## Abstract

A periodically reversing optic flow animation, experienced while standing, induces an involuntary sway termed visually induced postural sway (VIPS). Interestingly, VIPS is suppressed during light finger touch to a stationary object. Here, we explored whether VIPS is mediated by parietal field activity in the dorsal visual stream as well as by activity in early visual areas, as has been suggested. We performed a mobile brain/body imaging study using high-density electroencephalographic recording from human participants (11 men and 3 women) standing during exposure to periodically reversing optic flow with and without light finger touch to a stable surface. We also performed recording their visuo-postural tracking movements as a typical visually guided movement to explore differences of cortical process of VIPS from the voluntary visuomotor process involving the dorsal stream. In the visuo-postural tracking condition, the participants moved their center of pressure in time with a slowly oscillating (expanding, shrinking) target rectangle. Source-resolved results showed that alpha band (8–13 Hz) activity in the medial and right occipital cortex during VIPS was modulated by the direction and velocity of optic flow and increased significantly during light finger touch. However, source-resolved potentials from the parietal association cortex showed no such modulation. During voluntary postural sway with feedback (but no visual flow) in which the dorsal stream is involved, sensorimotor areas produced more theta band (4–7 Hz) and less beta band (14–35 Hz) activity than during involuntary VIPS. These results suggest that VIPS involves cortical field dynamic changes in the early visual cortex rather than in the posterior parietal cortex of the visual dorsal stream.

## INTRODUCTION

Maintenance of a bipedal standing posture is one of the fundamental motor skills of humans. Corrective responses in postural control are informed by visual as well as proprioceptive and vestibular information. Visual motion signals in peripheral vision (optic flow [OF]), experienced as a consequence of movements in head position and orientation (Gibson, 1979), affect postural control (Lee & Aronson, 1974). Periodically reversing forward/backward movement of the visual scene at low frequencies (0.02–0.2 Hz) induces an involuntary postural adjustment governed by ankle planter/dorsiflexion (Lestienne, Soechting, & Berthoz, 1977) typically called “visually induced postural sway” (VIPS). An interesting multimodal interaction is that maintaining a light touch of a fingertip (force < 1 N) to a stable object during standing (“light finger touch”) reduces the sway during VIPS, measured as movement of the center of pressure (COP; Jeka, Oie, & Kiemel, 2000; Jeka & Lackner, 1994). For the conflict between the visual motion and leg proprioceptive feedback, the proprioceptive feedback produced by light touch can enhance the certainty of the proprioceptive feedback. On the basis of a theoretical framework of optimal estimation to minimize the uncertainty (van der Kooij, Jacobs, Koopman, & van der Helm, 2001; van der Kooij, Jacobs, Koopman, & Grootenboer, 1999), the suppression of VIPS with light touch implies involvement of the sensory integration process that enhances the weight of the estimate from somatosensory feedback and reduces the weight of visual motion for the self-motion perception. However, few studies have reported on brain dynamics during VIPS suppression by light finger touch.

The subcortical neural networks of spinal cord, brain stem, and cerebellum play a vital role in postural control. Vestibular nuclei, which receive the vestibular, proprioceptive, and visual motion signals (Cullen, 2012), are one of possible sites to mediate multimodal signals for postural control. However, VIPS cannot be explained by the activity within the vestibular nuclei alone because the vision-sensitive neurons influence eye movement but not postural control (Bryan & Angelaki, 2009). Another possible site is cerebellar vermis (Colnaghi, Honeine, Sozzi, & Schieppati, 2017; Ouchi, Okada, Yoshikawa, Nobezawa, & Futatsubashi, 1999), which receives visual input from visual cortices and has projections to not only spinal descending pathways but also various cortical regions including motor-related areas (Baumann et al., 2015). In cortical level, the parietal regions are responsible for multimodal sensory integration and the activity of pyramidal tract neurons in the motor cortex contributes to the postural control (Beloozerova, Sirota, Orlovsky, & Deliagina, 2005). Because the vestibular nuclei and cerebellar vermis can communicate with various cortical regions, it should be useful for improving our understanding of multimodal integration to clarify cortical contribution reflected in its dynamics.

A possible model accounting for how and why VIPS is suppressed by light finger touch is the “parietal sensory integration model” involving activity in the dorsal stream (from visual cortex to parietal association cortex to motor cortex; Takakusaki, 2017; Nassi & Callaway, 2009). The middle temporal area, middle superior temporal area, and V6 are involved in inferring self-motion from OF (Gramann et al., 2011; Cardin & Smith, 2010; Pitzalis et al., 2010; Morrone et al., 2000; Probst, Plendl, Paulus, Wist, & Scherg, 1993). The ventral intraparietal area includes bimodal neurons that respond to both visual and somatosensory inputs (Schlack, Hoffmann, & Bremmer, 2002; Graziano & Gross, 1993). These areas are thus thought to process multimodal sensory integration for self-motion perception (Cardin & Smith, 2011; Holmes & Spence, 2004). In addition, both areas have connections with motor areas (Kruse, Dannenberg, Kleiser, & Hoffmann, 2002; Tanné-Gariépy, Rouiller, & Boussaoud, 2002). Thus, the dorsal stream could play a critical role in VIPS.

In a second possible model, the “occipital sensory integration model,” the primary visual cortex, motor cortex, and subcortical motor areas work together to support VIPS. Evidence shows that firing rates of primary visual cortical neurons of mice walking on a treadmill depend not only on OF speed but also on the speed of locomotion (Saleem, Ayaz, Jeffery, Harris, & Carandini, 2013). Roth et al. (2016) suggested that the locomotor information is conveyed to the cortex from the dorsolateral geniculate nucleus and lateral posterior nucleus of the thalamus. Thus, the early visual cortex may integrate both visual and sensorimotor signals to support self-motion perception. The sensory integration in early stage of the visual process is beneficial for rapid reaction. Actually, Nashner and Berthoz (1978) showed that the visual inputs influence involuntary postural adjustment within 100 msec, which is shorter than typical voluntary motor responses. A strength of the occipital sensory integration model is that omitting parietal sensory integration processes reduces the need for long-range cortical communication and cooperation required in the parietal sensory integration model.

In this study, we designed an electroencephalographic (EEG) experiment to determine whether the parietal sensory integration model or the occipital sensory integration model is better supported by the cortical brain dynamics measured during VIPS and control conditions. Mobile brain/body imaging (MoBI) is an approach to understanding neural processes supporting motivated actions by simultaneous recording of dense-array EEG and body and/or eye movements, particularly those accompanying recorded changes in the environment (Gramann et al., 2011; Makeig, Gramann, Jung, Sejnowski, & Poizner, 2009). Advanced EEG signal processing methods including independent component analysis (ICA; Makeig, Bell, Jung, & Sejnowski, 1996) and cortical source localization of ICA-derived effective EEG sources (Makeig et al., 2002) allow distinguishing cortical signals from extracranial artifact sources, such as blinking, eye movements, neck muscle activities, and so forth (Gramann, Gwin, Bigdely-Shamlo, Ferris, & Makeig, 2010; Jung et al., 2000), and can also separate functionally distinct cortical field activities arising in different cortical effective source areas (Delorme, Palmer, Onton, Oostenveld, & Makeig, 2012). The equivalent dipole source of each independent component (IC) is localized by using boundary element method with a neuroelectromagnetic head model (Akalin Acar & Makeig, 2010; Oostenveld & Oostendorp, 2002). On the basis of the MoBI approach, Gwin and Ferris (2012) found significant coherence between the electromyographies (EMGs) of ankle and knee muscles and ICs localized in the contralateral motor area during voluntary leg joint movements. Solis-Escalante et al. (2019) showed cortical contributions for postural recovery responses from loss of balance with MoBI approach. During the recoveries for rapid movements of support surface, ICs in motor-related cortical areas were modulated according to the strategies of stepping or feet in place.

Power spectral densities of brain-effective source components resolved by ICA decomposition of the continuous EEG data were evaluated to explore the brain dynamics accompanying visual stimuli. Modulations of the EEG power spectrum in specific frequency bands are associated with characteristic mental conditions. For example, increased (8–13 Hz) alpha band power in the occipital region indicates deactivated or idling states of visual processing (Pfurtscheller, Stancák, & Neuper, 1996), often associated with decreases in visual cortex blood oxygen level dependent activity in fMRI experiments (Moosmann et al., 2003; Goldman, Stern, Engel, & Cohen, 2002). In sensorimotor areas, (14–35 Hz) beta band activity is attenuated before and during movements (Zaepffel, Trachel, Kilavik, & Brochier, 2013; Tzagarakis, Ince, Leuthold, & Pellizzer, 2010; Alegre et al., 2003; Jasper & Penfield, 1949), but (4–7 Hz) theta band activity increases occur during cognitively controlled movements (Cavanagh & Frank, 2014). For example, Sipp, Gwin, Makeig, and Ferris (2013) reported that theta band power increased in several cortical areas (anterior cingulate, parietal, left, and right sensorimotor areas) during loss of balance in balance beam walking when the brain cognitive and motor system needs to plan and execute quick corrective movements.

Ishigaki, Ueta, Imai, and Morioka (2016) showed that light finger touch to a stable object was accompanied by reduced alpha (mu) band power in sensorimotor and parietal cortices during quiet standing. In their experiments, neither touching an unstable object nor simply attending to the same fingertip was as effective in improving postural balance or reducing alpha power. In the current study, we compared physical sway and EEG power modulation patterns during VIPS with and without light finger touch to a stable surface. If VIPS suppression by finger touch is correlated with an EEG spectral power modulation difference in the parietal cortex, this fact would support the parietal sensory integration model. Otherwise, if the EEG spectral power modulation condition difference arises in occipital areas, this would support the occipital sensory integration model. We also included a voluntary visuo-postural tracking condition known to engage the dorsal visual stream (Hatzitaki & Konstadakos, 2007) to determine how the neural process of VIPS is different from the typical visuomotor control.

