1 Introduction

Recently, with advances in display technologies (e.g., larger size, high definition, 4 K resolution, and stereoscopy), users can enjoy virtual experiences that provide the “sense of being there,” [1] whether at home or at an amusement parks. On the other hand, there has been an increase in opportunities to cause the symptoms that are similar to motion sickness, referred often to as visually induced motion sickness (VIMS) or cyber sickness [2, 3], which is experienced by as user during or after enjoying these virtual experiences. Stanney has shown that 88% of virtual environment participants developed VIMS when viewing virtual reality movies for an hour [4]. Thus, the current state of the virtual experience with the objective of amusement eventually becomes stressor in some instances.

The pathogenic mechanism and reason for onset of the complex symptoms are not sufficiently understood. However, one of the leading hypotheses for the pathogenic mechanism includes sensory conflict theory, which suggests that the presence of conflicts among afferent input from each sensoria (vision, equilibrium sense, somatosensory) and the subject’s experience causes the complex symptoms because of the irrelevant correction of differences in information [5, 6]. In particular, VIMS is evoked by differences in information between vision and other senses.

Allowing for the development of the complex symptoms caused by sensory conflict, we can set the hypothesis that positive correcting differences in information among afferent input from each sensoria lead to the suppression of these symptoms induced by sensory conflict. This hypothesis is easy to understand viscerally, and humans may be able to display behavior like positive correcting sensory conflict subconsciously. However, there is almost no scientific verification to demonstrate this human behavior.

Our study group has a system that gives artificially created independent information, which assumes the feeling of acceleration or obliquity to the vision and vestibular-labyrinth system. Thus, this system has the capability to provide the condition of sensor conflict or that of accordance artificially. The input to vision, such as visual information like motion in a movie (amount of movement, motion in the direction) is controlled by computer graphics software. For the input to the vestibular-labyrinth system, the method of delivering electrical stimulus to the vestibular-labyrinth system from body surface is utilized. This method is called galvanic vestibular stimulation (GVS), and provides a sense of obliquity or that of constant period fluctuation in the side direction. Our system creates quantitative condition of sensor conflict/accordance because it simultaneously provides input to both the vision and vestibular-labyrinth system. In this study, as fundamental verification of above hypothesis, we verified the effect on the human body by using our artificially providing the condition of sensor conflict/accordance system in case of stimulating only vision, only GVS, and vision and GVS, simultaneously.

2 Materials and Methods

2.1 Galvanic Vestibular Stimulation (GVS)

GVS (See the review of GVS [7]) is stimulation technique allows humans to perceive the sensation of acceleration or gradient by applying small electrical current behind the ear. The GVS technique is older, and has been used for over a century as a means to discover and understand the function of the vestibular-labyrinth system. Bohemian physiologist Johann Purkyne [8] reported that a galvanic current flowing through the head upset balance and equilibrium in his dissertation. GVS generally places electrodes on both mastoid processes, as shown in Fig. 1. Then, direct current or alternating current (AC) (<2 mA) is passed between the electrodes. Sinusoidal motion in the side (anode) direction is felt when AC is applied because GVS induces a sensation of acceleration or gradient according to the magnitude of the electrical current. This study archived GVS follows:

Fig. 1.
figure 1

Galvanic vestibular stimulation (GVS). In GVS, electrodes are placed on both mastoid processes, and a direct current or alternating current (AC) (<2 mA) is passed between electrodes. This study used GVS with an 0.25-Hz AC sine wave.

Full size image
  1. 1.

    Self-built stimulus waveform software based on LabVIEW 2016 (National Instruments, Austin, TX, USA) was used to generate the pre-input-waveform.

  2. 2.

    The pre-input waveform was inputted into an Isolator (SS-203J, NIHON KOUDEN, Tokyo, Japan) through the external power unit.

  3. 3.

    The Isolator adjusted the current value and then outputted GVS.

We perform GVS (maximum current value: 2 mA and period: 0.25 Hz) using this system. If the participant did not remain in the standing position during the GVS task, the maximum current of GVS was set at 1 mA (1 male and 1 female).

