Disentangling the visual, motor and representational effects of vestibular input.

TL;DR: It is suggested that vestibular information contributes to computation of egocentric representations by affecting the internal representation of the body midline.

About: This article is published in Cortex.The article was published on 2018-07-01 and is currently open access. It has received 10 citations till now. The article focuses on the topics: Vestibular system.

Summary

1. Introduction

• Judging the position of external objects relative to the body is essential for interacting with the external environment.
• Both semicircular canals and otolith organs constantly provide afferent information regarding body orientation and body movement.
• Here the authors aimed to clarify whether and how vestibular inputs contribute to egocentric spatial representation in healthy volunteers.
• Perilymphatic cathodal currents depolarize the trigger site and lead to excitation, whereas anodal currents hyperpolarize it resulting in inhibition (Goldberg et al., 1984) .

2.1.2. Galvanic Vestibular Stimulation

• Bipolar GVS was applied to deliver a boxcar pulse of 1 mA using a commercial stimulator (Good Vibrations Engineering Ltd., Nobleton, Ontario, Canada) .
• Postural studies confirm that this level of GVS activates the vestibular organs without effects persisting beyond the period of stimulation (Fitzpatrick and Day, 2004) .
• Carbon rubber electrodes (area 10 cm 2 ) coated with electrode gel were placed binaurally over the mastoid processes and fixed in place with adhesive tape.
• Both left-anodal/right-cathodal (L-GVS) and right-anodal/left-cathodal (R-GVS) configurations were used (Fig. 1B ).
• The authors also applied a sham stimulation using electrodes placed on the left and right side of the neck, about 5 cm below the GVS electrodes (Ferrè et al., 2013a , Ferrè et al., 2013b) , with a left-anodal/rightcathodal configuration.

2.1.2. Procedure

• Verbal and written instructions were given to participants at the beginning of the session.
• The participants' posture was monitored throughout the experiment to ensure that the body midline was always aligned with the center of the monitor.
• The stimuli were shown 4 cm above the center of the monitor at the eye-line of the participants.
• In the other half the hands were crossed so that the left hand pressed the right-sided button and vice versa (Fig. 1C ).
• If the principal locus of GVS effects was to bias responding towards one hand, then performance on the localization task should be strongly affected by crossing the hands.

2.2. Results

• Trials with implausibly short reaction times (< 183 ms: corresponding to 1 % of trials) were excluded from analysis.
• The point of subjective equality (PSE) and just-noticeable difference (JND) were calculated from the functions for each participant in each condition.
• The effect of GVS polarity was analyzed using a 2 (GVS polarity: L-GVS, R-GVS) x 2 (Hand posture: uncrossed, crossed) ANOVA.

2.3. Discussion

• These results suggested that GVS changed the participants' egocentric spatial representation, with the perceived body midline shifting towards the anodal side.
• There was no effect of crossing the hands, ruling out alternative explanations based on GVS affecting the selection or efficiency of lateralized motor output processes.
• The authors data therefore support a body-related theory suggesting that GVS selectively modulates their egocentric spatial representations (Fig. 3A ).

3.1.2. Procedure

• Thus, the task required a visual allocentric representation, centred on the reference, and egocentric location was irrelevant.
• Peripheral presentation of the reference ensured that the task difficulty was equal between both experiments.
• As there was no main effect of hand position in Experiment 1, participants completed this task with hands uncrossed.
• The two reference locations (left and right) and three GVS polarities (L-GVS, R-GVS and sham stimulation) gave six conditions, repeated twice to give 12 blocks of 150 trials.
• The experimental setup and all other procedures were as Experiment 1.

3.2. Results

• The effect of GVS polarity was analyzed using a 2 (GVS polarity: L-GVS, R-GVS) x 2 (Reference: left, right) ANOVA.
• The authors performed a two-way ANOVA with Stimulation (L-GVS and R-GVS) as within factor and Experiment (Exp.1 and Exp.2) as between factor.
• These data indicate that GVS effect differ between experiments.
• The authors subtracted Sham-GVS from R-GVS and L-GVS.

3.3. Discussion

• GVS significantly biased allocentric visual localization in the opposite direction compared to the egocentric localization bias of Experiment 1 (Fig. 3D ).
• This shift in attention would subsequently increase foveal mislocalisation for visual stimuli presented on the left, and reduce mislocalisation for stimuli on the right.
• Hence, GVS biased allocentric representations towards the left for R-GVS and towards the right for L-GVS.
• Taken together, their results highlight the relationship between vestibular information and egocentric body representation.
• A final alternative explanation is that GVS could induce a change in gaze location which may have affected the egocentric judgement task.

Participants

• None of the participants had participated in the previous experiments.
• Exclusion criteria were as the previous experiments.

Procedure

• Participants completed the egocentric localization task with GVS while their gaze location was recorded.
• The task was completed with hands uncrossed.
• Only L-GVS and R-GVS polarities were used.
• The two conditions were repeated twice giving 4 blocks of 150 trials.
• Gaze location was recorded with an Eyelink 2000 eye-tracker (sampling frequency 1 kHz).

4.1.2 Results

• The PSE was shifted leftwards by 1.04 cm during L-GVS versus R-GVS, replicating the results of Experiment 1.
• Eye-tracking data were processed by manual inspection and exclusion of trials with eye-blinks.
• Figure 4B shows a fixation map, obtained from a representative participant.
• Areas with longer fixation are shown with warmer colours (i.e. red).
• Figure 4C shows the change in gaze location across time, averaged across participants.

4.2.2 Results

• Figure 4D shows average PSEs across participants while gaze was directed either leftwards or rightwards.
• The authors verified that the gaze was directed to the correct location by estimating the average gaze location during the period from the onset of the target to the response (mean ± SD: gaze-left = -3.9 ± 0.26 cm; gaze-right 4.0 ± 0.36 cm).

4.3. Discussion

• The results of Experiment 3 showed that GVS can bias not only egocentric body representations (Fig. 4A ), but also shift horizontal gaze location (Fig. 4B ).
• Thus, one might argue that this GVS-induced shift in gaze location could have indirectly driven the bias in egocentric body representation, without any direct effect of GVS.
• In Experiment 4, egocentric judgements were not significantly altered when participants were explicitly asked to direct their gaze leftwards or rightwards by an amount equivalent to the GVS-induced bias (Fig. 4C ).
• Taken together, these results therefore suggest that the indirect effect of GVS mediated by gaze shifts is minimal or absent.
• Thus, the strong biasing effect of GVS on egocentric body representations is independent of changes in gaze location, and appears to represent a direct vestibular input to spatial representation of the body midline.

5. General Discussion

• In many situations appropriate motor responses must be chosen rapidly based on the location of external objects relative to the body midline.
• This bias has been successfully remediated by artificial vestibular stimulation, suggesting that the vestibular system contributes to body representation (Cappa et al., 1987; Vallar et al., 1993; Karnath et al., 1994) .
• Thus, in Experiment 1, hand posture did not change the GVS effect on perceptual judgments (Fig. 2A and 2B), and there was no GVS effect on reaction times.
• Voluntary gaze shift has been reported to modulate spatial representations (Cui et al., 2010) .
• This seems to suggest that the representation of the midline is established by constant online integration of ongoing sensory input, rather than being a systematic stored knowledge about one's own body.