## METHODS

### Participants

Fourteen participants (age: 32 ± 8.6 years, height: 170 ± 9.0 cm, weight: 64 ± 9.7 kg, 11 men and 3 women) participated in this experiment. In general, the sample size in the EEG studies in references ranges around 10–20. The mean (SD) of the sample size in the EEG studies in the references were 17.4 (11.7). Written informed consent was obtained from all participants, and the experimental design was approved by an institutional review board at the University of California San Diego. One of the participants reported Charcot–Marie–Tooth disease, but we found no abnormalities of the behavioral and EEG data for this participant and so included their data in our analysis. No other participant reported any neurological or musculoskeletal disease. Although preliminary trials with exposure of OF stimuli were performed as a screening test of motion sickness, no participants reported bad feeling and declined to participate. The Edinburgh handedness questionnaire was used to determine that all participants were right-handed. Eleven participants reported that they preferred to use their right leg to kick a ball, two participants reported they used both right and left legs equally, and one participant reported that they preferred to use the left leg.

### Apparatus

As shown in Figure 1, participants stood in front of a screen (1.08 m high and 1.58 m wide) on which the visual stimuli were projected (with a PDG-DWL100, Sanyo Electric Co. Ltd.) in a dark room. The resolution of the projector was 1024 × 768 pixels. The distance from screen to eye was 1.0 m, and the viewing angles of the screen were 76.6° in the horizontal direction and 53.1° in the vertical direction. The 3-D graphics were generated by the Panda3D graphic engine on the Simulation and Neuroscience Application Platform (github.com/sccn/SNAP), an open-source platform for presenting visual stimuli and sending event markers in psychophysical experiments.

Figure 1.

(A) Experimental setup. Participants stood in front of a screen on which visual stimuli were presented. A tripod was placed to the right side of the participant to provide a stationary platform on which to maintain light finger touch; a sensor recorded the touch pressure. Participants wore a 128-channel EEG electrode cap and a 32-channel EMG legging. A force plate measured postural sway during standing. (B) EMG legging and locations of the electrodes used to evaluate the activity of the left gastrocnemius muscle.

Figure 1.

(A) Experimental setup. Participants stood in front of a screen on which visual stimuli were presented. A tripod was placed to the right side of the participant to provide a stationary platform on which to maintain light finger touch; a sensor recorded the touch pressure. Participants wore a 128-channel EEG electrode cap and a 32-channel EMG legging. A force plate measured postural sway during standing. (B) EMG legging and locations of the electrodes used to evaluate the activity of the left gastrocnemius muscle.

Scalp EEG (128 channels) and EMG (32 channels over the left ankle plantar flexor and dorsiflexor muscles), plus an analog signal from a pressure sensor on the finger touch platform, were recorded simultaneously at 2048 Hz/channel (Active-Two, BioSemi). This EEG recorder uses common-mode sense and driven left leg, which serves as conventional ground and reference electrodes interchangeably. Technical details can be found in the manufacture site www.biosemi.com/faq/cms&drl.htm. These electrodes were located near vertex (Cz) in the current application. Because the VIPS in this study is a bilaterally symmetric movement, the muscle activities of only the left leg were measured. EEG and EMG signals were rereferenced, before further analysis, to their respective signal averages. The analog pressure sensor signal was used to confirm that finger touch force was less than 1 N during “light finger touch” (and was ∼0 N in other conditions). Before recording began, the locations of the EEG and EMG electrodes relative to skull and leg landmarks, respectively, were digitized using an electrode positioning system (ELPOS, Zebris Medical). Of the 32 recorded EMG channels, activities of the gastrocnemius muscle were derived by differentiating potentials (bipolar derivation) from two electrodes shown in Figure 1. Ground reaction force and moment were measured by a force plate (Accusway, AMTI), sampled at 200 Hz. The time course of participant COP was derived from the recorded force and moment by using the equilibrium equations of moment acting on the force plate. Event markers for the visual stimuli, as well ground reaction force, EEG, EMG, and finger touch sensor data, were synchronized using the open-source Lab Streaming Layer data acquisition and synchronization framework (available at github.com/sccn/labstreaminglayer).

Before beginning the experiment, each participant was asked to indicate their subjectively perceived direction of “straight ahead” to define the center of the participant's visual field. Figure 2 shows a background image of a virtual 3-D (3 m × 3 m) corridor. The image was presented as a 2-D projection on the screen with a distance from the participant of 1.0 m. The height of the eyes above the virtual floor was 1.6 m. The vanishing point of the virtual corridor was adjusted to be at the subjective center of the visual field, and an eye fixation point was presented at the vanishing point as a blue 2-cm square crosshair. The textures of the walls, floor, and ceiling were the same black-and-white checkerboard pattern. A gray fog with exponentially distributed density (RGB: 127, 127, 127) masked the central vision region.

Figure 2.

An image of a virtual 3-D corridor was presented on the screen. The textures of walls, floor, and ceiling were a checkerboard pattern in black and white over which a gray “fog” was applied. A blue crosshair, serving as an eye fixation point, was placed at the vanishing point correspondingto the subjective center of the visual field.

Figure 2.

An image of a virtual 3-D corridor was presented on the screen. The textures of walls, floor, and ceiling were a checkerboard pattern in black and white over which a gray “fog” was applied. A blue crosshair, serving as an eye fixation point, was placed at the vanishing point correspondingto the subjective center of the visual field.

As shown in Figure 3, three task conditions were used to examine visuomotor processing and sensory integration during VIPS. Behavior and brain activity during standing while exposed to sinusoidally varying (approaching–retreating) OF were compared to those in two other conditions: the same OF with light touch (OF-T) and visuo-postural tracking with feedback (VT-F).

Figure 3.

Experimental design. Behavior and brain activity during standing when exposed to sinusoidal OF were compared to those in two other conditions: OF-T and VT-F. In the OF-T condition, participants maintain light finger pressure with their right index finger on a stable platform surface (platform support not shown). Other task requirements were as in the OF condition. In the VT-F condition, no background OF was presented. Instead, a sinusoidally expanding/contracting target frame (red) was presented together with a moving indicator frame (black) that expanded or contracted according to the participant's current COP. Participants attempted to adjust their magnitude of forward/backward postural sway so that the size of the black indicator frame matched that of the red target frame.

Figure 3.

Experimental design. Behavior and brain activity during standing when exposed to sinusoidal OF were compared to those in two other conditions: OF-T and VT-F. In the OF-T condition, participants maintain light finger pressure with their right index finger on a stable platform surface (platform support not shown). Other task requirements were as in the OF condition. In the VT-F condition, no background OF was presented. Instead, a sinusoidally expanding/contracting target frame (red) was presented together with a moving indicator frame (black) that expanded or contracted according to the participant's current COP. Participants attempted to adjust their magnitude of forward/backward postural sway so that the size of the black indicator frame matched that of the red target frame.

#### OF-induced Sway

In the OF condition, the velocity of a visual motion was given by a sinusoidal function v(t) = a sin ωt where the frequency ω was 0.2 Hz and the amplitude a was 2 m/sec. The distance of the visual motion defined by the integral of the velocity was x(t) = a/ω(1 − cos ωt). The apparent velocity of the corridor relative to the participant in the first half of each cycle was in the forward direction away from the participant (produced by concentrating the 2-D image); in the latter half, the movement was backward (toward the participant, produced by image expansion). The presented OF was equivalent to the visual motion during sinusoidally varying head/body movements whose maximum speed was 2 m/sec. Note this was much faster than actual participant head movement produced during postural sway (<0.05 m/sec). Participants were asked to hold their right index finger above a platform placed conveniently near their wrist, without touching it.

#### OF with Stable Touch

In the OF-T condition, the participants were asked to maintain a light touch with their index finger to the top surface of a stationary platform. They were also instructed that the force applied by their finger to the platform did not exceed 1 N. Before the experiment, they performed several finger touch training trials to learn the feeling of 1 N of pressure before beginning the experiment. Other task requirements were the same as in the OF condition.

Figure 4A shows the visual presentation sequence in the OF and OF-T conditions in each block. In Phase 1, an eye fixation point was presented on a white background at the center of the participant visual field. The participants were asked to fix their gaze on this position and to minimize further eye movements. Next (Phase 2), a static image of the 3-D virtual corridor appeared on the screen. The durations of Phases 1 and 2 were randomized to between 6 and 10 sec each. EEG data acquired during Phase 2 were used as a baseline to evaluate EEG power modulation during the following conditions. During Phase 3, the participant performed the requested task during twenty-four 5-sec sinusoidal OF cycles. In the OF and OF-T conditions, participants were instructed to neither intentionally change their posture nor suppress their postural sway unless they perceived a danger of falling.

Figure 4.

The sequence of visual stimuli in each task condition. (A) The same sequence was presented in the OF and OF-T conditions. At the beginning of the trial (Phase 1), an eye fixation point was presented at screen center. Next (Phase 2), a stationary virtual 3-D corridor was displayed. The durations of Phase 1 and 2 were each randomized across a range of 6–10 sec. In Phase 3, OF with a sinusoidal pattern at 0.2 Hz was generated for 120 sec (24 visual motion cycles) by moving the virtual point of gaze in the corridor forward and backward. (B) The visual stimuli presented in the VT-F condition. Phases 1 and 2 were as in the OF and OF-T conditions. During Phase 3, a (red) target frame and a (black) COP indicator frame were presented. Target height and width were again varied sinusoidally at 0.2 Hz for 120 sec (24 cycles).

Figure 4.

