentile values of the electric field (mV/m) in the vestibular system (VS ) and the vestibulocochlear nerve calculated with 0 . 2 mA peak current amplitude for bilateral , left -sided unilateral and right-sided unilateral stimulation . Sample mean ± standard deviation over five sub j ects is shown .

Figure 7 - Spectrograms of lateral Center of Pressure (CoP_x) signals in response to bilateral , leftsided unilateral (L-unilat . ) and right-sided unilateral (R-unilat . ) frequency sweep trials organi zed into rows by the stimulation type ( i . e . electrode montage and peak current amplitude ) and columns by the subj ect (denoted with ’ S# ’ ) . Mean spectrogram over all subj ects is shown on the right . Measurements lasting less than 120 s were stopped by subj ects ' request, and those under 30 s were excluded from analysis . In most cases , vestibular stimulation produced a single piece-wise linear curve similar to the input current waveform shown Figure 4b . Figure 8 - Band bass filtered lateral Center of Pressure (CoP_x) signals (black) as a function of stimulation current amplitude with constant frequency, and linear regression line (red) fitted to signal peaks with corresponding p-values shown on top of each figure. Stimulation frequency for each row is marked on the left and the subject number over each column. Measurements lasting less than 120 s were stopped by subjects' request .

Figure 9 - Comparison of linear regression slopes (Figure 8) between different types of electrical vestibular stimuli for intensity sweep trials: bilateral, left- and right-sided unilateral, and sham. Student's t-test comparison of stimulation against sham of the same frequency: (*) p<0.05, (**) pcO.01, (no symbol) pho.05.

Figure 10 - Electric field modeling in a representative subject. Top row: visualized current flow between electrodes for bilateral and left-sided unilateral vestibular stimulation. Bottom row: Electric field distribution at 0.5 mm depth in the left and right vestibular systems (seen from above) with instantaneous field direction marked with arrows.

Figure 11 - Dependence of the postural sway on the electric field and frequency in the vestibular system. On the left, is a contour map of the magnitude of the slope as a function of mean electric field (E_mean) and frequency predicted by mixed linear effect model for all data points. In the panels on the right, corresponding to 0.1, 0.5, 1 and 2 Hz, the line segments show the linear mixed-effect model fits for each subject (marked by color) with the slope of the postural sway response as a function of the mean electric field (E_mean) • The markers indicate the observed values for each electrode montage and subject. P-values PJ and Pr were obtained from likelihood ratio tests for the fixed and random effects of the electric field, respectively.

Discussion

This study examined the effects of sinusoidal alternating current electrical vestibular stimulation (EVS) on postural sway in terms of frequency and intensity. The results show that bipolar EVS induces a full-body postural sway in lateral (i.e. left-right) direction, that can be seen as harmonic changes in the recorded center of pressure. The sway response is frequency specific, noticeable even at 10 Hz, and is amplified by increasing the intensity of the stimulus. The same effect can be achieved with both bilateral and unilateral stimulation in the lateral direction. Electrode montages that produce a stronger electric field in the vestibular system lead to a larger postural sway. Most previous studies on the effects of EVS on humans as a function of frequency have focused on small scale responses, such as eye movements or surface EMG of limbs (Dlugaiczyk et al., 2019) . While some studies have used CoP as a metric for the effects of EVS, those studies either used noisy current waveforms with a wide frequency range (Iwasaki et al., 2014; Hannan et al., 2021) or sinusoidal currents at discrete frequencies lower than 1 Hz (Latt et al., 2003; Coats, 1972) . The reason why the latter focused on frequencies below 1 Hz was that the response diminishes sharply at higher frequencies, even though postural torque measurements have been conducted to 4 Hz (Petersen et al., 1994) . However, according to the results presented here, postural sway response can be detected from lateral CoP at frequencies higher than previously thought, even at 10 Hz, which was the highest frequency studied. This could be due to the long measurement duration of 120 s, which enabled frequency-domain analysis using spectrograms and increased the signal-to-noise ratio. It should be noted that the results were achieved by making standing posture during the measurements as unstable as possible by requesting the subjects stand with their feet together, eyes closed, and arms crossed on an unstable surface. All these factors are well-known to increase the size of the sway response (Fitzpatrick and Day, 2004; Dlugaiczyk et al., 2019) .