2.2 Visual Stimulation

The visual stimulation used was a movie created using 3ds Max 2015 computer graphics software (Autodesk, San Rafael, CA, USA). A screenshot of the movie that was used in this study is shown in Fig. 2. The basic construction of the movie consisted of several color balls, which were displayed at random positions, and a green cross that was shown at the center position as the point of reference.

Fig. 2.
figure 2

Screenshot of the movie and the experimental setup. In order to measure the position of the COP, participants were asked to stand on a Wii Balance Board with Romberg’s posture. (Color figure online)

Full size image

The direction of motion in the movie was along the side direction (X-direction) because the conformation of both motions directions induced visual stimulation and GVS. The motion in the movie also was sinusoidal at 0.25 Hz and was generated by moving camera-simulated ocular globes (the balls themselves did not move). The amplitude of the sinusoidal motion was set to 150 software setting.

As regards the presentation of the movie, the movie was projected onto a rear projection screen that was 150 cm in front of the participant with a domestic three-dimensional (3D) projector (EH-TW5100, Seiko Epson Corporation, Suwa, Japan). The projected movie size was 157.5 cm × 280 cm and the matrix size was 1,920 × 1,080. The participants watched the experimental 3D movies using 3D glasses (ELPGS03, Seiko Epson Corporation, Suwa, Japan) as a parallax barrier. Our previous study reported that continuously viewing this motion movie induced body sway that was in sync with the motion in the movie [9].

2.3 Procedure and Design

Ten university students (4 males and 6 females; 20–24 years, motion sickness susceptibility questionnaires-short (MSSQ-short [10]) adult score: 10.58 (average) ± 5.9 (S.D.), total score: 23.04 (average) ± 14.0 (S.D.)) who did not have vision or equilibrium problems participated in this study. Through the MSSQ-short that the participant performed before examination, we confirmed that the distribution of participants did not have sensitivity bias attributable to sensory conflict. The study was approved by the Research Ethics Committee at Gifu University of Medical Science. Written consent was obtained from the participants after the purpose and significance of the study and the nature and risk of the measurements were explained, both orally and in writing. In addition, the study was conducted in line with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

The experimental setup utilized in this study is shown in Fig. 2. We performed the experiments in a controlled environment (illuminance: 10 lx) in order to avoid any variations that were caused by visual stimuli. In order to measure the position of the center of pressure (COP) as body sway, participants stood on a Wii Balance Board (Nintendo Co., Ltd., Kyoto, Japan) with Romberg’s posture. As regards the study protocol, first, a participant stood on the Wii Balance Board with eye-opening, and watched a static (nonmoving) movie for 60 s as the pretest (control). Further, participants performed four tasks in a random sequence to avoid the order effect. Task-interval was set at more than 5 min. (Table 1).

Table 1. Types of experimental tasks.
Full size table

2.4 Analysis

During the duration of the tasks, position of the COP was continuously recorded using the Wii Balance Board and custom-built stabilometry software WiimoteLib [11]. The COP measurements were recorded at 20 Hz, which is a basic sampling setting in clinical gravimetric tests. The continuous COP data were separated at intervals of 60 s of task time in order to only use stable second-half 60 s-data for various types of analysis. Our previous study reported that the position of the COP moved in synchronization with the phase in the motion movie, when the participants watched movie with low-frequency global motion [9]. Moreover, we assumed that the effect of the GVS on the change in the position of the COP indicated the same tendency. Thus, in order to evaluate the synchronization accuracy of the phase of movie and the GVS, the COP data unit underwent a frequency analysis with a fast Fourier transform with a Hamming window along both the X- and anteroposterior direction (Y-direction). Moreover, the total locus length, area, and standard deviation (S.D.) of the COP data for the two directions, which are generally indexes of body sway, were calculated. Then, a Tukey–Kramer method was performed using ORIGIN Pro 8.5 (OriginLab, Corporation, Northampton, MA, USA).