The sequence of visual stimuli in each task condition. (A) The same sequence was presented in the OF and OF-T conditions. At the beginning of the trial (Phase 1), an eye fixation point was presented at screen center. Next (Phase 2), a stationary virtual 3-D corridor was displayed. The durations of Phase 1 and 2 were each randomized across a range of 6–10 sec. In Phase 3, OF with a sinusoidal pattern at 0.2 Hz was generated for 120 sec (24 visual motion cycles) by moving the virtual point of gaze in the corridor forward and backward. (B) The visual stimuli presented in the VT-F condition. Phases 1 and 2 were as in the OF and OF-T conditions. During Phase 3, a (red) target frame and a (black) COP indicator frame were presented. Target height and width were again varied sinusoidally at 0.2 Hz for 120 sec (24 cycles).

#### Voluntary Sway

In the VT-F condition, the stationary background image was the same corridor and checkerboard walls, generating no OF. A sinusoidally expanding/contracting red rectangular target frame (height: 0.1 m) was centered at the screen superimposed on the static tunnel image. The width of the red rectangle on screen was varied in proportion to R(t) = A(1 − cos ωt) + B (A = 0.05 m, B = 0.05 m). Thereby, its minimum and maximum widths were 0.05 and 0.15 m, respectively. Its sinusoidal variations in width were at the same (0.2 Hz) frequency as the sinusoidal fluctuations in OF velocity in the OF condition.

The width of the black rectangular indicator was linearly related to the current COP position along the same front–back axis. This black indicator widened as their COP moved forward with respect to the participant's base support and contracted as their COP moved backward. The participant attempted to adjust the size of the black frame to match that of the red frame by appropriately varying the position of their COP forward or backward. In this VT-F condition, participants were asked to keep their right index finger above the platform without touching it.

As shown in Figure 4B, the visual presentation sequence during Phases 1 and 2 was as in the OF condition. In Phase 3, the size-varying (red) target and (black) COP indicators were superimposed on the static checkerboard corridor image. In a fourth task condition, OF velocities were randomized to test a separate hypothesis; the data from this condition will be published elsewhere.

Before the data recording, participants practiced their tasks in the OF and VT-F conditions. Training trials in the VT-F condition were performed to reduce tracking error (Hatzitaki & Konstadakos, 2007); each participant practiced this task until they felt their tracking error had stabilized. A round consisted of four blocks, one for each task condition; in each round, the four conditions were presented in random order. Six rounds, with breaks of self-selected duration between rounds, were carried out. The total duration of the experiment across all six rounds was about 60 min.

### Data Analysis

The measured EEG and EMG signals and ground reaction force data were processed using custom MATLAB scripts (The MathWorks, Inc.) using the EEGLAB toolbox (Delorme & Makeig, 2004). Movements of the participants' COP were calculated using the force and moment acting on the force plate. The magnitude of postural sway was measured as COP displacement in the anterior/posterior direction during each cycle. The 32-channel EMG data were preprocessed by applying a high-pass filter with a cutoff frequency of 3 Hz to eliminate slow drifts; line noise was removed using a band-stop filter at 60 Hz. The filtered EMG data were rereferenced to their average value across the 32 EMG channels. EMG amplitude of the gastrocnemius muscle in each cycle was estimated by the root mean square of the EMG signal from the representative electrode channels (Figure 1B).

The EEG preprocessing procedure for each participant consisted of high-pass filtering with a cutoff frequency of 0.5 Hz and band-stop filtering at the line frequency, 60 Hz. The EEG signals were then resampled to 512 Hz to reduce processing time and data storage requirements. Noisy channels (mean = 5.2 per participant, range = [1, 10] per participant) were rejected based on visual inspection. These were channel signals with abnormally large or small signal amplitude arising from unstable electrode connections. The remaining EEG channel data were rereferenced to their average across channels. The processed EEG signals were then decomposed into IC processes using adaptive mixture ICA (Palmer, Makeig, Kreutz-Delgado, & Rao, 2008); this method was chosen because adaptive mixture ICA was shown to produce stronger mutual information reduction in multichannel EEG data than 21 other tested blind source separation algorithms (Delorme et al., 2012). The signal source of each IC process was estimated using the DIPFIT plug-in in EEGLAB (sccn.ucsd.edu/wiki/A08:_DIPFIT; Oostenveld & Oostendorp, 2002). ICs related to nonbrain (artifact) processes including EOG, EMG, and cardiac activity were manually excluded based on their equivalent dipole locations, scalp map topographies, and power spectral densities. All ICs with scalp maps having high residual variance (>15%) from the scalp projection of the best-fitting equivalent dipole were also excluded. The mean number of retained cortical ICs was 8.71 per participant (SD = ±2.02).

For group-level analysis, manually identified brain ICs from all the participants were clustered by a K-means algorithm in the STUDY framework of EEGLAB based on their dipole locations (Delorme et al., 2011). The number of clusters was set to six. BioImage Suite web (bioimagesuiteweb.github.io/webapp/) was used to transform each dipole location in Talairach coordinates (Lacadie, Fulbright, Rajeevan, Constable, & Papademetris, 2008) and specify the nearest Brodmann's area. The mean dipole location (standard deviation) and the cortical region of each IC cluster were summarized in Table 1. For each cluster, the power spectral density and stimulus event-related spectral perturbations (ERSPs) were evaluated for all conditions (Makeig, 1993). The time-locking events used in computing the ERSPs were the zero-phase moments in the sinusoidal cycling of the visual stimulus (changing OF velocity or visual target size). The EEG data recorded during the task were segmented into overlapping time epochs of two visual flow cycles (10 sec) in length centered on each zero-phase moment. The common baseline power spectrum among the task conditions for each IC was taken to be the mean IC spectrum in Phase 2, as shown in Figure 4, when neither the OF nor the visual target was presented. The baseline log mean power spectrum that was shown in Figure 8 was subtracted from the log mean spectrum at each time point in the ERSP latency window to evaluate stimulus event-related spectral power modulations (event-induced “perturbations”) with respect to baseline. The temporal changes of power in each frequency were evaluated by regression of 0.2- and 0.4-Hz sinusoidal functions that correspond to the directional and nondirectional (absolute) visual motion. The correlation may quantify how the power modulation is associated with the visual motion and/or accompanied postural behavior.

Table 1.
Results of Clustering of Brain-based ICs and Centroid Location for Each of the Identified IC Clusters
Cluster Name# Participants (of 14) and # ICsTalairach CoordinatesNearest Brodmann's Area (% ICs in the BA)Cortical Region
xyz
Occipital 14 Ss, 28 ICs 16 (15) −73 (14) 0 (10) BA 18 (50%) Visual cortex
R. parietal 14 Ss, 22 ICs 29 (12) −36 (11) 34 (14) BA 40 (27%) Parietal cortex
L. parietal 12 Ss, 23 ICs −25 (13) −49 (12) 24 (16) BA 39 (26%) Parietal cortex
Med. frontal 10 Ss, 19 ICs −2 (12) 8 (8) 46 (13) BA 6 (63%) Sensory motor cortex
L. frontal 11 Ss, 14 ICs −43 (11) −1 (19) 8 (13) BA 13 (7%) Frontal cortex
R. frontal 10 Ss, 16 ICs 36 (11) −2 (16) 19 (14) BA 13 (13%) Frontal cortex
Cluster Name# Participants (of 14) and # ICsTalairach CoordinatesNearest Brodmann's Area (% ICs in the BA)Cortical Region
xyz
Occipital 14 Ss, 28 ICs 16 (15) −73 (14) 0 (10) BA 18 (50%) Visual cortex
R. parietal 14 Ss, 22 ICs 29 (12) −36 (11) 34 (14) BA 40 (27%) Parietal cortex
L. parietal 12 Ss, 23 ICs −25 (13) −49 (12) 24 (16) BA 39 (26%) Parietal cortex
Med. frontal 10 Ss, 19 ICs −2 (12) 8 (8) 46 (13) BA 6 (63%) Sensory motor cortex
L. frontal 11 Ss, 14 ICs −43 (11) −1 (19) 8 (13) BA 13 (7%) Frontal cortex
R. frontal 10 Ss, 16 ICs 36 (11) −2 (16) 19 (14) BA 13 (13%) Frontal cortex

The values in Talairach coordinates give the mean (standard deviation) of the dipole locations. BA = Brodmann's area; L. = left; R. = right; Med. = medial; Ss = subjects.

### Statistical Analysis

To sum up, the factorial design in this study was within-participant 1 × 3 with three levels: OF, OF-T, and VT-F. We planned comparisons of mean values for the OF condition with the OF-T and VT-F conditions in behavioral and EEG data. The differences between OF and OF-T indicate the effects of light finger touch. We compared the mean values of OF with VT-F to examine the differences of VIPS from the typical visuomotor process. Because the comparison between OF-T and VT-F is not our interest, the results of this comparison were excluded from figures and tables.

To test condition COP displacement and EMG amplitude differences, average values were compared using one-way repeated-measures ANOVA and post hoc Tukey–Kramer method with a p < .05 significance level. To test condition differences of signal power in theta, alpha, and beta bands within clustered ICs, a two-way repeated-measures ANOVA (Conditions × Frequency Bands) was used. The main effects and interaction were evaluated by p value of F distribution and partial eta squared (ηp2) as the measure of effect size. The post hoc test of Tukey–Kramer method was performed. The ANOVAs and post hoc tests were performed using the Statistics and Machine Learning Toolbox for MATLAB.

To test source-resolved ERSP condition differences, cluster-mean values for the OF condition were compared with those from the OF-T and VT-F conditions, respectively for each time–frequency pixel using a paired t test with p < .05. For the correction of FWE arising from these mass univariate multiple comparisons, false discovery rate correction (Benjamini & Yekutieli, 2001) was applied. FWE rate-controlled significance level was set at p < .05, which served as multiple-comparison-corrected final results. The statistical comparison of ERSP was performed using EEGLAB functions. In addition to the statistical tests, we evaluated another difference measure Cohen's d as the effect size of differences, which is defined as (mean_difference)/(pooled_standard_deviation). As ICA is a data-driven method, in the IC clustering, the numbers of participants contributing to each cluster were not the same across clusters. Here, between 71.4% and 100% of the participants were included in each cluster (mean = 84.5%, SD = 13.1%). Table 1 details the numbers of included participants and ICs.

