A Systematic Review and Meta-Analysis on The Use of Tactile Stimulation in Vection Research Lars Kooijman*, Houshyar Asadi, Shady Mohamed and Saeid Nahavandi, Institute for Intelligent Systems Research and Innovation, Deakin University, Australia. * Corresponding author. E-mail address: [email protected]. Authors’ affiliated emails: {L.Kooijman; Houshyar.Asadi; Shady.Mohamed; Saeid.Nahavandi}@deakin.edu.au ABSTRACT Vection is classically defined as the illusory perception of self-motion induced via visual stimuli. The utility of vection research lies in its potential to enhance simulation fidelity, as measured through presence, and reduce the probability that motion sickness symptoms occur. Recent studies have shown a multimodal interaction of various sensory systems in facilitating vection. Moreover, the utility of co-stimulating some of these sensory systems along with the presentation of visual stimuli have been reviewed. However, a review on the use of tactile stimulation in vection research appears to be missing from literature. The purpose of this review is to evaluate the current methodologies, and outcomes, of tactile stimulation in vection research. We searched for articles through EBSCOHost, Scopus and Web of Science. Studies were included only if they detailed an experiment on the effect of tactile stimulation on vection. Twenty-three studies were obtained and distilled in tabular form. Twenty studies contained sufficient information to be included in a meta-analysis. We identified that tactile stimulation has mostly been applied in the form of vibrational stimulation to the feet. Furthermore, tactile stimulation is most effective when it is presented in a temporally congruent manner to other sensory cues whereas tactile stimulation as a unisensory stimulus does not appear to be effective in eliciting vection. We discuss the need for more qualitative research to reduce methodological inhomogeneities and recommend future research in tactile-mediated vection to investigate stimulation to the torso and investigate the use of forces as a tactile stimulus. Keywords and Phrases: Vection, Tactile Stimulation, cue-integration, Systematic Review, Meta-analysis. 1 1 INTRODUCTION Sometimes when we are stationary, we experience the illusion that we are moving. The illusory experience of self-motion is often exemplified using historical anecdotal accounts. Wood (1895) detailed his perception of illusory self-motion, which he described as "goneness within”, based on his experiencing of the Haunted Swing illusion (Wood, 1895). In this illusion the contents of a room are rotated around you and thereby eliciting the feeling that you are moving yourself. The Haunted Swing illusion can nowadays be experienced in theme park rides such as the Madhouse or Villa Volta manufactured by Vekoma. Another common example of illusory self- motion perception is the train illusion, which can be traced back to William James’ book Principles of Psychology. James (1890) described that “...when another train comes alongside of ours in a station, and fills the entire window, and, after standing still awhile, begins to glide away, we judge that it is our train which is moving, and that the other train is still.” (James, 1890, p.91). These illusory experiences of self-movement are often referred to as vection in scientific research. The utility of vection research lies mainly in the understanding how vection is elicited and which sensory systems contribute to its perception, and subsequently using this information to 1) enhance the fidelity (e.g., immersiveness or presence) of virtual driving or flight simulators and 2) reduce the risk of eliciting motion sickness (MS). The fidelity of simulators is often assessed via measures of presence (e.g., Brackney & Priode, 2017), and previous research has shown a positive correlation between vection and spatial presence when participants viewed naturalistic moving stimuli (Riecke et al., 2005a). Currently, there is no consensus as to whether vection is a predeterminant of MS: findings either indicate that vection is (Nooij et al., 2017) or isn’t (Kuiper et al., 2019) a predeterminant of MS (see Keshavarz et al., 2015 for a review). The leading hypothesis is that MS occurs due to a mismatch between sensory signals and the neural storage of what these sensory signals are expected to be (Reason, 1978). For example, MS could occur when one perceives strong visual cues indicating self-motion while standing still. Since most motion simulators are incapable of generating large or fast physical motions of real vehicles to corroborate the virtual visual stimuli, there is a high probability that participants will experience MS in these simulators. Additionally, a novel theory by Palmisano et al. (2020) proposes