Beyond Human: Animals as an Escape from Stereotype Avatars in Virtual Reality Games
Andrey Krekhov, Sebastian Cmentowski, Katharina Emmerich, Jens Krüger
BBeyond Human: Animals as an Escape fromStereotype Avatars in Virtual Reality Games
Andrey Krekhov , Sebastian Cmentowski , Katharina Emmerich , Jens Krüger High Performance Computing Group, Entertainment Computing GroupUniversity of Duisburg-Essen, Germany{andrey.krekhov, sebastian.cmentowski, katharina.emmerich, jens.krueger}@uni-due.de
Figure 1. We explore the potential of nonhuman avatars in VR games. The evaluation of our three escape room games for different animal types revealsthat players enjoy the control over additional body parts, as such morphologies allow novel, refreshing interactions and enable superhuman abilities.
ABSTRACT
Virtual reality setups are particularly suited to create a tightbond between players and their avatars up to a degree wherewe start perceiving the virtual representation as our own body.We hypothesize that such an illusion of virtual body ownership(IVBO) has a particularly high, yet overlooked potential fornonhumanoid avatars. To validate our claim, we use the exam-ple of three very different creatures—a scorpion, a rhino, and abird—to explore possible avatar controls and game mechanicsbased on specific animal abilities. A quantitative evaluationunderpins the high game enjoyment arising from embodyingsuch nonhuman morphologies, including additional body partsand obtaining respective superhuman skills, which allows usto derive a set of novel design implications. Furthermore, theexperiment reveals a correlation between IVBO and gameenjoyment, which is a further indication that nonhumanoid
Permission to make digital or hard copies of all or part of this work for personal orclassroom use is granted without fee provided that copies are not made or distributedfor profit or commercial advantage and that copies bear this notice and the full citationon the first page. Copyrights for components of this work owned by others than theauthor(s) must be honored. Abstracting with credit is permitted. To copy otherwise, orrepublish, to post on servers or to redistribute to lists, requires prior specific permissionand/or a fee. Request permissions from [email protected].
CHI PLAY’19,
October 22–25, 2019, Barcelona, Spain© 2019 Copyright held by the owner/author(s). Publication rights licensed to ACM.ISBN 978-1-4503-6688-5/19/10. . . $15.00DOI: https://doi.org/10.1145/3311350.3347172 creatures offer a meaningful design space for VR games worthfurther investigation.
CCS Concepts • Human-centered computing → Virtual reality; • Softwareand its engineering → Interactive games;
Virtual worldssoftware;
Author Keywords
Animal avatars; virtual creatures; animal embodiment; IVBO;virtual reality games; avatar control.
INTRODUCTION
The choice of our virtual representation, our avatar, has astrong influence on how we perceive a game. Hence, introduc-ing novel avatar kinds, beyond stereotypes such as knights andwizards, is a viable option to create refreshing and engagingplayer experiences. This choice applies even more for virtualreality (VR) games, because such immersive setups are capa-ble of amplifying the bond with our virtual self. That bondcan be strong enough such that we start perceiving the virtualrepresentation as our own body—a phenomenon also knownas the illusion of virtual body ownership (IVBO) [61]. a r X i v : . [ c s . H C ] J u l y a smart choice of avatars, VR games could allow us tocollect impressions and experiences that would not be possi-ble or would be far less engaging in a nonimmersive setup.One prominent example is games focused on nonhumanoidcreatures, be it real animals or mythical creatures. Eventhough players enjoy “beastly” non-VR games, such as Black& White [32] and
Deadly Creatures [44], similar scenarios areoffered very rarely. Especially in VR, where presence and theIVBO effect could significantly intensify our experience whenusing animal abilities, games like
Eagle Flight [68] remain anexception.We see manifold reasons why that potential remains unful-filled, including the very few studies on creature embodimentin VR, which makes it difficult for game designers to predictwhether and how players will perceive animal avatars. Further-more, as only a few games have touched upon this topic, bestpractices and design guidelines for such avatars are lacking.In other words, we need further research to understand thechallenges and opportunities induced by the nonhuman mor-phology, e.g., additional limbs and their influence on IVBO,differing postures, and possible control approaches.Our paper makes two contributions. First, we explore nonhu-manoid avatars in VR using escape room games built aroundthree very different animals: a rhino, a scorpion, and a bird(cf. Figure 1). Each game explores a different control mech-anism and focuses on distinct “superhuman” skills that aretypical for these animals. Our evaluation underpins the re-sulting high player enjoyment, especially from these animalabilities and additional body parts, such as horns, tails, orwings. Accordingly, we draw design implications for animalavatars and present our lessons learned during the design ofsuch VR games.Our second contribution is the investigation of IVBO in suchscenarios. We study how the nonhuman morphology influ-ences our ability to embody such avatars in VR games. Inparticular, our evaluation reveals correlations between IVBO,game enjoyment, and presence, and confirms that additionalbody parts and skills are not an obstacle for inducing IVBO.Hence, we assume that our work will motivate researchersand practitioners to reconsider IVBO-enabled nonhumanoidavatars as an important component of player experience inVR.
RELATED WORK
As our research targets virtual environments, we begin with abrief introduction of the related VR terms before focusing onthe embodiment of nonhumanoid avatars. Nowadays, VR hasregained attention mostly because of affordable mainstreamHMDs, such as HTC Vive [11], which allow players to experi-ence games from a novel perspective. Thereby, researchers [5,56] usually refer to immersion [9] as the technical quality ofVR equipment and apply the term presence [62, 58] to de-scribe the impact of such devices on our perception. In ourcase, we are particularly interested in presence, which canbe measured as proposed by, e.g., IJsselsteijn et al. [17] andLombard and Ditton [34]. Immersive technologies not only allow us to experience sucha “feeling of being there” [16], but also increase our abilityto emphasize our virtual self-representation. We can embodyour avatar to a remarkable degree, which is also referred to asthe illusion of virtual body ownership (IVBO) [35], agency, orbody transfer illusion.IVBO originates in the effect of body ownership. The initialexperiments by Botvinick and Cohen [7] introduced the rubberhand illusion: the participant’s arm was hidden and replaced byan artificial rubber limb, and stroking both the real and virtualarms created the illusion of actually owning that artificiallimb. After further investigations [67], researchers proposed anumber of models [66, 13, 43, 30] to explain such an interplaybetween external stimuli and our internal body perception.Slater et al. [59] and Banakou et al. [3] transferred the originalbody ownership effect, including the underlying visuotactilestimulation, to virtual environments. However, in their laterwork, Slater et al. [61] and Sanchez-Vives et al. [53] revisitedthe stimuli correlations and concluded that sensorimotor cuesare more important than the visuotactile cues, which is animportant insight, as VR setups seldom include tactile stimu-lations. To complete the picture, apart from visuotactile andsensorimotor cues, the IVBO effect is mainly impacted by vi-suoproprioceptive cues (perspective, body continuity, postureand alignment, appearance, and realism) [60, 61, 42, 37].IVBO was mainly explored with anthropomorphic charactersand realistic representations [35, 31, 18]. For instance, relatedto the question of avatar customization in games, Waltemateet al. [69] showed that customizable representations lead tosignificantly higher IVBO effects.A strong IVBO can produce various changes in player behav-ior [20, 38], resembling the Proteus Effect by Yee et al. [76].For instance, the work by Peck et al. [41] revealed a signif-icant reduction in racial bias when players embody a blackcharacter. Similarly, virtual race can also affect the drummingstyle [22]. Other reactions are childish feelings arising fromembodying child bodies [3] and an increase in perceived sta-bility when having a robotic avatar [36]. Hence, prior workindicates that IVBO can be used to evoke specific feelings andattributes [25]. We suggest that a strong bond to the creaturecaused by IVBO can also increase our involvement with envi-ronmental issues [1, 4] and our empathy for animals, which,in turn, is transferable to human-human empathy, as shown byTaylor et al. [64].Researchers have also expressed interest in studying IVBObeyond human morphology. For instance, Riva et al. [48]posed the following question:
But what if, instead of simplyextending our morphology, a person could become somethingelse- a bat perhaps or an animal so far removed from thehuman that it does not even have the same kind of skeletonâ ˘A ˇTan invertebrate, like a lobster?