To test correlations of the power modulation with the time profiles of the sinusoidal visual stimulus (OF or target size) and/or accompanying postural behavior, the following sinusoidal regression model was adopted:
$yt=A+B1sin2πft+φ1+B2sin4πft+φ2+et−5≤t<5,$
(1)
where y(t) is event-related log power change from baseline for an IC process at a given frequency bin. The second term represents the stimulus effects of directional position and velocity (f = 0.2 Hz); the third term represents the effect of nondirectional (squared) velocity or speed of visual flow. e(t) is the residual error. The contributions of the second and third terms were evaluated using the following criteria:
$C1=varB1sin2πft+φ1varyt,$
(2)
$C2=varB2sin4πft+φ2varyt,$
(3)
where var[·] represents the variance operator for −5 ≤ t < 5. The contribution ratios C1 and C2 were averaged over the clustered ICs. Assuming that the EEG power is modulated by the visual stimulation and/or motor behavior, the EEG power is then increased or decreased at specific phases. Therefore, the phase distribution in sinusoidal changes of EEG power for the clustered ICs is expected to be nonuniform. The uniformity of the phase distribution was statistically examined by Rayleigh testing using a p < .05 significance threshold. Rayleigh tests were performed using the Circular Statistics Toolbox for MATLAB (Berens, 2009).

## RESULTS

### Behavioral Results

Figure 5 shows typical time profiles of the visual stimulus, COP, and EMG for the plantar flexor muscle. In the OF condition, the COP moved in phase with the OF. As expected, visual-flow-related fluctuations in COP location and muscle activity were much smaller in the OF-T condition. In the VT-F condition, the fluctuations in COP and muscle activities resembled those in the OF condition.

Figure 5.

Sample visual, behavioral, and EMG data. The top trace shows (in the OF and OF-T conditions) the sinusoidally varying apparent visual distance (approach/withdrawal) into the imaged corridor produced by OF of the corridor walls (beginning at Time 0). In the VT-F condition, this trace is proportional to the varying size of a (red) target rectangle; the participant voluntarily produces COP displacements to attempt to match the size of the visual feedback (black) rectangle to that of the target rectangle. The second through fourth traces show examples of correlated variations in behavioral COP. The bottom three traces show sample EMG traces for the gastrocnemius muscle in each condition.

Figure 5.

Sample visual, behavioral, and EMG data. The top trace shows (in the OF and OF-T conditions) the sinusoidally varying apparent visual distance (approach/withdrawal) into the imaged corridor produced by OF of the corridor walls (beginning at Time 0). In the VT-F condition, this trace is proportional to the varying size of a (red) target rectangle; the participant voluntarily produces COP displacements to attempt to match the size of the visual feedback (black) rectangle to that of the target rectangle. The second through fourth traces show examples of correlated variations in behavioral COP. The bottom three traces show sample EMG traces for the gastrocnemius muscle in each condition.

Figure 6A and B shows the mean amplitude of COP displacement and EMG amplitude change across all participants in the OF and OF-T conditions and in the OF and VT-F conditions, respectively. ANOVAs for COP displacement and EMG amplitude revealed significant differences among the task conditions (COP: F(2, 26) = 8.28, p = .002; EMG: F(2, 26) =14.21, p < .001). Table 2 shows the results of the post hoc tests and Cohen's d for differences in OF versus OF-T and OF versus VT-F. The comparison between OF and OF-T indicates the effect that experiencing OF-T reduced VIPS. In the VT-F condition, the voluntary postural sway was comparable with VIPS. The EMG amplitude in VT-F was higher than that in the OF condition, although the difference was not significant, caused by variability among participants.

Figure 6.

(Left) Average COP displacements and (right) mean EMG amplitudes for each participant. The top diagrams show the results for the OF and OF-T conditions and indicate statistical differences between them. The bottom diagrams similarly compare the OF and VT-F conditions.

Figure 6.

(Left) Average COP displacements and (right) mean EMG amplitudes for each participant. The top diagrams show the results for the OF and OF-T conditions and indicate statistical differences between them. The bottom diagrams similarly compare the OF and VT-F conditions.

Table 2.
Results of Statistical Tests of COP Displacement and EMG Amplitude of the OF-T and VT-F Conditions in Relation to the OF Condition
OF-T vs. OFVT-F vs. OF
Diff. (SD)p ValueCohen's dDiff. (SD)p ValueCohen's d
COP (m) −0.017 (0.011) < .001 1.45 0.002 (0.021) .81 0.11
EMG (μV) −3.62 (3.85) < .001 0.94 4.02 (5.99) .063 0.67
OF-T vs. OFVT-F vs. OF
Diff. (SD)p ValueCohen's dDiff. (SD)p ValueCohen's d
COP (m) −0.017 (0.011) < .001 1.45 0.002 (0.021) .81 0.11
EMG (μV) −3.62 (3.85) < .001 0.94 4.02 (5.99) .063 0.67

The columns for difference give mean difference (standard deviation) from the OF condition.

### EEG Results

Figure 7 shows the averaged topographic scalp maps of each IC cluster (Figure 7A). The centroid locations and standard deviations of the IC cluster equivalent dipoles are shown in Table 1. We focused on the following four IC clusters (leftmost in Figure 7A) related to the visuomotor control in the dorsal stream: the occipital cluster, left and right parietal clusters, and medial frontal cluster. The locations of the equivalent current dipoles for the ICs in each cluster of interest are shown in Figure 7B.

Figure 7.

Results of clustering brain-based ICs. (Top) IC-mean scalp map projections for each cluster. (Bottom) Locations of equivalent current dipoles for ICs in four clusters of interest.

Figure 7.

Results of clustering brain-based ICs. (Top) IC-mean scalp map projections for each cluster. (Bottom) Locations of equivalent current dipoles for ICs in four clusters of interest.

Figure 8 shows comparisons of cluster-mean power spectral densities. Tables 3 and 4 summarize the results of two-way repeated-measures ANOVAs of the power spectrum densities among three task conditions and three frequency bands. In addition, Figure 9 shows the condition differences of EEG power change from the baseline across the clustered ICs for each frequency band and IC cluster. There were significant main effects of the experiment conditions in occipital, left, and right parietal clusters and significant interactions in all clusters. Although the visual stimuli in the OF and OF-T conditions were the same, the post hoc tests indicated that alpha and beta band power in the occipital cluster was significantly lower in the OF condition than in the OF-T condition (top row of the Figure 9A and Table 4). On the other hand, the power spectra in the other three (parietal and frontal) clusters did not differ significantly between these two conditions in which the degree of postural sway and EMG amplitude variation differed significantly. Note the similarity of the left and right parietal cluster power spectra. Comparison between the VT-F and OF (voluntary and involuntary sway) conditions revealed significant differences in spectra in all the four clusters (right column of Figure 8). Although the differences of COP displacement and EMG activations were not significant in these two conditions, the medial frontal cluster showed condition differences in power spectra in theta and beta bands (bottom row of the Figure 9B and Table 4).

Figure 8.

Cluster-mean power spectral densities over IC processes in each cluster. The left spectral plots compare the OF and OF-T conditions. The right plots compare the OF and VT-F conditions. The areas with colored shading indicate frequency bands with statistically significant condition differences (p < .05). The color of the shading indicates the condition that is significantly higher than the other condition (pink: OF-T > OF; blue: VT-F > OF; gray: OF > OF-T/VT-F).

Figure 8.

Cluster-mean power spectral densities over IC processes in each cluster. The left spectral plots compare the OF and OF-T conditions. The right plots compare the OF and VT-F conditions. The areas with colored shading indicate frequency bands with statistically significant condition differences (p < .05). The color of the shading indicates the condition that is significantly higher than the other condition (pink: OF-T > OF; blue: VT-F > OF; gray: OF > OF-T/VT-F).

Table 3.
Comparison of Cluster-Mean Power Spectrum Densities with Two-way Repeated-Measures ANOVA (3 Conditions × 3 Frequency Bands)
ConditionsFrequency BandInteraction
F Valuep Valueηp2F Valuep Valueηp2F Valuep Valueηp2
Occipital F(2, 54) = 28.98 < .001 .51 F(2, 54) = 3.00 .063 .10 F(4, 108) = 25.12 < .001 .48
L. parietal F(2, 44) = 9.90 < .001 .31 F(2, 44) = 7.13 < .001 .24 F(4, 88) = 16.75 < .001 .43
R. parietal F(2, 42) = 20.54 < .001 .49 F(2, 42) = 24.03 < .001 .53 F(4, 84) = 30.18 < .001 .59
Med. frontal F(2, 36) = 0.30 .75 .02 F(2, 36) = 7.90 < .001 .31 F(4, 72) = 18.53 < .001 .51
ConditionsFrequency BandInteraction
F Valuep Valueηp2F Valuep Valueηp2F Valuep Valueηp2
Occipital F(2, 54) = 28.98 < .001 .51 F(2, 54) = 3.00 .063 .10 F(4, 108) = 25.12 < .001 .48
L. parietal F(2, 44) = 9.90 < .001 .31 F(2, 44) = 7.13 < .001 .24 F(4, 88) = 16.75 < .001 .43
R. parietal F(2, 42) = 20.54 < .001 .49 F(2, 42) = 24.03 < .001 .53 F(4, 84) = 30.18 < .001 .59
Med. frontal F(2, 36) = 0.30 .75 .02 F(2, 36) = 7.90 < .001 .31 F(4, 72) = 18.53 < .001 .51

The values of partial eta squared (ηp2) are the effect size to the ANOVA results.