In order to evaluate heart rate variability (HRV) and RR interval variability (RRIV), a lead II ECG and thoracic movement, which is an index of respiration, were monitored and recorded (DL-320, S&ME, Tokyo, Japan). These data were separated at 60-s intervals in order to only use stable second-half 60 s-data for various types of analysis, as was the case with the COP data. Then, we calculated heart rate (HR), normalized low-frequency component (LF, 0.04–0.15 Hz)/high-frequency component (HF, 0.15–0.40 Hz), Normalized HF and Total power by using Memcalc/win (Ver.1.2, GMS, Tokyo, Japan), which is a time series analysis software that uses the maximum entropy method. Then, the Tukey–Kramer method was performed out using ORIGIN Pro 8.5.

As for the subjective measurements, motion sickness or VIMS symptoms were measured. The participants completed a simulator sickness questionnaire (SSQ) [12], which has been used in a number of previous studies, after each task. The total score and three subscores (nausea, oculomotor discomfort, disorientation,) were calculated for each task (See for the SSQ calculation methods [12]). Then, the Scheffe’s multiple comparison was performed using ORIGIN Pro 8.5.

3 Results

3.1 Stabilometry

A typical stabilogram result calculated from the second-half 60 s-COP data (22 years old, male) is shown in Fig. 3a–d. The stabilogram indicated a continuous change along the side direction. Comparison of the Movie task (Fig. 3b) and other tasks (Fig. 3a, c, and d) show that the COP size in the side direction in the tasks with GVS was wider than that without GVS. Moreover, the change in the COP in the side direction of the GVS task with movie (Fig. 3c and d) increased than that without movie (Fig. 3a). A significant difference between the In phase task and the Reversed phase task was not identified in the stabilogram.

Fig. 3.
figure 3

Stabilogram result of the second-half 60 s-COP. (22 years old, male). (a) GVS task, (b) Movie task, (c) Reversed task, (d) In phase task.

Full size image

The index of the COP (body sway) in each task was calculated, as shown in Fig. 4a–e. Figure 4a–d show the results for body sway indexes calculated from the second-half 60 s-COP; Fig. 4e shows the body sway indexes for power spectral density (PSD) at 0.25 Hz, which were obtained from the frequency analysis using the second-half 60 s-COP. The all body sway indexes and the PSD at 0.25 Hz show similar results. First, the In phase task indicated the highest index value among all tasks, and the value of the Reversed phase task was the second highest. Second, compared the result for In phase task with that for the Movie task, the index values of the In phase showed significantly higher than that of the Movie task. Third, comparison of result for S.D. in the X-direction and in the Y-direction showed that body sway spread not only in the X-direction but also in the Y-direction owing to the effect of tasks; however, each task had an effect on the body sway only along the X-direction.

Fig. 4.
figure 4

Body-sway index of the each task and the PSD at 0.25 Hz. (a) Total locus length, (b) Area, (c) S. D. at X-direction, (d) S.D. at Y-direction, and (e) PSD at 0.25 Hz.

Full size image

3.2 HRV and RRIV

Figure 5a–d show the results for the HRV and the RRIV using of the second half. No significant changes were observed. Moreover, a consistent tendency that could be attributed to the difference in the type of task was not also observed.

Fig. 5.
figure 5

The results of the HRV and RRIV in each task. (a) Hart Rate, (b) Normalized LF/HF, (c) Normalized LF/HF, and (d) Total power.

Full size image

3.3 SSQ

For the SSQ results (total score and three subscores), a Scheffe’s multiple comparison as shown in Fig. 6a–d. All SSQ scores indicated the same tendencies, regardless of kind of SSQ score. First, score of the In phase task was the highest, while, that of the Movie task was the smallest. Second, both scores of the In phase task and Reversed task were significantly higher than that of the Movie task except for the Oculomotor score. Third, both the In phase task and the Reversed task had similar average SSQ score.

Fig. 6.
figure 6

The results of Simulator sickness questionnaire (SSQ). (a) Nausea (Subscore), (b) Oculomotor (Subscore), (c) Disorientation (Subscore), and (d) Total score.