Interestingly, embodying a batis even being discussed in philosophy [39]. If we considerexotic body compositions, as in the case of a lobster, that havefew properties in common with our human body, the idea ofsensory substitution [2] might play an important role. Onemight also consider such substitution mechanisms as playfulnteractions: e.g., the echolocation feature of a bat could bereplaced by tactile feedback in a VR game.Given the extreme diversity of real and fictional creatures, itis difficult or even impossible to research IVBO for virtualanimals as a whole. Instead, previous research tackled isolatedmodifications of body parts. For instance, Kilteni et al. [24]were able to stretch the virtual arm up to four times its originallength without losing IVBO. Normand et al. [40] used IVBOto induce the feeling of owning a larger belly than in reality.As a first step toward generalization, Blom et al. [6] concludedthat strong spatial coincidence of real and virtual body part isnot mandatory to produce IVBO.Certain animals, such as scorpions or rhinos in our study, haveadditional body parts that players might want to control. Inthis respect, prior work [14, 15] confirmed that having anadditional arm preserves IVBO and induces a double-touchfeeling. Steptoe et al. [63] reported effects of IVBO uponattaching a virtual tail-like body extension to the userâ ˘A ´Zsvirtual character. Clearly, these findings are relevant for aplethora of real and fictive nonhumanoids, such as dragons.The authors also discovered higher degrees of IVBO when thetail movement is synchronized with the real body.To remain briefly with the example of a dragon as an avatar:Egeberg et al. [12] proposed different ways wing control couldbe coupled with sensory feedback, and Sikström et al. [57] as-sessed the influence of sound on IVBO in such scenarios. Wonet al. [74] further analyzed our ability to inhabit nonhumanoidavatars that have additional body parts.Closely related to our research, the works-in-progress paperby Krekhov et al. [27] also suggested embodying virtual an-imals in VR games. In their preliminary, explorative study,the authors implemented different control approaches for vir-tual tigers, bats, and spiders, and reported tendencies thatIVBO remains intact for such avatars. We continue that workby building on the lessons learned regarding full-body andhalf-body control approaches, yet focus on embedding thisknowledge into games research.Naturally, we need a way to measure and compare IVBOstrength in order to investigate whether and how IVBO influ-ences player experience. In this regard, we point readers tothe recent work by Roth et al. [49] that introduced the alphaIVBO questionnaire based on a fake mirror scenario study.The authors suggested acceptance, control, and change as thethree factors that determine IVBO. As the study by Krekhovet al. [27] relied on this questionnaire to study animal em-bodiment in VR, we applied the same process to generatecomparable results.A body of literature related to the control of animal avatarsshould be mentioned in this context. Leite et al. [29] experi-mented with virtual silhouettes of animals that were used likeshadow puppets and controlled by body motion. For 3D cases,Rhodin et al. [45] applied sparse correspondence methods tocreate a mapping between player movements and animal be-havior and tested their approach with species such as spidersand horses. As a next step, Rhodin at al. [46] experimentedwith the generalization of wave gestures to create control possi-
Figure 2.
Bird Cage.
Players embodied a bird that was caught in acage and had to escape through the top right door (marked yellow). Theblue marked rods could be used for rests between the exhausting flightsperformed by full-body controls (bottom right). Finally, wind gusts hadto be created to turn a lamp into a wrecking ball (top right). bilities for, e.g., caterpillar crawling movements. Our researchextends these methods by presenting additional mechanismstailored to animal avatar control.
THE VIRTUAL ANIMAL EXPERIENCE
Our main goal is to understand the benefits and limitationsof animal avatars in VR games. Unfortunately, prior workindicates that we cannot overgeneralize such research, becauseanimals vary greatly among themselves, be it regarding theirskeletons, or postures, or motion. Hence, we focus on a soundmethodology for a few sufficiently distinct representatives andprovide in-depth insights how such avatars can be embeddedin a gaming context. In particular, this section describes ourreasoning regarding the choice of animals and their controls,as well as a quantitative evaluation of the outcomes.
Choosing Virtual Animals
One of the main questions to be asked when designing a gamewith nonhumanoid avatars is which creature to pick. Ob-viously, this choice is determined by various game designaspects that are not specific to animal avatars. However, theinclusion of such creatures adds degrees of freedom that needto be considered. We focus on two main aspects: the increasedinteraction design space and the induced challenges in control-ling such avatars.In the first place, playing an animal allows us to naturallyinhibit the respective superhuman skills, such as flying as abird or exploring underwater scenes as a dolphin. We postulatethat such natural interactions could be intuitive and easy tolearn when done right. Furthermore, the IVBO effect canintensify [3, 36] our perception of such actions due to theincreased bond to our avatar compared to non-VR games.These additional skills are often bound to additional body partsof nonhumanoid creatures. Fortunately, prior work [74] indi-cates that such additions do not necessarily destroy the IVBO igure 3.
Rhino Room.
Players had to mimic the rhino posture (left) and escape from a burning zoo. The blue marked water tap (middle) had to beremoved from the wall (right) to extinguish the fire. To open the yellow marked door, players had to use the horn and remove a lock bar (cf. Figure 1). effect and can still be intuitively controlled by players [12]. Inthis respect, we recommend designing the avatar such that thealtered morphology is perceived as an extension to our body,instead of being a restriction. For instance, Krekhov et al. [27]reported that players liked the large wings of a bat becausethey felt like arm extensions and helpful tools, but dislikedtiger paws that felt shorter than their actual limbs.A second important aspect to be considered when designingsuch games is how the creature should be controlled by theplayer. To embody animal avatars, it is reasonable to synchro-nize the movements of the players as precisely as possiblewith their virtual representation [53]. However, typical room-scale VR equipment tracks only the players’ heads and hands.We see three approaches to overcome that barrier: relyingon only three tracked positions, including markerless track-ing [75], or requiring tracking extensions, such as the HTCVive Trackers [11].Even when full-body tracking is available, the question still re-mains how animals with significantly different postures shouldbe controlled. A prominent example is creatures with non-upright postures, such as typical mammals. A straightforwardway would be to crawl on all fours as a player to achieve themost realistic mapping. However, this might cause exhaus-tion over a longer period of time. As a remedy, half-bodycontrols [27] can be applied to remain in an upright posturewithout noticeable sacrifice of IVBO. Half-body mapping ap-proaches have either no direct mapping between players’ legsand the limbs of an animal at all, or one leg is mapped tomultiple limbs, which allows us to control creatures like spi-ders. Apart from fatigue, such controls are beneficial for caseswhere full-body tracking is not available.To summarize, finding an optimal virtual creature is a mul-tifaceted process, and we suggest the impact of additionalbody parts and resulting superhuman skills, as well as possiblecontrol approaches be considered during the decision-making.To illustrate that process in more detail, we will showcase apossible selection approach of avatars and game mechanics inthe next section.