Table 4.
Results of Multiple Comparisons of Theta, Alpha, and Beta Power between the OF-T and VT-F Conditions in Relation to the OF Condition
OF-T vs. OFVT-F vs. OF
Diff. (SD)p ValueCohen's dDiff. (SD)p ValueCohen's d
Occipital θ 0.097 (0.322) .264 0.301 −0.457 (1.258) .152 0.363
α 0.607 (0.672) < .001 0.903 −1.750 (1.640) < .001 1.067
β 0.170 (0.222) .001 0.765 −0.599 (0.863) .003 0.693
L. parietal θ 0.085 (0.443) .636 0.191 0.114 (0.618) .658 0.184
α −0.072 (1.184) .954 0.061 −1.669 (1.895) .001 0.881
β −0.082 (0.570) .774 0.143 −0.851 (0.984) .001 0.866
R. parietal θ 0.049 (0.262) .658 0.188 0.048 (0.684) .941 0.071
α −0.009 (0.650) .998 0.013 −1.741 (1.574) < .001 1.106
β −0.131 (0.321) .159 0.409 −1.177 (0.839) < .001 1.402
Med. frontal θ 0.081 (0.251) .351 0.326 0.403 (0.385) .001 1.047
α 0.003 (0.233) .998 0.012 −0.063 (0.468) .828 0.135
β −0.003 (0.276) .998 0.011 −0.402 (0.474) .004 0.848
OF-T vs. OFVT-F vs. OF
Diff. (SD)p ValueCohen's dDiff. (SD)p ValueCohen's d
Occipital θ 0.097 (0.322) .264 0.301 −0.457 (1.258) .152 0.363
α 0.607 (0.672) < .001 0.903 −1.750 (1.640) < .001 1.067
β 0.170 (0.222) .001 0.765 −0.599 (0.863) .003 0.693
L. parietal θ 0.085 (0.443) .636 0.191 0.114 (0.618) .658 0.184
α −0.072 (1.184) .954 0.061 −1.669 (1.895) .001 0.881
β −0.082 (0.570) .774 0.143 −0.851 (0.984) .001 0.866
R. parietal θ 0.049 (0.262) .658 0.188 0.048 (0.684) .941 0.071
α −0.009 (0.650) .998 0.013 −1.741 (1.574) < .001 1.106
β −0.131 (0.321) .159 0.409 −1.177 (0.839) < .001 1.402
Med. frontal θ 0.081 (0.251) .351 0.326 0.403 (0.385) .001 1.047
α 0.003 (0.233) .998 0.012 −0.063 (0.468) .828 0.135
β −0.003 (0.276) .998 0.011 −0.402 (0.474) .004 0.848

The second and fifth columns provide mean (standard deviation) differences from the OF condition for the OF-T and VT-F conditions.

Figure 9.

Condition differences of theta, alpha, and beta band power from baseline in comparisons between the OF-T and OF conditions and between the VT-F and OF conditions. The numbers of plotted lines correspond to the numbers of ICs in the clusters.

Figure 9.

Condition differences of theta, alpha, and beta band power from baseline in comparisons between the OF-T and OF conditions and between the VT-F and OF conditions. The numbers of plotted lines correspond to the numbers of ICs in the clusters.

Figure 10A shows ERSP images during two cycles of the OF/target movements in the OF, OF-T, and VT-F conditions. Periodic power modulations in the OF and OF-T conditions are clearly seen in the occipital and left parietal cluster ERSPs. The statistical differences and Cohen's d between OF-T and OF are shown in Figure 10B. As well as the power spectral density, alpha and beta band power of the ICs in the occipital cluster in the OF-T condition (as shown in the ERSP images) was significantly higher than that in the OF condition. The Cohen's d image in the occipital cluster also indicated an obvious difference around the alpha band. Figure 11 shows the profiles of the power modulation in the 10-Hz bin for the OF and OF-T conditions. The waveforms in OF-T and OF-T were similar, and their difference (green dash-dotted line) shows an approximately constant profile. It is likely that the alpha power increase with light touch was independent from the OF velocity. The ERSP images for the other three clusters do not indicate any significant differences between conditions. Figure 10C compares ERSPs in the OF and VT-F conditions. All clusters show significant differences between these conditions. Similar to the power spectral density (Figure 8), in the VT-F condition, the estimated EEG sources in the medial frontal cluster produced higher theta band power and smaller beta band power than that during VIPS (OF).

Figure 10.

Cluster-mean ERSPs compared across the OF, OF-T, and VT-F conditions within the occipital (top), left parietal (top center), right parietal (bottom center), and medial frontal (bottom) IC clusters. The left panels show equivalent dipole locations of the clustered ICs. The sinusoidal waves on the top panels show OF velocity or change rate of target size. The vertical dashed lines demark successive half cycles in the sinusoidal variation. (A) ERSPs of the OF, OF-T, and VT-F conditions. (B and C) False-discovery-rate-corrected p values (left) and Cohen's d (right) for comparisons for OF-T versus OF and VT-F versus OF, respectively.

Figure 10.

Cluster-mean ERSPs compared across the OF, OF-T, and VT-F conditions within the occipital (top), left parietal (top center), right parietal (bottom center), and medial frontal (bottom) IC clusters. The left panels show equivalent dipole locations of the clustered ICs. The sinusoidal waves on the top panels show OF velocity or change rate of target size. The vertical dashed lines demark successive half cycles in the sinusoidal variation. (A) ERSPs of the OF, OF-T, and VT-F conditions. (B and C) False-discovery-rate-corrected p values (left) and Cohen's d (right) for comparisons for OF-T versus OF and VT-F versus OF, respectively.

Figure 11.

ERSP profiles at the 10-Hz component in the occipital cluster. The black-solid and red-dashed lines denote the profiles of the OF and OF-T conditions, respectively. The green-dash-dotted lines denote the difference of the two profiles, which indicates constant power difference (OF-T > OF).

Figure 11.

ERSP profiles at the 10-Hz component in the occipital cluster. The black-solid and red-dashed lines denote the profiles of the OF and OF-T conditions, respectively. The green-dash-dotted lines denote the difference of the two profiles, which indicates constant power difference (OF-T > OF).

We evaluated periodically altered power modulation in occipital and parietal clusters by sinusoidal regression (Figures 12 and 13). As seen in Figure 12, for the occipital cluster, the contribution ratios C1 and C2 in the alpha band were higher than in other frequency bins in all three conditions. At both 0.2 and 0.4 Hz, entrainment to the sinusoidal variations explained approximately 20–30% of the changes in the alpha and beta bands. In total, the two sinusoidal functions covered about 50% of the variation of power modulation in these frequency bands. The contribution of the 0.4-Hz component (nondirectional velocity) was higher than that of the 0.2-Hz component in the OF and OF-T conditions, whereas the 0.2-Hz component was higher in the VT-F condition. Furthermore, phase nonuniformity test revealed nonuniform distribution of phase differences φ1 and φ2 in the alpha and beta bands. For the left parietal cluster, as shown in Figure 13, the correlations were smaller than those in the occipital cluster. In the OF and OF-T conditions, the contribution ratios and their phase distributions appear to be modulated by the speed of OF, as for the occipital cluster. On the other hand, in the VT-F condition, the contribution ratios and nonuniformity of the phase distribution show smaller effects of the visual stimulation than for the occipital cluster.

Figure 12.

Results of sinusoidal regression analysis on ERSPs for the occipital IC cluster. Left column panels show the cluster-mean ERSPs in the three conditions. Center and right column panels show frequency-wise contribution ratio (means and SDs) at the 0.2- and 0.4-Hz modulation frequencies, respectively. Pink-shaded areas indicate frequencies at which the phase distribution is significantly different from uniform (p < .05). The high contribution of 0.4-Hz components and significant nonuniformity in and around the alpha and beta bands indicates that ICs in the occipital cluster were influenced by the magnitude of the OF velocity.

Figure 12.

Results of sinusoidal regression analysis on ERSPs for the occipital IC cluster. Left column panels show the cluster-mean ERSPs in the three conditions. Center and right column panels show frequency-wise contribution ratio (means and SDs) at the 0.2- and 0.4-Hz modulation frequencies, respectively. Pink-shaded areas indicate frequencies at which the phase distribution is significantly different from uniform (p < .05). The high contribution of 0.4-Hz components and significant nonuniformity in and around the alpha and beta bands indicates that ICs in the occipital cluster were influenced by the magnitude of the OF velocity.

Figure 13.

Results of sinusoidal regression analysis on ERSPs for the left parietal IC cluster. Other details as in Figure 11. These results indicate that the ICs in the left parietal cluster were also influenced by OF velocity.

Figure 13.

Results of sinusoidal regression analysis on ERSPs for the left parietal IC cluster. Other details as in Figure 11. These results indicate that the ICs in the left parietal cluster were also influenced by OF velocity.

## DISCUSSION

The results of the current study support the occipital sensory integration model—the hypothesis that VIPS is mediated by primary visual and sensorimotor areas rather than by cortical areas in the dorsal stream. Alpha band power of sources in the occipital cluster was modulated by the sinusoidal variations in visual stimulation both during light finger touch (OF-T) and in its absence (OF; Figure 12). Not only the visual motion but also the presence of finger touch influenced occipital alpha band activity; alpha band suppression was attenuated during the OF-T condition relative to the OF condition (Figures 811). These results suggest that the visual cortex is involved in multisensory integration in VIPS.

Another piece of evidence that refutes the parietal sensory integration model are the ERSP results for the medial frontal cluster. The pattern of increased theta and suppressed beta band power during intentional visuo-postural tracking (VT-F) was significantly altered in VIPS (Figures 810), despite the similar sway behavior in the two conditions (Figure 6). Furthermore, the medial frontal ERSP did not differ significantly between the OF and OF-T conditions (Figure 10), although postural changes and concomitant muscle activities were significantly attenuated when participants touched the finger touch platform (Figure 6).