Full size image

4 Discussion

4.1 Body Sway

We compared the GVS task with the Movie task for the body sway test (Fig. 4) and found that instability and synchronization accuracy of the GVS task was higher than that of the Movie task. The reason was attributed to between relationship the current setting of GVS and motion settings in the experimental motion movie as follows. The effect of the GVS task (2 mA current setting) was higher than that of the Movie task (150 software setting). Thus, the body sway in both the In phase task and Reversed phase task was mainly due to the GVS. We compared the In phase task and the Reversed phase task and found that the indexes of the body sway in the In phase task was higher than that in the Reversed task, although no significant difference was found. We considered that the vision stimulation was treated as accessory component on body sway in this study. In the In phase task, vision stimulation worked as adding component toward GVS. By contrast, that in the Reversed task worked as subtraction component toward that because both stimulations impinged on body sway, mutually. All experimental settings, such as the current of GVS, movie contents, viewing condition, and difference in phases, were factors of change in body sway. Therefore, we assumed that adjustment of these settings enabled the control of changes in the body sway freely.

S.D. (standard deviation), which is a body sway index, shows instability for body sway along x- and y-directions. In this study, GVS and visual stimulation were designed to cause periodic body sway along the side direction. However, the result of this study also showed a significant increase in S.D. along the Y-direction. Therefore, we assumed that instability of body sway was not independent in each direction, and it had a reciprocal relationship.

The all-average value of indexes of body sway and all average SSQ scores indicated the same tendency (Figs. 4 and  6, respectively). On the other hand, no significant difference was found between SSQ scores in the In phase task and that of Reversed task. We believe that the In Phase task increased in intensity of the sensory conflict (incense in posture gap between static upright posture and posture during the In phase task), and the Reversed phase task decreased in that (decrease in posture gap from that). However, a significant difference was not found. Therefore, it was assumed the Reversed task had other factors of the sensory conflict that were not considered in this study.

4.2 HRV and RRIV

In the analysis of HRV and RRIV, the results did not show a significant increase/decrease. Marco et al. reported that LF/HF power ratio and total power increased in a stressful environment [13]. In addition, Abe et al. and Bonnet et al. reported that evaluation of autonomic nervous system such as HF, LF/HF power ratio was useful in detection of condition of the VIMS or motion sickness as subjective detection method [14, 15]. However, significant changes in conditions from the control condition (upright posture with opening eye) in each task were not found from the HRV and RRIV analysis. Considering previous studies, the results of HRV and RRIV indicated that the task did not cause a change in the condition regardless of the kind of task. However, the SSQ scores in both the In phase task and Reversed task increased significantly, as compared to the Movie task. Thus, the disruption between subjective and objective evaluations was recognized. Previous studies have shown that the HR and LF component in the upright posture is relatively higher than that in sitting posture. Considering previous studies, it was assumed that effect of the task was overshadowed by the effect of difference in the posture. Therefore, usefulness of the detection of symptoms induced by the sensory conflict using HRV and RRIV requires further verification with changing study conditions such as posture.

5 Conclusion

In this study, we verified the hypothesis that positive correcting differences in information among afferent input from each sensoria lead to the suppression of various symptoms induced by the sensory conflict. In addition, we verified effect on the human body in case where only vision, only GVS, and vision and GVS were stimulated simultaneously. The following conclusions can be drawn:

  1. 1.

    The instability of body sway and the synchronization accuracy increased in the following order: Movie < GVS < Reversed phase < In phase. Moreover, the result of SSQ score changed in the following order: Movie < GVS < Reversed phase = In phase. Therefore, the disruption between subjective and objective evaluation was recognized. It was assumed that the Reversed task had other factor of the sensory conflict, which was not considered in this study.

  2. 2.

    From HRV and RRIV analysis, significant changes between each task were not confirmed. In addition, there was disruption between the results of subjective and objective evaluation. Therefore, further verification is required with changing study conditions and settings such as posture.