Example Realization
To study animal avatars in a game context, we created a diversetestbed that supports multiple creatures with Unity 3D [65].The main idea is based on so-called escape rooms [70]: play-ers are placed in a room filled with challenges that have to be solved in order to win/escape. We picked that setting for twoparticular reasons. First, if virtual and real rooms match in sizeand shape, locomotion can be achieved by natural walking,which has a positive impact on presence [50, 26] and removesthe need for additional, artificial navigation techniques, suchas teleportation. Hence, players can focus more on the actualanimal experience and are less distracted by accompanyingfunctionalities. Second, escape room games are similar intheir concept, which allows us to implement multiple, yet com-parable scenarios, i.e., different rooms with different animalavatars. Each room contained two to three quests that involvenavigation and object manipulation. In contrast to common es-cape games, we did not impose any time limitation to removecompetition as a factor from our studies.After picking the overarching game type, we focused on thedesign of the underlying game mechanics. We set ourselvesthe objective of building the individual room quests arounddistinct animal abilities. We selected three animals based theirsuperhuman skills and/or additional body parts: a rhino, ascorpion, and a bird. In particular, our selection includedmorphologies with different degrees of similarity comparedto our human body. A bird has a straightforward mapping,i.e., our arms become wings, and our legs become bird’s feet.The horn of the rhino has no direct counterpart and requiresa head-oriented interaction that is exotic for human beings.Finally, the scorpion comes with additional limbs, a tail with asting, and two claws, which is the most differing morphologywith at least two nonhumanoid interactions.
Rhino Room
We chose a rhino mainly because of its horn and the relatedcapabilities. We suggest that such head-centered interactionsoccur seldom in VR games and could offer a unique playerexperience. In our case, players should use the horn (andpaws) to escape from a burning cage, as shown in Figure 3. Inparticular, the horn was needed to move crates and clear thearea in front of a water tap, to remove the tap from the wall torelease a jet of water, and, finally, to lift and remove a lock barthat kept the door closed.We utilized full-body controls with 1:1 motion mapping, i.e.,players had to crawl on all fours during the game. Therefore,we positioned additional trackers at the hip and both anklesand wrists, i.e., no Vive controllers were used. We relied oninverse kinematics (IK) [8] to reconstruct the player posture. igure 4.
Scorpion Room.
Players remained in an upright posture (left) and used the controllers to open and close the claws and initiate a tail strike.To escape from the labyrinth, players had to cut away several branches (right). The exit-blocking emperor scorpion (middle) had to be pelted withpoisoned fruits. The avatar tail was used to pick up these fruits. Aiming during the throwing process was done via a proper hip orientation.
In particular, we applied a combination of closed-form anditerative solvers to provide the required degrees of freedomyet minimize jittering caused by unavoidable tracking errors.The horn was always visible to the players and placed slightlybelow the camera, as can be seen in Figure 3.
Scorpion Room
A scorpion offers even more unique interactions comparedwith a rhino if we allow players to control its tail and claws. Inour scenario, depicted in in Figure 4, players had to use thesetechniques to cut their way through a labyrinth and defeat agiant enemy by throwing a poisoned fruit at it (cf. Figure 1).The fruit had to be picked up and thrown via the sting at theend of the scorpion tail.To explore a variety of control approaches, we relied on half-body tracking, i.e., an upright posture, instead of 1:1 mapping,as done in the rhino game. We used the tracking data fromthe HMD, two Vive controllers, and an additional tracker atthe hip position. Player arm movement was transferred to thevirtual claws via IK. Trigger buttons could be used to openand close the claws to perform cutting. The circular track padbutton initiated a tail strike, whereas aiming was performedby hip alignment. We did not track players’ legs. Hence, thelimbs of the scorpion were equipped with predefined “walking”animations matching the speed of player movement.
Bird Cage
To complete the diverse set of our virtual animals, we alsoincluded a flying creature, as can be seen in Figure 2. Being abird, players could use their virtual wings for two purposes:flying and creating gusts of wind to move objects. To escapefrom their virtual cage, players had to gain altitude, reach thehighest point, and flutter with their wings in sync with themovement of a ceiling lamp, which then gained momentumand broke the cage door. Gaining altitude required significanteffort, and players had to rest on rods between their flights.We used the same tracking setup as in the rhino game, i.e.,trackers at hip, wrists, and ankles. Players remained in anupright posture, and their arms were mapped to the wings, i.e.,flapping was achieved via rapid up and down arm movements.To create gusts, players moved their arms horizontally instead.Flying around in the cage consisted of two components: flap-ping to gain altitude, and walking to perform a horizontal transition. We explicitly enforced that horizontal physicalmovement to minimize cybersickness [28] by reducing thecognitive mismatch between physical and visual feedback. Ifplayers stopped waving their arms mid-air, a “falling” proce-dure was applied. That transition was performed rapidly toprevent cybersickness [26].
EVALUATIONResearch Questions and Hypotheses
The main purpose of our study was to investigate how playersexperience the animal avatars in our three game scenariosto draw conclusions about which aspects of representation,control, and interaction are perceived positively or negatively.Accordingly, our main research questions are:1: Do animal avatars induce positive player experiences?2: How do players evaluate the different design decisionsregarding posture, visible body parts, and control map-ping in our three games?We assume that slipping into the role of an animal is anovel and interesting experience, and that the control of non-humanoid body parts and the use of related special abilities canraise players’ enjoyment and engagement. Our three differentrealizations allow us to investigate whether a realistic postureand locomotion technique (e.g., crawling), the visibility ofcertain body parts, and the type of control mapping contributeto or interfere with a positive experience.Besides the general acceptance of animal avatars and the eval-uation of the respective player experiences, we also considerthe concept of IVBO. Based on prior findings indicating thatIVBO is not limited to human-like bodies [74, 27], we hypoth-esize that our virtual animal bodies are capable of inducingIVBO as well, and that higher IVBO can be associated withhigher perceived presence and game enjoyment. Hence, wewant to test the following hypothesis:H1: IVBO is positively correlated with game enjoymentand perceived presence.