Examining the relationship between the strength of the stimulus and the magnitude of the response, the results were similar to previous studies with DC (Wardman et al., 2003a, b; Yang et al., 2015) and AC (Latt et al., 2003) , where increasing the current amplitude resulted in a larger response, starting from 0.2 mA. The use of linearly increasing current amplitude showed a similar increasing trend in the harmonic CoP displacement. However, unlike in the study of Latt et al. (2003) , where the magnitude of the response did not increase with the sinusoidal current above 0.5 mA, here the trend continued to 1- 1.5 mA depending on the subject. The reason could be that higher frequencies required higher currents than 1.5 mA, or that such saturation point varies greatly between individuals and was thus not detected in these five subjects.

Using 0.2 mA peak amplitude as an estimate for the lowest current capable of affecting the vestibular system, in situ electric fields were calculated in subject-specific, anatomically realistic volume conductor models. Around the approximate location of the otolith organs, responsible for sensing linear accelerations, the field strength was found to be between 20 to 30 mV/m. For crista of the semicircular canals, which detect changes in angular acceleration, the field was approximately 40- 60 mV/m. In the modeled section of the vestibulocochlear nerve , the mean field strength was 20 - 30 mV/m. A study by Thomas et al . (2020 ) showed similar results in regards to the location of maximum field strength and that the field strength in semicircular canals was higher than around otoliths . Although the focus of that study was more on the current flow and field distribution, the numerical values achieved with 1 mA input current scaled down to 0 . 2 mA were similar in magnitude as the results presented here . The remaining difference could be explained with the number of segmented tissues and different electrical conductivity values used . Currently, it is thought that EVS acts directly on the vestibular nerve afferents (Fitzpatrick and Day, 2004 ; Goldberg et al . , 1984 ) , but several studies suggest that vestibular hair cells are also activated by the stimulation and contribute to the evoked response (Norris et al . , 1998 ; Aw et al . , 2008 ; Gens- berger et al . , 2016) . Given that electric field strengths in the excess of 6 V/m are needed to trigger action potentials in the endings of large myelinated axons (Reilly, 1989 ) , the strength of the electric field in the vestibular nerve is far too low to affect the afferent neurons directly . Therefore , our data provides indirect support for the hypothesis that EVS affects hair cells without a direct effect on the vestibular afferents .

Linear mixed-effect analysis suggested that the mean field strength in the vestibular system had a statistically significant positive effect on the si ze of postural sway . The bilateral electrode montage , which produced a stronger mean electric field in the vestibular systems also led to larger postural sway compared with the unilateral electrode montages . Comparison of the mean and maximum electric field models further demonstrated that the said effect was more likely due to the s imultaneous stimulation of both left and right vestibular systems , and not due to the higher field strength on only one side . This agrees with the statement that bilateral EVS modulates the function of the vestibular system on both sides of the head, increasing the firing rate of vestibular hair cells on one side and decreasing it on the opposite side (Goldberg et al . , 1984 ) . Unilateral EVS modulates only the targeted side and the other side remains relatively unaffected due to low field strength ( Figure 7 ) . The ana lysis also indicated that , while all subj ects swayed from side to side when stimulated with a sufficiently high electric field, the effect of the electric field on the si ze of postural sway significantly varied between subj ects , such that the same electric field produced variable sway amplitude in each individual . The causes for these differences are still unclear, but they might be related to differences in the subj ect ' s physical characteristics , e . g . , height and weight , level of alertness , or the sensitivity of the vestibular control of balance . It should also be noted that the function of the subj ects ' vestibular system were not evaluated before the measurements , and although the subj ects were relatively young and did not report any issues with balance , a part of the sensitivity differences could have been caused by undetected vestibular dysfunction .