that motion sickness in virtual reality (VR), which is also known as Cybersickness (CS), could be due to the lag in the rendering of visual feedback present in modern Head Mounted Displays (HMDs), which creates a discrepancy between participants’ virtual and physical head position. Although the use of HMDs has shown to be more immersive compared to screen-based methods (Shu et al., 2019), simulators utilizing HMDs are more likely to elicit CS due to this inherent rendering lag, which could negatively impact simulator fidelity. A review by Weech et al. (2019) concluded that there appears to be a negative relationship between presence and CS, and the authors highlighted the need to measure vection, CS and presence concurrently (e.g., see D’Amour et al. 2017; Fauville et al., 2021). Thus, it is worthwhile investigating if, and how, multisensory stimulation can elicit the appropriate (illusory) sensation of self-motion in a high-fidelity environment with the lowest probability of eliciting MS/CS. As such, it is necessary to disentangle the physiological and behavioural responses to MS/CS, presence and vection and compare them to subjective measures targeted at evaluating each phenomenon. Lastly, vection research can assist in our understanding of the functionality of, and correlation between, received information from multisensory systems that facilitate human perception. For example, it has recently been shown that vection can facilitate perspective switching which contributes to our spatial orientation (Riecke et al., 2015). Such fundamental information on human perception highlights the functional significance of vection research outside of the field of simulator development. Vection has classically been identified as a visually-induced illusion (Palmisano et al., 2015) and this inference can presumably be traced back to Mach (1875), who concluded that it was highly unlikely that sensations from connective tissue and bones (i.e., ‘Bindegewebe und Knochen’, pp. 65-66), skin (i.e., ‘Haut’, pp. 66-69), muscles (‘Muskeln’, pp. 69-76), blood (‘Blut’, pp. 76-79) and cerebellum (‘Kleinhirn’, pp. 90-94) contributed to our motion perception (Bewegungsempfindungen). However, the contribution and integration of various sensory signals involved in vection have been acknowledged (e.g., Britton & Arshad, 2019; Campos & Bülthoff, 2012; Greenlee et al., 2016; Väljamäe, 2009). Accordingly, Palmisano et al. (2015) analysed the definitions of vection that are used in literature and distilled their findings into four definitions. Throughout this paper, we shall adhere to the fourth definition, namely that vection is “…[a] conscious subjective experience of self-motion”, as it 2 encompasses both illusory as real self-motion perception identified through a multitude of sensory organs (Palmisano et al., 2015). One of the senses involved in self-motion perception is the human tactile sense. The contribution of tactile stimulation to vection is not a novel concept. Riecke and Schulte-Pelkum (2013) noted that it was Helmholtz who suggested as early as “…1866 that vibrations and jerks that naturally accompany self-motions play an important role for self-motion illusions, in that we expect to experience at least some vibrations or jitter...”. Furthermore, the potential of tactile stimulation in eliciting or enhancing vection has been highlighted in the well-cited work of Dichgans and Brandt (1978) over four decades ago. However, tactile-vection research does not appear to be as prevalent in literature as, for example, vection research using visual and auditory stimulation. Visual, vestibular, and auditory vection have been reviewed extensively (see Britton & Arshad, 2019; Dichgans & Brandt, 1978; Hettinger et al., 2014; Lappe et al., 1999; Väljamäe, 2009) and Harris et al. (2002) briefly discuss the relation of gravitoinertial forces and proprioception to vection. Additionally, the works of Amemiya (2018) and Lécuyer (2017) contain a discussion on the utility of tactile stimulation in vection research, however, the content is brief and mainly focussed on the research output by their respective groups. Lastly, Hettinger et al. (2014) discussed the use of haptic and tactile cues in the genesis of vection, however, it comprised of only one page. Thus, to the best of our knowledge, no extensive review is present in literature on tactile-mediated vection. The aim of the current article is to review the literature on the methodologies around tactile stimulation in vection research. Furthermore, we aim to qualitatively and quantitively, through a meta-analysis, assess the outcomes of tactile stimulation in vection research. The article is structured as follows: our research questions, search methodology and statistical methods are detailed