Study Procedure and Applied Measures
We applied a within-subjects design with the game scenarioas the independent variable with three levels (rhino, scorpion,ird). After being informed about the study procedure andsigning an informed consent, participants filled in a first ques-tionnaire about their demographic data, gaming habits, andprior experiences with VR headsets. We also administered theImmersive Tendencies Questionnaire (ITQ) [72] to check par-ticipants’ individual tendencies to get immersed in an activityor fiction.We then introduced the participants to the first game scenario.The three games were played in varying order to avoid biasesdue to sequence effects. In particular, we counterbalancedthe sequence of the three game scenarios across subjects. Allsessions followed the same procedure. First, the examinerexplained the goal and controls of the game and applied theVR headset, an HTC Vive Pro [10] with a wireless adapter,and HTC Vive Trackers [11]. Subsequently, a neutral mirrorscene was started, in which participants saw their animal bodyavatar and were able to get used to the controls by observingtheir movements in a big mirror, as can be seen in Figure 5.This scene was displayed for 2 minutes to enable embodiment.The duration is a common choice for IVBO studies, and priorwork indicates that even 15 seconds are enough to cause thateffect [33]. After the mirror scene, we asked the participants toremove the HMD and administered the acceptance and controlsubscales of the alpha IVBO questionnaire [49]. We weremainly interested in the IVBO experience while playing andnot in the subsequent effects on players’ bodily perception,so the change dimension of the IVBO questionnaire was notapplied. We decided against performing threat tests for captur-ing IVBO, because we expected significant sequence effects.Note there is no unified procedure for measuring IVBO, and athreat test is not the only possibility [23, 49]. We decided touse the alpha IVBO questionnaire and checked its reliabilityby calculating Cronbach’s alpha for both subscales (all alphas> 0.82).Upon completion, we re-equipped the participants with theHMD and launched the main game. Each gaming sessionlasted about 7 to 10 minutes, depending on how quickly play-ers were able to solve all riddles. After each session, we askedthe participants to fill in a questionnaire asking about theirexperiences during play. We administered the enjoyment sub-scale of the Intrinsic Motivation Inventory (IMI) [51] to assessgeneral game enjoyment, as well as the Player Experience ofNeed Satisfaction (PENS) questionnaire [52, 47, 19] to testexperienced autonomy, competence, and intuitiveness of con-trols. We measured the feelings of presence by the PresenceQuestionnaire [73, 71] and the Igroup Presence Questionnaire(IPQ) [55, 54]. To test for negative physiological effects ofusing the immersive HMD, we also administered the Simu-lator Sickness Questionnaire (SSQ) [21]. Finally, we posedsome custom, game-specific questions to assess how playersevaluated the controls, the required posture during play, aswell as the visibility of certain body parts. We also askedwhether participants could imagine using this kind of avatarcontrol in other VR games. All administered questionnaireitems had to be rated on a unipolar scale ranging from 0 to 6(“completely disagree” to “completely agree”), except fromthe SSQ, which had to be rated on a unipolar 4-point scale.
Figure 5. Before each game, players were asked to act in front of awall-sized mirror for about two minutes to get familiar with their virtualrepresentation and to answer the alpha IVBO questionnaire [49].
Results
In total, 32 persons (19 female, 13 male) with a mean ageof 23.7 years ( SD = 5.18) participated in our study. Due torecruiting at a university, most of them were students ( N = 25),whereas the others were employees. Many participants re-ported prior experiences with VR headsets ( N = 22), but onlytwo of them used VR systems on a regular basis. All par-ticipants were familiar with digital games and the majority( N = 24) reported playing digital games regularly. Players’ Experiences with the Three Animal Avatars
Following our research questions, we analyzed participants’ratings of the different animal avatars and their experiencesin the three game scenarios. Mean values of all applied ques-tionnaires can be found in Table 1. Considering the scales’range from 0 to 6, almost all aspects were rated above average,indicating a positive experience in all three game scenarios. Inparticular, IMI scores and perceived presence as measured bythe PQ show that players enjoyed the games and felt as if theywere actually being and acting in the virtual world. Scores hino Scorpion Bird M ( SD ) M ( SD ) M ( SD ) F χ p IMI
Enjoyment/Interest 4.46 (0.96) 4.17 (1.37) 4.08 (1.36) - 0.065 .968
PENS
Competence 3.66 (1.85) 3.03 (1.85) 3.33 (1.66) - 3.000 .223Autonomy 3.82 (1.51) 3.27 (1.53) 3.14 (1.62) 3.562 - .034 *Intuitive controls 5.14 (1.15) 4.50 (1.74) 4.53 (1.35) - 11.608 .003 * PQ Realism 4.09 (1.15) 3.93 (1.19) 3.92 (1.09) 0.493 - .613Possibility to act 4.55 (0.83) 3.98 (0.99) 3.85 (1.06) 7.607 - .001 *Quality of interface 4.91 (1.02) 4.54 (1.21) 4.97 (1.02) 4.142 - .020 *Possibility to examine 4.44 (1.05) 3.89 (1.29) 3.72 (1.24) - 22.709 < .001 *Self-evaluation of performance 4.53 (1.51) 4.19 (1.44) 4.34 (1.24) - 4.019 .134Total 4.42 (0.92) 4.06 (1.02) 4.09 (0.91) 4.162 - .020 *
IPQ
General 3.88 (1.52) 3.56 (1.68) 3.63 (1.70) - 2.092 .351Spatial presence 4.22 (1.19) 4.07 (1.24) 3.94 (0.99) 1.339 - .270Involvement 3.34 (1.28) 3.02 (1.21) 2.96 (1.26) 2.373 - .102
IVBO
Acceptance 3.34 (1.15) 3.26 (1.29) 3.46 (1.31) 0.502 - .608Control 4.64 (1.42) 4.40 (1.53) 4.88 (1.26) - 3.576 .167
SSQ
Nausea 0.21 (0.32) 0.27 (0.38) 0.26 (0.40) - 2.590 .274Oculomotor 0.24 (0.27) 0.35 (0.45) 0.26 (0.34) - 9.968 .007 *Disorientation 0.11 (0.18) 0.19 (0.41) 0.17 (0.42) - 3.405 .182 * significant main effect at a significance level of α = .05 Table 1. Mean scores and standard deviations of the IMI, PENS, PQ, IPQ, IVBO, and SSQ subscales for the three game scenarios (all scales rangefrom 0 to 6, except from SSQ, which ranges from 0 to 3). Significant differences of mean values between conditions were tested by calculating repeatedmeasures ANOVA ( F ) or Friedman tests ( χ ), if data was not normally distributed. of all three subscales of the SSQ—nausea, oculomotor, anddisorientation—were very low in all conditions (all M < 0.36),and thus cybersickness was not an issue and can be excludedas a potential confounding variable.We compared players’ experiences in the three game scenariosin terms of the subscales of IMI, PENS, PQ, IPQ, and IVBO toinvestigate whether the different avatars and interactions wereperceived differently. Our analysis of covariance indicated nosignificant influence of immersive tendencies (ITQ) on ourdependent variables, hence we did not further elaborate onthat. In advance, we performed Kolmogorov-Smirnov teststo assess all scales for normal distribution as a requirementfor parametric calculations. If violated, results of Friedmantests are reported instead of repeated measures ANOVA forcomparing the three game scenarios. Bonferroni correctionwas applied for all post-hoc tests. The main test statistics canbe found in Table 1.Regarding players’ need satisfaction (PENS), we found signif-icant differences between the three game scenarios in terms ofautonomy and intuitive controls. Post-hoc tests indicate thatperceived autonomy was significantly higher when playingwith the rhino compared to the scorpion scenario ( p = .049).The intuitiveness of controls was rated significantly higher inthe rhino scenario than in both the scorpion ( p = .012) and thebird condition ( p = .004). Comparisons of the PQ subscales show further significantdifferences. According to post-hoc tests, participants reportedsignificantly better experiences regarding the possibility toact and the possibility to examine in the rhino scenario thanin the other two games (all p < .004). Moreover, the totalscore for presence was significantly higher for the rhino thanfor the scorpion ( p = .017). The interface quality, in contrast,was rated significantly lower in the scorpion game comparedto the bird scenario ( p = .049). All other measures did notdiffer significantly, i.e., we did not find significant differencesregarding general game enjoyment or IVBO. Insights About the Different Postures