According to these results , a weak vestibular electric field below 60 mV/m can disturb the sense of balance and cause false sensations of movement . To provide some perspective , the basic restriction for human exposure to time varying electric and magnetic fields recommended by the International Commission on Non- Ioni zing Radiation Protection ( ICNIRP) for all tissues at frequencies lower than 3 kHz are 400 mV/m for the general public and 800 mV/m for occupational exposure ( ICNIRP, 2010) . These basic restrictions aim to protect against adverse health effects that exposure to external electromagnetic fields can cause through peripheral and central nerve stimulation. For tissues of the central nervous system for fields within 10- 25 Hz, the basic restriction is 50 mV/m, which is set to prevent the induction of retinal phosphenes. An earlier study by Nissi and Laakso (2022) examined retinal electric fields at the phosphene threshold for electric and magnetic stimulation. The retinal electric fields (5- 80 mV/m) were of similar magnitude to the vestibular electric field of the current study. While the hair cells of the vestibular system rely on mechanotransduction, they share similarities with the retinal photoreceptors (Lagnado and Schmitz, 2015; Kremer et al., 2006; Cosgrove and Zallocchi, 2014) , stimulation of which is suspected to cause phosphenes. By our results, both structures may be similarly sensitive to applied electric fields, and also the frequency ranges where the effects occur are overlapping. Therefore, examining vestibular electric fields could provide valuable information for the development of human exposure limits, which are presently based on the threshold electric fields for phosphenes (ICNIRP, 2010; IEEE, 2019) . Additionally, because the quasistatic field assumption implies linearity of the electric field solution, the results can be scaled to other stimulus intensities and extended for examination of other EVS responses. For example, Severac Cauquil et al. (2003) reported causing ocular responses with 0.1 mA GVS and a half as strong vestibular field.

This study has some limitations. Firstly, the original aim was to see if there exists a threshold point where EVS begins to interfere with balance by using a gradually increasing current or decreasing frequency. No such point was found, but it is possible the current amplitude increased or frequency decreased too quickly and detection of the threshold could require a slower rate change or even a di fferent type of waveform . For the same reason, the detection of the highest frequencies in the response was limited to visual inspection of the CoP signal spectrograms instead of more precise methods . Secondly, the results of the dosimetric analysis are influenced by the number of segmented tissues and chosen conductivity values . Data on the dielectric properties for the fluids in the inner ear are limited ( Parazzini et al . , 2006 ) . Here , they were estimated to have conductivity similar to cerebrospinal fluid and the vestibular system was modeled as one homogeneous structure . Lastly, future work should include a greater number of subj ects in order to examine how anatomical differences , such as skull thickness , affect the induced vestibular electric field and the evoked postural response .

Conclusion

The effects of EVS on full-body balance were investigated using amplitude- and frequency-modulated sinusoidal current waveforms with bilateral and unilateral electrode configurations . The purpose of the study was to determine thresholds in regards to lowest current amplitudes that would elicit a measurable response and use the thresholds to calculate the in situ electric field induced inside the vestibular system through dosimetric analysis . Based on our results , a distinct threshold point did not exist . Instead, the stimulation appeared to affect balance almost instantly and slowly increasing current amplitude simply magnified the response , increasing the si ze of postural sway at the tested frequencies under 10 Hz . Using a value of 0 . 2 mA as an estimate for stimulus current that produced a noticeable effect across all subjects and frequencies, the strength of the electric field in the vestibular system was 20- 60 mV/m. These values were calculated using high resolution, personalized computational head models of multiple volunteer subjects with detailed vestibular systems. The results of the study provide realistic and scalable approximations for the vestibular electric field, which can be used to further investigate the working mechanisms of EVS and human balance.

Figure 12 During a measurement, a subject stands on a forceplate while electrical transcranial stimulation is used to noninvasively stimulate the vestibular system.

Acknowledgements

This work was partially supported by the Academy of Finland under Grant no. 325326.

References

Aw, S., Todd, M., Aw, G., Weber, K., and Hal- magyi, G. (2008) . Gentamicin vestibulotoxicity impairs human electrically evoked vestibulo-ocular reflex. N-e urology, 71 (22) :1776- 1782.