These results suggest that the motor process in VIPS is different from motor processes linked to activity in the dorsal stream. It is probable that sensory integration occurring in the visual perception enables perception of self-motion with less need for distributed cortical network activity than during responses to other kinds of perceptual challenges, such as visually guided movements, in which cortical areas in the dorsal stream must become involved. In addition, the net portion of local field activity in sensorimotor areas recorded in the EEG scalp signals is reduced during the OF condition compared to the VT-F condition (Figure 10), which also suggests that VIPS is driven by a more time- and energy-efficient motor system process than that supporting voluntary motor activity processes. The more efficient visuomotor process is likely beneficial for stabilizing posture in light of perceived visual motion. We conclude that involuntary postural adjustment from visual motion is driven by early-stage visual and subcortical processes rather than by processes in the dorsal stream that support voluntary movements.

In the occipital and parietal clusters, correlation analysis of the alpha power modulation showed high contribution of 0.2- and 0.4-Hz components of sinusoidal variation and nonuniform distribution of phase difference in the OF and OF-T conditions (Figures 12 and 13). The common features between the OF and OF-T conditions imply that the sinusoidal variations of alpha power were induced by visual motion. The results of correlation analysis correspond to previous EEG studies for OF (Vilhelmsen, van der Weel, & van der Meer, 2015; Delon-Martin et al., 2006; Probst et al., 1993). The ERSP images of right parietal clusters in the OF and OF-T conditions show an alpha power increase from baseline, whereas those of left parietal cluster do not. The asymmetric ERSPs may be related to laterality (Mutha, Haaland, & Sainburg, 2012), such as handedness and/or leg dominance, but there seems no clear evidence that supports this speculation.

Increases in posterior scalp alpha band power are typically interpreted as indexing deactivated or idling states of the visual processing system (Pfurtscheller et al., 1996). In our study, the smaller alpha band power in the visuo-postural tracking task may be associated with more visual processing demands to compare the sizes of the target and COP feedback rectangles simultaneously. In addition, more demand on the visual attention to the tracking target and COP feedback is another factor of the alpha band suppression. The higher alpha band power in the finger touch condition may imply less active cortical processing of visual motion. On the other hand, previous investigations of the effects of light finger touch on EEG indicated contributions of parietal and sensorimotor cortices (Ishigaki et al., 2016; Bolton, McIlroy, & Staines, 2011). During standing while touching a stable surface, somatosensory potentials evoked by stimulation of the median nerve differed from those in a no-touch condition (Bolton et al., 2011). Ishigaki et al. (2016) reported that alpha band power in the sensorimotor and parietal cortices was significantly smaller during light finger touch. The participants in the previous studies were instructed to close their eyes during the light touch task, whereas in this study, participants were instructed to watch the linear OF in a large field-of-view to cause intense dynamic visual experience. Such influx of visual information could have caused different cortical activities from the previous studies.

One may wonder if the differences in occipital alpha band power between the OF and OF-T conditions might have resulted from retinal visual motion accompanying postural sway. However, such an explanation would be inconsistent with the behavioral and EEG power spectral data. Because the direction of the COP movement in VIPS was approximately the same as the OF (Figure 5), the velocity of virtual corridor texture relative to the eye position during VIPS should be slower than during standing with light finger touch. If the velocity of visual motion on the retina directly affected alpha power in occipital cluster sources, the alpha power should have been more suppressed in the OF-T than the OF condition, because retinal visual motion in the OF-T (touch) condition was faster than in the OF (no-touch) condition. However, our analysis gave opposite results: Alpha band power increased in the finger touch (OF-T) condition. Therefore, the occipital cluster alpha increase in the OF-T condition cannot be simply accounted for by a difference in visual motion related to head movement.

In addition, motion-related artifacts from neck muscle activities are not plausible explanations for the occipital alpha power increase in the OF-T condition. These artifacts would contain other frequency components; EEG motion artifacts from postural sway have power in a lower frequency band (1–7 Hz; Kline, Huang, Snyder, & Ferris, 2015), whereas the spike-like EMG waveforms produce strong EMG power at 30 Hz and above, hence they do not overlap. We conclude that the occipital alpha band power reduction during OF alone and its enhancement during light finger touch reflect neural activities associated with the visual cortex. In comparisons between the OF and VT-F conditions, motion-related EEG artifacts were not a plausible interpretation of the EEG power spectral differences, as the magnitude of the postural sway was comparable between these conditions. In addition to the behavioral data, eye movements might not be the cause of these differences, as the gaze was fixed at the eye fixation point during both tasks. In the VT-F condition, because the rectangle enlargements and contraction occurred around the eye fixation point, the participants could perceive the size of the rectangles without eye movements. Therefore, it is also likely that the EEG power spectral differences between the OF and VT-F conditions were caused by differences in cortical dynamics.

Alpha power differences between lateral motion-sensitive and central primary visual areas might be expected to appear in the OF and VT-F conditions. To test this, we divided the occipital cluster (cf. Figure 7B) into central and (right) lateral occipital subclusters. However, no significant alpha power differences were found between the two subclusters. Possibly, an EEG study involving more participants might reveal source-resolved alpha activity differences between occipital areas. Although from our analysis, it is difficult to specify the portion of the visual cortex that engages the VIPS, the primary visual cortex is a plausible area as its information processing is affected by both visual motion and sensorimotor signals (Roth et al., 2016; Saleem et al., 2013).

Saleem et al. (2013) reported that, in mice, the level of neuronal activation in V1 during exposure to OF could be accounted for by the weighted sum of the speed of visual motion plus the actual speed of locomotion. A report of Roth et al. (2016) revealed the contribution of the thalamus to sensory integration in the visual cortex. Interestingly, the lateral posterior nucleus of the thalamus signals the discrepancy between retinal visual motion and the degree of motion expected from sensorimotor signals. Here, in both the OF and OF-T conditions, there was a discrepancy between behavioral and visual inputs. The visual motion provided an illusion of postural sway while the somatosensory feedback signaled the actual postural movement. The upper limb and hand feedback received through finger touch may improve certainty in postural perception and enable more accurate detection of the discrepancy. It is not clear which signal modulates the primary visual cortex, direct somatosensory signals from arm proprioception or indirect cortical feedback from parietal and sensorimotor cortices (Ishigaki et al., 2016; Bolton et al., 2011). We speculate that, in the finger touch condition, thalamic and visual cortical networks may have modulated occipital alpha power related to visual motion irrespective of self-motion.

Cavanagh and Frank (2014) suggested that medial frontal theta band power in their experiments was positively correlated with demand for cognitive control. In postural control, Hülsdünker, Mierau, Neeb, Kleinöder, and Strüder (2015) and Hülsdünker, Mierau, and Strüder (2016) showed that medial frontal theta band power was larger during less stable standing conditions (unipedal standing and standing on an unstable platform) than during bipedal standing on a stable platform. Malcolm, Foxe, Butler, Molholm, and De Sanctis (2018) investigated the effects of cognitive load under OF during walking using MoBI recording methods. They found more cautious walking behavior and higher theta band power in ACC during a dual-task condition (OF walking during performance of a go/no-go task) than during OF walking alone.

Beta band desynchronization in the sensorimotor cortex reflects motor planning and preparation of voluntary movements (Zaepffel et al., 2013; Tzagarakis et al., 2010; Alegre et al., 2003; Jasper & Penfield, 1949). Therefore, higher medial frontal theta and lower beta power in the VT-F condition may be interpreted as indexing involvement of cognitive and voluntary motor processes during target tracking by postural adjustment. The absence of this increased theta and decreased beta band power effects during VIPS suggests the influence of a lower-order visuomotor process not involving prefrontal cognitive effort- and motor-planning-related activity. There are reports that interactions of the medial frontal and parietal regions with the pyramidal tracts contribute to postural control (Beloozerova et al., 2005), suggesting important contributions of the dorsal pathway to postural control. The postural reaction to OF would then be a specific process within the postural control system.

Our methodology did not allow us to suggest involvement of an alternative pathway in responding to visual motion. Nashner and Berthoz (1978), who demonstrated the short latency of VIPS, suggested that extrapyramidal tracts such as the reticulospinal and tectospinal tracts might drive the short-latency response. Reticulospinal tracts are known to contribute to the regulation of muscle tone and posture (Takakusaki, 2017). The vestibulocerebellum also contributes to the postural balance through the vestibular nuclei and vestibulospinal tract. There are bimodal neurons responding to visual and vestibular stimuli in the vestibular nuclei and the fastigial nuclei of the cerebellum in goldfish (Allum, Graf, Dichgans, & Schmidt, 1976), rats (Precht, 1981), and monkeys (Büttner, Fuchs, Markert-Schwab, & Buckmaster, 1991; Waespe & Henn, 1977). However, the visuo-vestibular bimodal neurons contribute to only eye movement (optokinetic reflex); neurons not related to eye movements are not directly modulated by the large field visual motion in mice (Beraneck & Cullen, 2007) and macaques (Cullen, 2012; Bryan & Angelaki, 2009). Our behavioral and EEG results are consistent with a thalamocortical connection to the visual cortex (Roth et al., 2016). Cerebellar vermis as a part of spinocerebellum receives visual input from the visual cortex via pontine nuclei and influences leg muscles through reticulospinal and vestibulospinal tracts. These subcortical networks can be involved in the rapid postural reaction from large-field visual motion. However, human neural activities in these deep areas of the midbrain, brainstem, and cerebellum cannot be measured with EEG. To clarify the roles of these deep structures could require use of invasive recording in animals.

In conclusion, we observed that visuomotor cortical processing is involved in VIPS and that parietal cortical areas in the dorsal pathway are not primarily involved. Our result suggests that the early visual cortex plays a central role in VIPS rather than the dorsal visual processing stream. The time constant of dorsal visual processes, which were dominantly involved in the voluntary visuomotor tracking task, was found to be too slow to account for VIPS. The brain network underlying VIPS should include a subcortical postural control network to the visual cortex to enable quick and efficient postural control in responding to changing visual input.