As our different scenarios required different postures, we askedhow the actual gaming posture was perceived and if partic-ipants would have preferred another posture. For the birdand scorpion avatars, participants agreed that the upright pos-ture was comfortable (bird: M = 4.50, SD = 1.48; scorpion: M = 5.03, SD = 1.26), whereas the kneeling posture in therhino condition was rated ambiguously and perceived as be-ing physically demanding by several participants ( M = 3.44, SD = 1.98). However, when asked whether they would preferan upright playing posture to control the rhino, the majority ofplayers tended to disagree ( M = 2.56, SD = 2.41). In contrast,they agreed that the kneeling posture contributed to the realismof the game ( M = 4.16, SD = 1.99). MI PQ IPQ e n j oy m e n t r ea li s m po ss i b ilit y t o ac t i n t e rf ace qu a lit ypo ss i b ilit y t o e x a m i n e p e rf o r m a n ce t o t a l g e n e r a l s p a ti a l p r e s e n ce i nvo l v e m e n t r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) r s ( p ) IVBO - Rhino
Acceptance 0.068(.710) 0.511*(.003) 0.312*(.049) 0.480*(.005) 0.404*(.022) 0.329(.066) 0.508*(.003) 0.192(0.291) 0.354*(.047) 0.046(.805)Control 0.442*(.011) 0.734*(<.001) 0.557*(.001) 0.689*(<.001) 0.612*(<.001) 0.467*(.007) 0.760*(<.001) 0.559*(.001) 0.590*(<.001) 0.022(.903)
IVBO - Scorpion
Acceptance 0.357*(.045) 0.428*(.015) 0.438*(.012) 0.272(.132) 0.461*(.008) 0.399*(.024) 0.460*(.008) 0.560*(.001) 0.496*(.004) 0.349(.051)Control 0.572*(.001) 0.742*(<.001) 0.563*(.001) 0.562*(.001) 0.642*(<.001) 0.586*(<.001) 0.769*(<.001) 0.460*(.008) 0.504*(.003) 0.190(0.297)
IVBO - Bird
Acceptance 0.397*(.024) 0.481*(.005) 0.585*(<.001) 0.451*(.010) 0.557*(.001) 0.391*(.027) 0.618*(<.001) 0.436*(.013) 0.480*(.005) 0.237(.192)Control 0.387*(.029) 0.358*(.044) 0.412*(.019) 0.591*(<.001) 0.478*(.006) 0.324(.071) 0.501*(.004) 0.474*(.006) 0.520*(.002) 0.163(.374) * significant correlation at a significance level of α = .05 Table 2. Spearman’s rank-order correlation coefficients r s and p -values that indicate correlations among the IVBO subscales and IMI, PQ, and IPQ. The bird posture and the mechanics of locomotion (flappingwith the arms to move up combined with walking to movehorizontally) was also perceived as being realistic ( M = 4.84, SD = 1.08). Accordingly, participants did not wish for anotherposture ( M = 1.09, SD = 1.65).Similar ratings were given for the scorpion: participants ratedthe posture as being realistic ( M = 4.13, SD = 1.95) and didnot wish for another posture such as kneeling ( M = 1.78, SD = 2.01), although a kneeling posture would be objectivelymore realistic. When asked whether they had the feelingof being stuck in the ground (due to the low head position),participants were rather inconclusive ( M = 3.25, SD = 2.17).During the experiment, we observed that some participantswere indeed a bit irritated at the beginning, but got used to themismatch between their own and the avatar’s body size quitequickly. Controls
Overall, the high ratings for PENS’ intuitive controls confirmthat participants had no problems moving and interacting inthe game world and using the animals’ abilities in all three sce-narios. Although our three animal avatars are rather differentin terms of posture and control mapping, participants statedin all three cases that they could very well imagine using thiskind of avatar control in other VR games (rhino: M = 4.56, SD = 1.78; scorpion: M = 4.94, SD = 1.44; bird: M = 4.75, SD = 1.55). Visibility of Body Parts
Regarding the visibility of certain body parts, we were inter-ested in players’ opinions about the usefulness of such visu-alizations and the possible interferences. In the rhino game,the horn was displayed in the players’ sight throughout thegame. However, the horn was neither perceived as being dis-ruptive ( M = 0.69, SD = 1.18) nor resulted in the perceptionof a constrained field of view ( M = 0.97, SD = 1.26). In con-trast, participants enjoyed using the horn as a tool ( M = 4.78, SD = 1.60).In the bird cage scenario, apart from the wings, the bird’sfeet were also displayed. Participants appreciated this display,because they rated this feature as being helpful for landing onthe rods ( M = 5.22, SD = 1.49). Special Abilities
We also asked players about their opinions regarding the spe-cial abilities they could use as animals. Participants agreed thatthe use of the horn of the rhino enriched the whole experience( M = 4.91, SD = 1.33). The scorpion’s sting was rated as veryinteresting ( M = 4.44, SD = 1.59), and players also liked touse the claws ( M = 3.81, SD = 1.86). Moreover, players ratedhe experience of flying as a bird as very interesting ( M = 5.12, SD = 1.36) as they did the bird’s ability to create gusts of wind( M = 4.19, SD = 1.93). Correlations between IVBO, Enjoyment, and Presence
Mean values of IVBO control are rather high, and mean val-ues of IVBO acceptance are above average, as well, whichindicates that players have experienced IVBO while control-ling our animal avatars. To test our hypothesis regarding therelation between IVBO and the player experience (H1), weanalyzed the correlations between the two subscales of theIVBO questionnaire and the subscales of IMI, PQ, and IPQ.We calculated Spearman’s rank correlation coefficients (Spear-manâ ˘A ´Zs rho) due to a lack of normal distribution of somescales. Table 2 summarizes the results for each game scenario.Overall, we found significant positive correlations betweenIVBO and nearly all PQ and IPQ subscales: ratings of experi-enced realism, the possibility to act, the possibility to examine,PQ total, and spatial presence are consistently significantlycorrelated with both IVBO dimensions in all three games. Fur-thermore, IVBO control also significantly correlates with theperceived interface quality and the general feeling of presenceas measured by the IPQ. IVBO acceptance correlates with theinterface quality except from the scorpion scenario, and withgeneral presence except from the rhino scenario. The onlyscale not significantly correlated with IVBO in any scenario isIPQ involvement.Regarding game enjoyment, our analysis shows significantpositive correlations between IVBO control and IMI enjoy-ment scores in all three scenarios. The correlation betweenIVBO acceptance and enjoyment is significant in the birdcage and the scorpion room, whereas there is no correlationin the rhino condition. In sum, our results mainly support ourhypothesis H1.