Bresciani, J.-P., Blouin, J., Popov, K. , Sarlegna, F., and Bourdin, C. (2002) . Vestibular signals contribute to the online control of goal-direct ed. Curr. Psychol. Cogn., 21 (2-3) :263 - 280.

Cauquil, A. S., Gervet, M. F. T., and Ouaknine, M. (1998) . Body response to binaural monopolar galvanic vestibu lar stimulat ion in humans. Ne -uro- science letters, 245 (1) :37 - 40.

Coats, A. (1972) . The sinusoidal galvanic bodysway response. Acta Oto- laryngol., 74 (1- 6) :155- 162. Coats, A. C. (1973) . Effect of varying stimulus parameters on the galvanic body-sway response. Ann. Otol. Rhinol. Laryngol., 82 (1) :96- 102.

Cosgrove, D. and Zallocchi, M. (2014) . Usher protein functions in hair cells and photoreceptors. Int. J. Biochem., 46:80- 89.

Dakin, C. J., Luu, B. L., van den Doel, K. , Inglis, J. T., and Blouin, J.-S. (2010) . Frequency - specific modulation of vestibular-evoked sway responses in humans. J. Neurophysiol., 103 (2) :1048- 1056.

Dalton, B. H., Blouin, J.-S., Allen, M. D., Rice, C. L., and Inglis, J. T. (2014) . The altered vestibular-evoked myogenic and whole-body postural responses in old men during standing. Exp. Gerontol., 60:120- 128.

Dlugaiczyk, J., Gensberger, K. D., and Straka, H. (2019) . Galvanic vestibular stimulation: from basic concepts to clinical applications. J. Neurophysiol., 121 (6) :223 7- 2255.

Fitzpatrick, R. C. and Day, B. L. (2004) . Probing the human vestibular system with galvanic stimulation. J. Appl . Physiol., 96(6) :2301- 2316.

Gensberger, K. D., Kaufmann, A.-K., Dietrich, H., Branoner, F., Banchi, R., Chagnaud, B. P., and Straka, H. (2016) . Galvanic vestibular stimulation: cellular substrates and response patterns of neurons in the vestibulo-ocular network. J. Neurosci., 36(35) :9097- 9110.

Goldberg, J. M. , Smith, C. E., and Fernandez, C. (1984) . Relation between discharge regularity and responses to externally applied galvanic currents in vestibular nerve afferents of the squirrel monkey. J. Neurophysiol., 51 (6) :1236 - 1256.

Hannan, K. B., Todd, M. K., Pearson, N. J., Forbes, P. A., and Dakin, C. J. (2021) . Vestibular attenuation to random-waveform galvanic vestibular stimulation during standing and treadmill walking. Sci. Rep. , 11 (1) : 8127. lacono, M. I., Neufeld, E., Akinnagbe, E., Bower, K. , Wolf, J., Vogiatzis Oikonomidis, I., Sharma, D., Lloyd, B., Wilm, B. J., Wyss, M., Pruessmann, K. P., Jakab, A., Makris , N. , Cohen, E. D., Kuster, N., Kainz, W., and Angelone, L. M. (2015) . Mida: A multimodal imaging-based detailed anatomical model of the human head and neck. PLoS One, 10 (4) : 1-35.

ICNIRP (2010) . Guidelines for limiting exposure to time-varying electric and magnetic fields (1 Hz to 100 kHz) . Health Phys., 99 ( 6) : 818-836.

IEEE (2019) . IEEE standard for safety levels with respect to human exposure to electric, magnetic, and electromagnetic fields, 0 Hz to 300 GHz. IEEE Std 095.1-2019 (Revision of IEEE Std C95.1-2005/ Incorporates IEEE Std C95.1-2019/Cor 1-2019) , pages 1-312.

Iwasaki, S., Yamamoto, Y., Togo, F., Kinoshita, M., Yoshifuji, Y., Fujimoto, C., and Yamasoba, T. (2014) . Noisy vestibular stimulation improves body balance in bilateral vestibulopathy. Neurology, 82 (11) : 969-975.