## Author Contributions

Takahiro Kagawa: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing. Makoto Miyakoshi: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—original draft; Writing–review & editing. Scott Makeig: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing.

## Funding Information

The Swartz Foundation, a gift to The Swartz Center, UCSD. Japan Society for the Promotion of Science (http://dx.doi.org/10.13039/501100001691), Advancing Strategic International Networks, KAKENHI grant number: 18K12173.

## Acknowledgments

This study was supported by the JSPS Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers, JSPS KAKENHI Grant Number 18K12173, and a gift to The Swartz Center, UCSD from The Swartz Foundation (Old Field, NY). We thank Dr. David Medine for technical support related to measurement systems; Nicole Wells for assistance during experiments; and Dr. John Iversen, Dr. Johanna Wagner, Dr. Hiroyuki Kambara, Dr. Natsue Yoshimura, Dr. Hirokazu Tanaka, Dr. Yoji Uno, and Dr. Yasuharu Koike for their valuable suggestions and considerable support for performing the experiment at the Swartz Center of Computational Neuroscience, UCSD.

Reprint requests should be sent to Takahiro Kagawa, Faculty of Engineering, Department of Mechanical Engineering, Aichi Institute of Technology, 1247 Yachigusa, Yakusa-cho, Toyota 470-0392, Japan, or via e-mail: t_kagawa@aitech.ac.jp.