Discussion and Design Implications
Our results indicate that animal avatars in VR games caninduce positive player experiences. We implemented threegames with animal avatars that are very different regardingbody features and abilities, and in all cases players reportedhigh enjoyment and high presence, i.e., the feeling of actuallybeing in the virtual world and being the rhino, scorpion, orbird. Participants particularly appreciated the novel body ex-periences and nonhumanoid perspectives, as well as the use ofthe special animal abilities.
Special abilities
The feedback of participants on our three games shows thatplayers are very interested in performing actions that they arenot able to perform in real life. For instance, they were fasci-nated by the ability to fly upwards using their wings as a bird,and they enjoyed testing how they could manipulate objectswith their rhino horn. We reason that such superhuman abil-ities significantly contribute to players’ enjoyment and theirmotivation to play. Hence, the main game mechanics of gamesfeaturing animal avatars should foster the animal’s specificcharacteristics and abilities to create novel, fanciful experi-ences. Designers should take advantage of players’ curiosityand expose unique animal features.
Player Posture and Controls
In all three games, the adopted postures were perceived posi-tively and without an explicit desire for alternatives. In otherwords, there is no indication that a realistic yet uncomfortableposture (rhino) is better or worse than a convenient and uprightbut unrealistic posture (scorpion). However, the statementsregarding crawling on the floor as a rhino were quite ambiva-lent, i.e., some of the participants enjoyed such an experience,whereas others became rapidly exhausted by that activity. Notethat the rhino, however, outperformed other animals in cer-tain subscales, such as autonomy (PENS), intuitive controls(PENS), and the possibility to act (PQ). Hence, a 1:1 mappingwhere players have to behave exactly like they would expectfrom their animal avatar is easier to grasp and is perceived asvery realistic.Similarly, our results did not disqualify or favor any particularcontrol approach – all three controls were rated as very intu-itive and participants could imagine using such approachesin other VR games. Hence, we suggest controls be designedbased on the game-related animal abilities and the targetaudience . For instance, we assume that children are morewilling to spend their time crawling on the floor compared toelderly adults. In general, transferring as many player move-ments as possible onto the avatar is a reasonable approach,especially considering the positive influence on IVBO [53].However, as we have seen in the scorpion case, less straight-forward mappings can be equally engaging and fun withoutenforcing an uncomfortable posture. Furthermore, such imple-mentations can be achieved with less tracking equipment.
Visible body parts
Game designers have different approaches regarding the visi-bility of the avatar’s body in first-person mode. From our expe-rience, we would not recommend visualizing the whole body,as the avatar head position often leads to confusing viewportswhen players look down on them. Instead, we suggest thevisualization be limited to body parts that can be directlycontrolled by the player , e.g., claws, tails, wings, and horns.In particular, the additional body parts, although reducing thevisible area, are not perceived as disturbing. For instance,participants rated the horn as a helpful tool and reported thatthe bird’s feet facilitate the landing on thin rods. Furthermore,seeing animal body parts like claws moving in sync with ourown body increases our awareness of embodiment.
Morphology
Considering the morphology of our three animal avatars, ourresults indicate that players had no problems with controllingbodies that are not similar to the human shape. Even the con-trol of the scorpion, which has several additional limbs, claws,and a tail, was perceived as intuitive and did not cause any con-fusion. In contrast, we observed that players particularly likedadditional body parts such as the scorpion’s sting or the hornof the rhino. Hence, we challenge game designers to considerextraordinary animal shapes and derive innovative gamemechanics. We should not back off from adopting complexbody compositions as long as they are associated with inter-esting possibilities for interaction design. In our three games,we always focused on the outstanding bodily features of thenimals and linked them to certain player abilities (e.g., creategusts of wind) to give significance to them . We suggest thatadditional or missing body parts compared to the human bodyshould enrich players’ opportunities to examine and interactwith the virtual world and not appear as an impairment. Thisway, we can foster players’ experience of having superhumancapabilities.
IVBO
Finally, we conclude that additional body parts or a nonhumanbody shape do not inhibit an avatar’s potential to induce IVBO.Our three exemplary animal avatars illustrate that IVBO isnot limited to body models that are similar to the humanbody. With regard to our hypothesis H1, our results reveal thatIVBO—which was measured prior to the gaming sessions andis, thus, not biased by the subsequent game experience—ispositively correlated with game enjoyment and perceived pres-ence. This finding indicates that IVBO may contribute to apositive player experience. Hence, we conclude that
IVBOis a considerable factor when designing nonhumanoid VRavatars . To foster IVBO in a game, we suggest that gamedesigners provide players with possibilities to see their vir-tual body (e.g., in mirrors or water reflections) to increaseawareness of their virtual representation.
LIMITATIONS OF THE STUDY
Our derived design implications are based especially on thethree evaluated scenarios. Hence, we need to consider a set ofassociated limitations to prevent possible misinterpretation ofthe findings. In the first place, our main goal was to expose acomplete pipeline of embedding animal avatars into VR games.We aimed to raise the awareness regarding the wide varietyof decisions (e.g., posture, animal type, mapping/controls,special abilities, morphologies, locomotion) that have to beconsidered during such a game development process. As aresult, our evaluated scenarios are rather complex games witha number of possibly influential variables that might limit thegeneralizability. For instance, the general appeal of an animal,e.g., a dangerous scorpion vs. a domestic bird, might impactour game enjoyment. Well-known species, such as a rhino,might be more intuitive to control than exotic creatures withabilities unknown to us. And although we removed artificialVR navigation techniques (e.g., teleportation) by matching thesize of the virtual environment to the physical room, the differ-ent locomotion (flying vs. crouching vs. walking) could stillhave a considerable impact on the player experience. Finally,although all three scenarios were escape games, the particularquests could have influenced the outcome. In other words,we emphasize that the direct comparison of the three studyconditions should be interpreted with these limitations in mindand that the reason behind our variety of scenarios was notthe comparison per se, but our strive to cover as much of theanimal avatar design space as possible to create a comprehen-sive starting point for future explorations. Also, comparativestudies in the future would benefit from an additional controlgroup with a human avatar for a better assessment of the ani-malsâ ˘A ´Z influence on IVBO and game enjoyment comparedto a rather traditional virtual representation. Another important limitation to be mentioned is that we didnot involve animal/domain experts during the design phase andpretests of our study. Our decision making regaring the choiceof animal avatars and, even more important, their abilities andinteractions, was made without the input of an expert. The lat-ter could have provided additional input regarding the realisticbehavior of animals and our perception of such species.As a next step, we suggest to focus on particular avatar compo-nents in a more targeted study to build a theoretical frameworkthat provides an isolated in-depth exploration of major factors,such as locomotion, altered or additional body parts, and ap-peal. We suggest that such isolated insights should be gatheredas a second step after seeing the â ˘AIJwhole pictureâ ˘A˙I, i.e.,how such animals work or do not work in games. For instance,prior work [27] reported that embodying a tiger while crawlingon all fours was disliked by the participants, whereas a rhino,being a very similar mammal, provided the highest enjoymentin our scenario. Hence, we suppose that it is not just the fa-miliarity with an animal or the intuitive locomotion, but rathersubtle details, e.g., the additional horn, that can significantlyalter our experience of such avatars and, thus, need furtherresearch.For further studies, we also recommend expanding the agerange of the participants. In our case, most participants werestudents due to the acquisition at the university, which limitsthe applicability of our findings. Instead, it is likely that as-pects such as the necessary physical effort or the perceivedavatar appeal are experienced differently by other age groups.Consequently, the age of the target audience might be an im-portant design consideration and should be explored in futurework.