Kremer, H., van Wijk, E., Marker, T., Wolfrum, U., and Roepman, R. (2006) . Usher syndrome: molecular links of pathogenesis, proteins and pathways. Hum. Mai. Genet., 15 (suppl_2) : R262-R270.

Laakso, I. and Hirata, A. (2012) . Fast multi- grid-based computation of the induced electric field for transcranial magnetic stimulation. Phys. Med. Biol., 57 (23) :7753-7765.

Lagnado, L. and Schmitz, F. (2015) . Ribbon synapses and visual processing in the retina. Annu . Rev. Vis. Sci., 1 (1) :235 - 262. Latt, L., Sparto, P., Furman, J., and Redfern, M. (2003) . The steady-state postural response to continuous sinusoidal galvanic vestibular stimulation. Gait €3 posture, 18 (2) : 64 - 72.

Mackenzie, S. W. and Reynolds, R. F. (2018) . Ocular torsion responses to sinusoidal electrical vestibular stimulation. J. Neurosci. Methods, 294:116- 121.

Nissi, J. and Laakso, I. (2022) . Magneto-and electrophosphene thresholds in the retina: a dosimetry modeling study. Phys. Med. Biol. , 67 (1) : 015001.

Norris, C. H., Miller, A. J., Perin, P., Holt, J. C., and Guth, P. S. (1998) . Mechanisms and effects of transepithelial polarization in the isolated semicircular canal. Hear. Res., 123 (1- 2) :31- 40.

Parazzini, M., Tognola, G., Franzoni, C., Grander!, F., and Ravazzani, P. (2006) . Modeling of the internal fields distribution in human inner hearing system exposed to 900 and 1800 MHz. IEEE. Trans. Biomed. Eng. , 54 (1) : 39 - 48.

Petersen, H., Magnusson, M., Fransson, P.-A., and Johansson, R. (1994) . Vestibular disturbance at frequencies above 1 hz affects human postural control. Acta Otolaryngol., 114 (3) :225- 230.

Reilly, J. P. (1989) . Peripheral nerve stimulation by induced electric currents: Exposure to timevarying magnetic fields. Med. Biol. Eng. Compyt . , 27 (2) : 101- 110.

Scinicariello, A. P., Inglis, J. T., and Collins, J. (2003) . The effects of sto chas- tic monopolar galvanic vestibular stimulation on human postural sway. J. Vestibular Res., 12 (2-3) : 77- 85.

Severac Cauquil, A., Faldon, M., Popov, K. , Day, B. L., and Bronstein, A. M. (2003) . Short-latency eye movements evoked by near-threshold galvanic vestibular stimulation. Exp. Brain Res., 148:414-418. Soldati, M. and Laakso, I. (2020) . Computational errors of the induced electric field in voxelized and tetrahedral anatomical head models ex-posed to spatially uniform and localized magnetic fields. Phys. Med. Biol., 65 (1) :015001.

Thomas, C., Truong, D., Clark, T. K. , and Datta, A. (2020) . Understanding current flow in galvanic vestibular stimulation: A computational study. In 42nd Annual International Conference of the IEEE Engineering & Medicine & Biology Society (EMBC) , pages 2442- 2446.

Wardman, D. L., Day, B. L., and Fitzpatrick, R. C. (2003a) . Position and velocity responses to galvanic vestibular stimulation in human subjects during standing. J. Physiol., 547 (1) :293- 299.

Wardman, D. L., Taylor, J. L., and Fitzpatrick, R. C. (2003b) . Effects of galvanic vestibular stimulation on human posture and perception while standing. J. Physiol., 551 (3) :1033- 1042.

Wuehr, M., Decker, J., and Schniepp, R. (2017) . Noisy galvanic vestibular stimulation: an emerging treatment option for bilateral vestibulopathy. J. Neurol., 264:81-86.