## REFERENCES

Akalin Acar
,
Z.
, &
Makeig
,
S.
(
2010
).
.
Journal of Neuroscience Methods
,
190
,
258
270
.
Alegre
,
M.
,
Gurtubay
,
I. G.
,
Labarga
,
A.
,
Iriarte
,
J.
,
Malanda
,
A.
, &
Artieda
,
J.
(
2003
).
Alpha and beta oscillatory changes during stimulus-induced movement paradigms: Effect of stimulus predictability
.
NeuroReport
,
14
,
381
385
.
Allum
,
J. H.
,
Graf
,
W.
,
Dichgans
,
J.
, &
Schmidt
,
C. L.
(
1976
).
Visual-vestibular interactions in the vestibular nuclei of the goldfish
.
Experimental Brain Research
,
26
,
463
485
.
Baumann
,
O.
,
Borra
,
R. J.
,
Bower
,
J. M.
,
Cullen
,
K. E.
,
Habas
,
C.
,
Ivry
,
R. B.
, et al
(
2015
).
Consensus paper: The role of the cerebellum in perceptual processes
.
Cerebellum
,
14
,
197
220
.
Beloozerova
,
I. N.
,
Sirota
,
M. G.
,
Orlovsky
,
G. N.
, &
Deliagina
,
T. G.
(
2005
).
Activity of pyramidal tract neurons in the cat during postural corrections
.
Journal of Neurophysiology
,
93
,
1831
1844
.
Benjamini
,
Y.
, &
Yekutieli
,
D.
(
2001
).
The control of the false discovery rate in multiple testing under dependency
.
Annals of Statistics
,
29
,
1165
1188
.
Beraneck
,
M.
, &
Cullen
,
K. E.
(
2007
).
Activity of vestibular nuclei neurons during vestibular and optokinetic stimulation in the alert mouse
.
Journal of Neurophysiology
,
98
,
1549
1565
.
Berens
,
P.
(
2009
).
CircStat: A Matlab toolbox for circular statistics
.
Journal of Statistical Software
,
31
,
1
21
.
Bolton
,
D. A. E.
,
McIlroy
,
W. E.
, &
Staines
,
W. R.
(
2011
).
The impact of light fingertip touch on haptic cortical processing during a standing balance task
.
Experimental Brain Research
,
212
,
279
291
.
Bryan
,
A. S.
, &
Angelaki
,
D. E.
(
2009
).
Optokinetic and vestibular responsiveness in the macaque rostral vestibular and fastigial nuclei
.
Journal of Neurophysiology
,
101
,
714
720
.
Büttner
,
U.
,
Fuchs
,
A. F.
,
Markert-Schwab
,
G.
, &
Buckmaster
,
P.
(
1991
).
Fastigial nucleus activity in alert monkey during slow eye and head movements
.
Journal of Neurophysiology
,
65
,
1360
1371
.
Cardin
,
V.
, &
Smith
,
A. T.
(
2010
).
Sensitivity of human visual and vestibular cortical regions to egomotion-compatible visual stimulation
.
Cerebral Cortex
,
20
,
1964
1973
.
Cardin
,
V.
, &
Smith
,
A. T.
(
2011
).
Sensitivity of human visual cortical area V6 to stereoscopic depth gradients associated with self-motion
.
Journal of Neurophysiology
,
106
,
1240
1249
.
Cavanagh
,
J. F.
, &
Frank
,
M. J.
(
2014
).
Frontal theta as a mechanism for cognitive control
.
Trends in Cognitive Sciences
,
18
,
414
421
.
Colnaghi
,
S.
,
Honeine
,
J.-L.
,
Sozzi
,
S.
, &
Schieppati
,
M.
(
2017
).
Body sway increases after functional inactivation of the cerebellar vermis by cTBS
.
Cerebellum
,
16
,
1
14
.
Cullen
,
K. E.
(
2012
).
The vestibular system: Multimodal integration and encoding of self-motion for motor control
.
Trends in Neurosciences
,
35
,
185
196
.
Delon-Martin
,
C.
,
Gobbelé
,
R.
,
Buchner
,
H.
,
Haug
,
B. A.
,
Antal
,
A.
,
Darvas
,
F.
, et al
(
2006
).
Temporal pattern of source activities evoked by different types of motion onset stimuli
.
Neuroimage
,
31
,
1567
1579
.
Delorme
,
A.
, &
Makeig
,
S.
(
2004
).
EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis
.
Journal of Neuroscience Methods
,
134
,
9
21
.
Delorme
,
A.
,
Mullen
,
T.
,
Kothe
,
C.
,
Akalin Acar
,
Z.
,
Bigdely-Shamlo
,
N.
,
Vankov
,
A.
, et al
(
2011
).
EEGLAB, SIFT, NFT, BCILAB, and ERICA: New tools for advanced EEG processing
.
Computational Intelligence and Neuroscience
,
2011
,
130714
.
Delorme
,
A.
,
Palmer
,
J.
,
Onton
,
J.
,
Oostenveld
,
R.
, &
Makeig
,
S.
(
2012
).
Independent EEG sources are dipolar
.
PLoS One
,
7
,
e30135
.
Gibson
,
J. J.
(
1979
).
The ecological approach to visual perception
.
Boston
:
Houghton Mifflin
.
Goldman
,
R. I.
,
Stern
,
J. M.
,
Engel
, Jr.,
J.
, &
Cohen
,
M. S.
(
2002
).
Simultaneous EEG and fMRI of the alpha rhythm
.
NeuroReport
,
13
,
2487
2492
.
Gramann
,
K.
,
Gwin
,
J. T.
,
Bigdely-Shamlo
,
N.
,
Ferris
,
D. P.
, &
Makeig
,
S.
(
2010
).
Visual evoked responses during standing and walking
.
Frontiers in Human Neuroscience
,
4
,
202
.
Gramann
,
K.
,
Gwin
,
J. T.
,
Ferris
,
D. P.
,
Oie
,
K.
,
Jung
,
T.-P.
,
Lin
,
C.-T.
, et al
(
2011
).
Cognition in action: Imaging brain/body dynamics in mobile humans
.
Reviews in the Neurosciences
,
22
,
593
608
.
Graziano
,
M. S.
, &
Gross
,
C. G.
(
1993
).
A bimodal map of space: Somatosensory receptive fields in the macaque putamen with corresponding visual receptive fields
.
Experimental Brain Research
,
97
,
96
109
.
Gwin
,
J. T.
, &
Ferris
,
D. P.
(
2012
).
Beta- and gamma-range human lower limb corticomuscular coherence
.
Frontiers in Human Neuroscience
,
6
,
258
.
Hatzitaki
,
V.
, &
,
S.
(
2007
).
Visuo-postural adaptation during the acquisition of a visually guided weight-shifting task: Age-related differences in global and local dynamics
.
Experimental Brain Research
,
182
,
525
535
.
Holmes
,
N. P.
, &
Spence
,
C.
(
2004
).
The body schema and multisensory representation(s) of peripersonal space
.
Cognitive Processing
,
5
,
94
105
.
Hülsdünker
,
T.
,
Mierau
,
A.
,
Neeb
,
C.
,
Kleinöder
,
H.
, &
Strüder
,
H. K.
(
2015
).
Cortical processes associated with continuous balance control as revealed by EEG spectral power
.
Neuroscience Letters
,
592
,
1
5
.
Hülsdünker
,
T.
,
Mierau
,
A.
, &
Strüder
,
H. K.
(
2016
).
Higher balance task demands are associated with an increase in individual alpha peak frequency
.
Frontiers in Human Neuroscience
,
9
,
695
.
Ishigaki
,
T.
,
Ueta
,
K.
,
Imai
,
R.
, &
Morioka
,
S.
(
2016
).
EEG frequency analysis of cortical brain activities induced by effect of light touch
.
Experimental Brain Research
,
234
,
1429
1440
.
Jasper
,
H.
, &
Penfield
,
W.
(
1949
).
Electroeorticograms in man: Effect of voluntary movement upon the electrical activity of the precentral gyrus
.
Archiv für Psychiatrie und Zeitschrift Neurologie
,
183
,
163
174
.
Jeka
,
J.
, &
Lackner
,
J. R.
(
1994
).
Fingertip contact influences human postural control
.
Experimental Brain Research
,
100
,
495
502
.
Jeka
,
J.
,
Oie
,
K. S.
, &
Kiemel
,
T.
(
2000
).
Multisensory information for human postural control: Integrating touch and vision
.
Experimental Brain Research
,
134
,
107
125
.
Jung
,
T.-P.
,
Makeig
,
S.
,
Humphries
,
C.
,
Lee
,
T.-W.
,
McKeown
,
M. J.
,
Iragui
,
V.
, et al
(
2000
).
Removing electroencephalographic artifacts by blind source separation
.
Psychophysiology
,
37
,
163
178
.
Kline
,
J. E.
,
Huang
,
H. J.
,
Snyder
,
K. L.
, &
Ferris
,
D. P.
(
2015
).
Isolating gait-related movement artifacts in electroencephalography during human walking
.
Journal of Neural Engineering
,
12
,
046022
.
Kruse
,
W.
,
Dannenberg
,
S.
,
Kleiser
,
R.
, &
Hoffmann
,
K.-P.
(
2002
).
Temporal relation of population activity in visual areas MT/MST and in primary motor cortex during visually guided tracking movements
.
Cerebral Cortex
,
12
,
466
476
.
,
C. M.
,
Fulbright
,
R. K.
,
Rajeevan
,
N.
,
Constable
,
R. T.
, &
,
X.
(
2008
).
More accurate Talairach coordinates for neuroimaging using non-linear registration
.
Neuroimage
,
42
,
717
725
.
Lee
,
D. N.
, &
Aronson
,
E.
(
1974
).
Visual proprioceptive control of standing in human infants
.
Perception & Psychophysics
,
15
,
529
532
.
Lestienne
,
F.
,
Soechting
,
J.
, &
Berthoz
,
A.
(
1977
).
Postural readjustments induced by linear motion of visual scenes
.
Experimental Brain Research
,
28
,
363
384
.
Makeig
,
S.
(
1993
).
Auditory event-related dynamics of the EEG spectrum and effects of exposure to tones
.
Electroencephalography and Clinical Neurophysiology
,
86
,
283
293
.
Makeig
,
S.
,
Bell
,
A. J.
,
Jung
,
T.-P.
, &
Sejnowski
,
T. J.
(
1996
).
Independent component analysis of electroencephalographic data
. In
D.
Touretzky
,
M.
Mozer
, &
M.
Hasselmo
(Eds.),
Advances in neural information processing systems
(
Vol. 8
, pp.
145
151
).
Cambridge, MA
:
MIT Press
.
Makeig
,
S.
,
Gramann
,
K.
,
Jung
,
T.-P.
,
Sejnowski
,
T. J.
, &
Poizner
,
H.
(
2009
).
.
International Journal of Psychophysiology
,
73
,
95
100
.
Makeig
,
S.
,
Westerfield
,
M.
,
Jung
,
T.-P.
,
Enghoff
,
S.
,
Townsend
,
J.
,
Courchesne
,
E.
, et al
(
2002
).
Dynamic brain sources of visual evoked responses
.
Science
,
295
,
690
694
.
Malcolm
,
B. R.
,
Foxe
,
J. J.
,
Butler
,
J. S.
,
Molholm
,
S.
, &
De Sanctis
,
P.
(
2018
).
Cognitive load reduces the effects of optic flow on gait and electrocortical dynamics during treadmill walking
.
Journal of Neurophysiology
,
120
,
2246
2259
.
Moosmann
,
M.
,
Ritter
,
P.
,
Krastel
,
I.
,
Brink
,
A.
,
Thees
,
S.
,
Blankenburg
,
F.
, et al
(
2003
).
Correlates of alpha rhythm in functional magnetic resonance imaging and near infrared spectroscopy
.
Neuroimage
,
20
,
145
158
.
Morrone
,
M. C.
,
Tosetti
,
M.
,
Montanaro
,
D.
,
Fiorentini
,
A.
,
Cioni
,
G.
, &
Burr
,
D. C.
(
2000
).
A cortical area that responds specifically to optic flow, revealed by fMRI
.
Nature Neuroscience
,
3
,
1322
1328
.
Mutha
,
P. K.
,
Haaland
,
K. Y.
, &
Sainburg
,
R. L.
(
2012
).
The effects of bran lateralization on motor control and adaptation
.
Journal of Motor Behavior
,
44
,
455
469
.
Nashner
,
L.
, &
Berthoz
,
A.
(
1978
).
Visual contribution to rapid motor responses during postural control
.
Brain Research
,
150
,
403
407
.
Nassi
,
J. J.
, &
Callaway
,
E. M.
(
2009
).
Parallel processing strategies of the primate visual system
.
Nature Reviews Neuroscience
,
10
,
360
372
.
Oostenveld
,
R.
, &
Oostendorp
,
T. F.
(
2002
).
Validating the boundary element method for forward and inverse EEG computations in the presence of a hole in the skull
.
Human Brain Mapping
,
17
,
179
192
.
Ouchi
,
Y.
,
,
H.
,
Yoshikawa
,
E.
,
Nobezawa
,
S.
, &
Futatsubashi
,
M.
(
1999
).
Brain activation during maintenance of standing postures in humans
.
Brain
,
122
,
329
338
.
Palmer
,
J. A.
,
Makeig
,
S.
,
,
K.
, &
Rao
,
B. D.
(
2008
).
Newton method for the ICA mixture model
. In
Proceedings of the 33rd IEEE International Conference on Acoustics and Signal Processing
(pp.
1805
1808
).
Las Vegas, NV
:
IEEE
.
Pfurtscheller
,
G.
,
Stancák
, Jr.,
A.
, &
Neuper
,
C.
(
1996
).
Event-related synchronization (ERS) in the alpha band—An electrophysiological correlate of cortical idling: A review
.
International Journal of Psychophysiology
,
24
,
39
46
.
Pitzalis
,
S.
,
Sereno
,
M. I.
,
Committeri
,
G.
,
Fattori
,
P.
,
Galati
,
G.
,
Patria
,
F.
, et al
(
2010
).
Human V6: The medial motion area
.
Cerebral Cortex
,
20
,
411
424
.
Precht
,
W.
(
1981
).
Visual–vestibular interaction in vestibular neurons: Functional pathway organization
.
Annals of New York Academy of Sciences
,
374
,
230
248
.
Probst
,
T.
,
Plendl
,
H.
,
Paulus
,
W.
,
Wist
,
E. R.
, &
Scherg
,
M.
(
1993
).
Identification of the visual motion area (area V5) in the human brain by dipole source analysis
.
Experimental Brain Research
,
93
,
345
351
.
Roth
,
M. M.
,
Dahmen
,
J. C.
,
Muir
,
D. R.
,
Imhof
,
F.
,
Martini
,
F. J.
, &
Hofer
,
S. B.
(
2016
).
Thalamic nuclei convey diverse contextual information to layer 1 of visual cortex
.
Nature Neuroscience
,
19
,
299
307
.
Saleem
,
A. B.
,
Ayaz
,
A.
,
Jeffery
,
K. J.
,
Harris
,
K. D.
, &
Carandini
,
M.
(
2013
).
Integration of visual motion and locomotion in mouse visual cortex
.
Nature Neuroscience
,
16
,
1864
1869
.
Schlack
,
A.
,
Hoffmann
,
K.-P.
, &
Bremmer
,
F.
(
2002
).
Interaction of linear vestibular and visual stimulation in the macaque ventral intraparietal area (VIP)
.
European Journal of Neuroscience
,
16
,
1877
1886
.
Sipp
,
A. R.
,
Gwin
,
J. T.
,
Makeig
,
S.
, &
Ferris
,
D. P.
(
2013
).
Loss of balance during balance beam walking elicits a multifocal theta band electrocortical response
.
Journal of Neurophysiology
,
110
,
2050
2060
.
Solis-Escalante
,
T.
,
van der Cruijsen
,
J.
,
de Kam
,
D.
,
van Kordelaar
,
J.
,
Weerdesteyn
,
V.
, &
Schouten
,
A. C.
(
2019
).
Cortical dynamics during preparation and execution of reactive balance responses with distinct postural demands
.
Neuroimage
,
188
,
557
571
.
Takakusaki
,
K.
(
2017
).
Functional neuroanatomy for posture and gait control
.
Journal of Movement Disorders
,
10
,
1
17
.
Tanné-Gariépy
,
J.
,
Rouiller
,
E. M.
, &
Boussaoud
,
D.
(
2002
).
Parietal inputs to dorsal versus ventral premotor areas in the macaque monkey: Evidence for largely segregated visuomotor pathways
.
Experimental Brain Research
,
145
,
91
103
.
Tzagarakis
,
C.
,
Ince
,
N. F.
,
Leuthold
,
A. C.
, &
Pellizzer
,
G.
(
2010
).
Beta-band activity during motor planning reflects response uncertainty
.
Journal of Neuroscience
,
30
,
11270
11277
.
van der Kooij
,
H.
,
Jacobs
,
R.
,
Koopman
,
B.
, &
Grootenboer
,
H.
(
1999
).
A multisensory integration model of human stance control
.
Biological Cybernetics
,
80
,
299
308
.
van der Kooij
,
H.
,
Jacobs
,
R.
,
Koopman
,
B.
, &
van der Helm
,
F.
(
2001
).
An adaptive model of sensory integration in a dynamic environment applied to human stance control
.
Biological Cybernetics
,
84
,
103
115
.
Vilhelmsen
,
K.
,
van der Weel
,
F. R. R.
, &
van der Weer
,
A. L. H.
(
2015
).
A high-density EEG study of differences between three high speeds of simulated forward motion from optic flow in adult participants
.
Frontiers in Systems Neuroscience
,
9
,
146
.
Waespe
,
W.
, &
Henn
,
V.
(
1977
).
Neuronal activity in the vestibular nuclei of the alert monkey during vestibular and optokinetic stimulation
.
Experimental Brain Research
,
27
,
523
538
.
Zaepffel
,
M.
,
Trachel
,
R.
,
Kilavik
,
B. E.
, &
Brochier
,
T.
(
2013
).
Modulations of EEG beta power during planning and execution of grasping movements
.
PLoS One
,
8
,
e60060
.