CONCLUSION AND FUTURE WORK
Our work investigated the hidden potential of animal avatars.We focused on virtual reality games because of the relatedIVBO effect that allows us to embody our avatar and perceivecertain player interactions in a more intensive way. Accord-ingly, our studies supported our general assumption that gamescreated around animal avatars could lead to great enjoyment.In particular, players liked the interactions resulting from addi-tional body parts, such as wings and horns. In this regard, weproposed different ways to control animals with such differingmorphologies and discussed related design implications foranimal-centered VR games.As a particular finding, we reported a correlation among IVBO,presence, and game enjoyment. Since our studies had a dif-ferent emphasis, i.e., the general usefulness of animal avatars,we cannot disentangle these relations in detail. However, wesee our results as evidence for the importance of IVBO for VRgames in general, be it human or animal avatars. Hence, wepropose an in-depth investigation of that overarching topic aspossible follow-up research. Ultimately, we assume that a fur-ther exploration will encourage researchers and practitionersto consider IVBO as a helpful tool that allows the creation ofnovel, engaging player experiences that cannot be realized innon-VR games.
EFERENCES [1] Sun Joo Ahn, Joshua Bostick, Elise Ogle, Kristine LNowak, Kara T McGillicuddy, and Jeremy N Bailenson.2016. Experiencing nature: Embodying animals inimmersive virtual environments increases inclusion ofnature in self and involvement with nature.
Journal ofComputer-Mediated Communication
21, 6 (2016),399–419.[2] Paul Bach-y Rita and Stephen W Kercel. 2003. Sensorysubstitution and the human–machine interface.
Trends incognitive sciences
7, 12 (2003), 541–546.[3] Domna Banakou, Raphaela Groten, and Mel Slater.2013. Illusory ownership of a virtual child body causesoverestimation of object sizes and implicit attitudechanges.
Proceedings of the National Academy ofSciences
Environmentand Behavior
39, 2 (2007), 269–283.[5] Frank Biocca and Ben Delaney. 1995. Communicationin the Age of Virtual Reality. L. Erlbaum AssociatesInc., Hillsdale, NJ, USA, Chapter Immersive VirtualReality Technology, 57–124. http://dl.acm.org/citation.cfm?id=207922.207926 [6] Kristopher J Blom, Jorge Arroyo-Palacios, and MelSlater. 2014. The effects of rotating the self out of thebody in the full virtual body ownership illusion.
Perception
43, 4 (2014), 275–294.[7] Matthew Botvinick and Jonathan Cohen. 1998. Rubberhands â ˘AŸfeelâ ˘A ´Ztouch that eyes see.
Nature
IEEE Journal of Robotics andAutomation
17, 1-19 (2004), 16.[9] Paul Cairns, Anna Cox, and A Imran Nordin. 2014.Immersion in digital games: review of gamingexperience research.
Handbook of digital games (2014),337–361.[10] HTC Corporation. 2018. HTC Vive. Website. (2018).Retrieved October 12, 2018 from .[11] HTC Corporation. 2019. HTC Vive Tracker. Website.(2019). Retrieved March 10, 2019 from .[12] Mie C. S. Egeberg, Stine L. R. Lind, Sule Serubugo,Denisa Skantarova, and Martin Kraus. 2016. Extendingthe Human Body in Virtual Reality: Effect of SensoryFeedback on Agency and Ownership of Virtual Wings.In
Proceedings of the 2016 Virtual Reality InternationalConference (VRIC ’16) . ACM, New York, NY, USA,Article 30, 4 pages.
DOI: http://dx.doi.org/10.1145/2927929.2927940 [13] H Henrik Ehrsson. 2007. The experimental induction ofout-of-body experiences.
Science
Perception
38, 2 (2009), 310–312.[15] Arvid Guterstam, Valeria I Petkova, and H HenrikEhrsson. 2011. The illusion of owning a third arm.
PloSone
6, 2 (2011), e17208.[16] Carrie Heeter. 1992. Being there: The subjectiveexperience of presence.
Presence: Teleoperators &Virtual Environments
1, 2 (1992), 262–271.[17] Wijnand A. IJsselsteijn, Huib de Ridder, JonathanFreeman, and Steve E. Avons. 2000. Presence: concept,determinants, and measurement. (2000).
DOI: http://dx.doi.org/10.1117/12.387188 [18] Dongsik Jo, Kangsoo Kim, Gregory F Welch, WoojinJeon, Yongwan Kim, Ki-Hong Kim, andGerard Jounghyun Kim. 2017. The impact ofavatar-owner visual similarity on body ownership inimmersive virtual reality. In
Proceedings of the 23rdACM Symposium on Virtual Reality Software andTechnology . ACM, 77.[19] Daniel Johnson and John Gardner. 2010. Personality,motivation and video games. In
Proceedings of the 22ndConference of the Computer-Human Interaction SpecialInterest Group of Australia on Computer-HumanInteraction - OZCHI ’10 , Margot Brereton, StephenViller, and Ben Kraal (Eds.). ACM Press, New York,New York, USA, 276–279.
DOI: http://dx.doi.org/10.1145/1952222.1952281 [20] Joohee Jun, Myeongul Jung, So-Yeon Kim, andKwanguk Kenny Kim. 2018. Full-Body OwnershipIllusion Can Change Our Emotion. In
Proceedings of the2018 CHI Conference on Human Factors in ComputingSystems . ACM, 601.[21] Robert S Kennedy, Norman E Lane, Kevin S Berbaum,and Michael G Lilienthal. 1993. Simulator sicknessquestionnaire: An enhanced method for quantifyingsimulator sickness.
The international journal of aviationpsychology
3, 3 (1993), 203–220.[22] Konstantina Kilteni, Ilias Bergstrom, and Mel Slater.2013. Drumming in immersive virtual reality: the bodyshapes the way we play.
IEEE Transactions onVisualization & Computer Graphics
Presence: Teleoperators and Virtual Environments
21, 4(2012), 373–387.[24] Konstantina Kilteni, Jean-Marie Normand, Maria VSanchez-Vives, and Mel Slater. 2012b. Extending bodyspace in immersive virtual reality: a very long armillusion.
PloS one
7, 7 (2012), e40867.25] Martijn JL Kors, Gabriele Ferri, Erik D Van Der Spek,Cas Ketel, and Ben AM Schouten. 2016. A breathtakingjourney. On the design of an empathy-arousingmixed-reality game. In
Proceedings of the 2016 AnnualSymposium on Computer-Human Interaction in Play .ACM, 91–104.[26] Andrey Krekhov, Sebastian Cmentowski, KatharinaEmmerich, Maic Masuch, and Jens Krüger. 2018b.GulliVR: A Walking-Oriented Technique for Navigationin Virtual Reality Games Based on Virtual BodyResizing. In
Proceedings of the 2018 Annual Symposiumon Computer-Human Interaction in Play (CHI PLAY’18) . ACM, New York, NY, USA, 243–256.