Yang, Y., Pu, F., Lv, X., Li, 8., Li, J., Li, D., Li, M., Fan, Y., et al. (2015) . Comparison of postural responses to galvanic vestibular stimulation between pilots and the general populace. BioMed Res. Int., 2015.

It is obvious to a person skilled in the art that with the advancement of technology, the basic idea of the invention may be implemented in various ways. The invention and its embodiments are thus not limited to the examples described above, instead they may vary within the scope of the claims.

This application claims the benefit of and priority to Finnish patent application 20247071, filed on 13 May 2024 , the contents of which are incorporated herein by reference in their entirety . Any subj ect matter disclosed in said priority application is fully included as if set forth herein verbatim and forms part of the present disclosure .

Claims

1. A method comprising:

- utilizing electrical vestibular stimulation (EVS) on a person; and

- measuring postural sway for the person.

2. The method of claim 1, wherein a waveform for the EVS is of a varying frequency.

3. The method of claim 2, wherein the waveform for the EVS is of a decreasing frequency.

4. The method of claim 2 or 3, wherein the waveform for the EVS is at least 60 seconds of length, in particular at least 120 seconds of length.

5. The method of any preceding claim, wherein two waveforms, in particular with time-varying frequencies, are combined for the EVS.

6. The method of any preceding claim, comprising determining a frequency threshold as a highest frequency at which the EVS affects the postural sway.

7. The method of claim 6, wherein the frequency threshold is 5-6 Hz or above, in particular 10 Hz or above .

8. The method of any preceding claim, comprising evaluating the effects of the EVS on the postural sway through changes in center of pressure (CoP) and/or linear acceleration.

9. The method of any preceding claim, wherein the EVS is performed with left- and/or right-sided unilateral stimulation.

10 . The method of any preceding claim, wherein the measuring is performed utilizing a force platform, such as a force plate , and/or one or more wearable sensors , such as linear acceleration sensors .

11 . The method of any preceding claim, wherein sinusoidal alternating current is used for the EVS .

12 . A method of electrical vestibular stimulation to induce and measure postural sway in a person, comprising : a . providing a stimulator ; b . providing at least two stimulation electrodes connected to the stimulator, wherein each stimulation electrode is non-invasive ; c . providing means for measuring postural sway, wherein the means for measuring postural sway measure directional acceleration or pressure ; d . placing the at least two stimulation electrodes near or over the left and right mastoid processes of the person, and optionally placing one stimulation electrode on the midline of the head of the person and/or the back of the neck of the person ; e . connecting the person to the means for measuring postural sway; f . applying stimulation to the at least two stimulation electrodes via the stimulator for a time period, thus inducing postural sway, wherein the stimulation is a frequency-modulated waveform, in particular a frequency- modulated sinusoidal alternating current ; and g . measuring, during the time period of the stimulation, a response comprising a time series measurement of the pressure or acceleration detected by the means for measuring postural sway .

13 . The method of claim 12 , wherein at least three stimulation electrodes are provided, and two are placed near or over both left and right mastoid processes of the person and one on the midline of the head of the person and/or the back of the neck of the person .

14 . The method of claim 12 or 13 , wherein the means of measuring postural sway comprise a force platform, comprising at least four pressure sensors , whereby the force detected by the four pressure sensors can be measured when the person stands on the force platform .

15 . The method of claim 14 , wherein a foam pad is provided and placed on the force platform, whereby the person would stand on the foam pad, which is on the force platform .

16 . The method of claims 12 - 15 , wherein the means of measuring postural sway comprise one or more wearable sensors , which comprise or are acceleration sensors , which may be placed on one upper arm of the person and/or on the top of the head of the person, whereby directional acceleration can be measured .

17 . The method of claim 16 , wherein the one or more wearable sensors comprises or consists of two wearable sensors , which may be placed on one upper arm of the person and on the top of the head of the person .

18 . The method of any of claims 12 - 17 , wherein the time period of the stimulation is in the range of 60 - 180 s , or the time period is 120 s , or the time period is split into two sub-periods of 0 - 60 s and 60

- 120 s.

19. The method of any of claims 12-19, wherein the frequency-modulated waveform has a peak amplitude in the range of 0.2 mA - 1.5 mA, or 0.2 mA - 1 mA, or 0.2 mA - 0.75 mA, or 0.2 mA - 0.5 mA, or 0.5 mA - 1.5 mA, or 0.5 mA - 1 mA, or 0.5 - 0.75 mA, or the peak amplitude is 0.2 mA, or 0.5 mA, or 0.75 mA, or 1.0 mA, or 1.5 mA .

20. The method of any of claims 12-20, wherein the frequency-modulated waveform has a frequency in the range of 10 - 0.1 Hz, or 10 - 2 Hz, or 2 - 0.1 Hz, or 4

- 12 Hz, or 6 - 14 Hz, or 2 - 6 Hz over the time period.

21. The method of any of claims 12-20, wherein the frequency-modulated waveform has a linearly decreasing frequency in the range of 10 - 0.1 Hz, or 10 - 2 Hz, or 2 - 0.1 Hz over the time period.

22. The method of claim 21, wherein the frequency-modulated waveform has a linearly increasing frequency between two frequencies selected from: 2 - 10 Hz, 4 - 12 Hz and 6 - 14 Hz over a sub-period followed by a linearly decreasing frequency between the same two frequencies 10 - 2 Hz, 12 - 4 Hz and 14 - 6 Hz respectively over a sub-period.

23. The method of claim 22, wherein the stimulation comprises a combination waveform which comprises two frequency-modulated waveforms, in particular two frequency-modulated sinusoidal alternating currents, that have linearly increasing or decreasing frequencies that vary in frequence in the range of 2 - 6 Hz or 6 - 2 Hz respectively over two sub-periods.

24. The method of any of claims 12-23, wherein the stimulation further comprises 0.5 s rectangular DC- pulses before and after the frequency-modulated waveform.

25 . The method of any of claims 12 -24 , wherein the threshold of the frequency or peak amplitude that induces a response in the person is calculated .

26 . The method of any of claims 12 -25 , wherein three stimulation electrodes are provided, and two are placed near or over both left and right mastoid processes of the person and one on the midline of the head or the back of the neck of the person, wherein stimulation is applied to the two stimulation electrodes over the mastoid processes (bilateral ) , to the stimulation electrode over the left mastoid process and the stimulation electrode on the midline of the head or the back of the neck ( left-sided unilateral ) , or to the stimulation electrode over the right mastoid process and the stimulation electrode on the midline of the head or the back of the neck ( right-sided unilateral ) .

27 . The method of any of claims 12 -26 , wherein the stimulation of step f is repeated, with an initial step f followed by one or more subsequent steps f , and measurement of step g is repeated, with an initial step g followed by one or more subsequent steps g, optionally with a break between repetitions .

28 . The method of claim 26 and 27 , wherein when the stimulation of steps f is repeated, during each of the one or more subsequent steps f the stimulation is applied to the same stimulation electrodes as during the initial step f or to stimulation electrodes independently selected from : bilateral , left-sided unilateral , or right-sided unilateral .

29 . The method of claim 27 or 28 , wherein in the initial step f and each of the one or more subsequent steps f , the frequency-modulated alternating waveform has the same or different frequency and amplitude in at least one subsequent step f .

30. The method of claim 27, 28, or 29, wherein during the initial step f and/or at least one of the one or more subsequent steps f the stimulation is a sham signal, whereby no frequency- modulated waveform is provided .

31. An apparatus arranged for carrying out the method of any preceding claim.

32. The apparatus of claim 31, comprising an electrical vestibular stimulation system, in particular at least two stimulation electrodes and a stimulator.

33. The apparatus of claim 31 or 32, comprising a postural sway detector, in particular a force platform comprising at least four pressure sensors and/or one or more wearable acceleration and/or voltage sensors.

34. The apparatus of claims 31-33, wherein the one or more wearable acceleration and/or voltage sensors comprises or consists of two wearable acceleration and/or voltage sensors.

35. Use of the apparatus of any of claims 31- 34 to measure postural sway.