DOI: http://dx.doi.org/10.1145/3242671.3242704 [27] Andrey Krekhov, Sebastian Cmentowski, and JensKrüger. 2018a. VR Animals: Surreal Body Ownershipin Virtual Reality Games. In
Extended AbstractsPublication of the Annual Symposium onComputer-Human Interaction in Play . ACM, to appear.[28] Joseph J LaViola Jr. 2000. A discussion ofcybersickness in virtual environments.
ACM SIGCHIBulletin
32, 1 (2000), 47–56.[29] Luís Leite and Veronica Orvalho. 2012. Shape yourbody: control a virtual silhouette using body motion. In
CHI’12 Extended Abstracts on Human Factors inComputing Systems . ACM, 1913–1918.[30] Bigna Lenggenhager, Tej Tadi, Thomas Metzinger, andOlaf Blanke. 2007. Video ergo sum: manipulatingbodily self-consciousness.
Science
Proceedings of the ACM Symposium on AppliedPerception . ACM, 69–76.[32] Lionhead Studios. 2001.
Black & White . Game[Windows]. (30 March 2001). Electronic Arts, RedwoodCity, California, U.S. Last played August 2003.[33] Donna M Lloyd. 2007. Spatial limits on referred touchto an alien limb may reflect boundaries of visuo-tactileperipersonal space surrounding the hand.
Brain andcognition
64, 1 (2007), 104–109.[34] Matthew Lombard and Theresa Ditton. 1997. At theheart of it all: The concept of presence.
Journal ofComputer-Mediated Communication
3, 2 (1997), 0–0.[35] J-L Lugrin, Johanna Latt, and Marc Erich Latoschik.2015. Anthropomorphism and illusion of virtual bodyownership. In
Proceedings of the 25th InternationalConference on Artificial Reality and Telexistence and20th Eurographics Symposium on Virtual Environments .Eurographics Association, 1–8.[36] Jean-Luc Lugrin, Ivan Polyschev, Daniel Roth, andMarc Erich Latoschik. 2016. Avatar anthropomorphism and acrophobia. In
Proceedings of the 22nd ACMConference on Virtual Reality Software and Technology .ACM, 315–316.[37] Antonella Maselli and Mel Slater. 2013. The buildingblocks of the full body ownership illusion.
Frontiers inhuman neuroscience
Extended Abstracts Publicationof the Annual Symposium on Computer-HumanInteraction in Play . ACM, 599–605.[39] Thomas Nagel. 1974. What is it like to be a bat?
Thephilosophical review
83, 4 (1974), 435–450.[40] Jean-Marie Normand, Elias Giannopoulos, BernhardSpanlang, and Mel Slater. 2011. Multisensorystimulation can induce an illusion of larger belly size inimmersive virtual reality.
PloS one
6, 1 (2011), e16128.[41] Tabitha C Peck, Sofia Seinfeld, Salvatore M Aglioti, andMel Slater. 2013. Putting yourself in the skin of a blackavatar reduces implicit racial bias.
Consciousness andcognition
22, 3 (2013), 779–787.[42] Daniel Perez-Marcos, Maria V Sanchez-Vives, and MelSlater. 2012. Is my hand connected to my body? Theimpact of body continuity and arm alignment on thevirtual hand illusion.
Cognitive neurodynamics
6, 4(2012), 295–305.[43] Valeria I Petkova and H Henrik Ehrsson. 2008. If I wereyou: perceptual illusion of body swapping.
PloS one
Deadly Creatures . Game[Nintendo Wii]. (13 February 2009). THQ, Agoura Hills,California, U.S. Last played January 2010.[45] Helge Rhodin, James Tompkin, Kwang In Kim, KiranVaranasi, Hans-Peter Seidel, and Christian Theobalt.2014. Interactive motion mapping for real-time charactercontrol. In
Computer Graphics Forum , Vol. 33. WileyOnline Library, 273–282.[46] Helge Rhodin, James Tompkin, Kwang In Kim, EdilsonDe Aguiar, Hanspeter Pfister, Hans-Peter Seidel, andChristian Theobalt. 2015. Generalizing wave gesturesfrom sparse examples for real-time character control.
ACM Transactions on Graphics (TOG)
34, 6 (2015),181.[47] C. Scott Rigby and Richard M. Ryan. 2007. The PlayerExperience of Need Satisfaction (PENS): An appliedmodel and methodology for understanding keycomponents of the player experience. (2007).[48] Giuseppe Riva, John Waterworth, and Dianne Murray.2014.
Interacting with Presence: HCI and the Sense ofPresence in Computer-mediated Environments . Walterde Gruyter GmbH & Co KG.49] Daniel Roth, Jean-Luc Lugrin, Marc Erich Latoschik,and Stephan Huber. 2017. Alpha IVBO-construction ofa scale to measure the illusion of virtual body ownership.In
Proceedings of the 2017 CHI Conference ExtendedAbstracts on Human Factors in Computing Systems .ACM, 2875–2883.[50] Roy A Ruddle and Simon Lessels. 2009. The benefits ofusing a walking interface to navigate virtualenvironments.
ACM Transactions on Computer-HumanInteraction (TOCHI)
16, 1 (2009), 5.[51] Richard M Ryan and Edward L Deci. 2000.Self-determination theory and the facilitation of intrinsicmotivation, social development, and well-being.
American psychologist
55, 1 (2000), 68.[52] Richard M. Ryan, C. Scott Rigby, and AndrewPrzybylski. 2006. The Motivational Pull of VideoGames: A Self-Determination Theory Approach.
Motivation and Emotion
30, 4 (2006), 344–360.[53] Maria V Sanchez-Vives, Bernhard Spanlang, AntonioFrisoli, Massimo Bergamasco, and Mel Slater. 2010.Virtual hand illusion induced by visuomotor correlations.
PloS one
5, 4 (2010), e10381.[54] Thomas Schubert, Holger Regenbrecht, and FrankFriedmann. 2018. Igroup Presence Questionnaire (IPQ).(2018). [55] Thomas W. Schubert. 2003. The sense of presence invirtual environments: A three-component scalemeasuring spatial presence, involvement, and realness.
Zeitschrift für Medienpsychologie
15, 2 (2003), 69–71.[56] William R Sherman and Alan B Craig. 2002.
Understanding virtual reality: Interface, application,and design . Elsevier.[57] Erik Sikström, Amalia De Götzen, and Stefania Serafin.2014. The role of sound in the sensation of ownership ofa pair of virtual wings in immersive VR. In
Proceedingsof the 9th Audio Mostly: A Conference on InteractionWith Sound . ACM, 24.[58] Mel Slater. 2003. A note on presence terminology.
Presence connect
3, 3 (2003), 1–5.[59] Mel Slater, Daniel Pérez Marcos, Henrik Ehrsson, andMaria V Sanchez-Vives. 2008. Towards a digital body:the virtual arm illusion.
Frontiers in humanneuroscience
Frontiers in neuroscience
PloS one
5, 5 (2010), e10564.[62] Mel Slater, Martin Usoh, and Anthony Steed. 1995.Taking steps: the influence of a walking technique onpresence in virtual reality.
ACM Transactions onComputer-Human Interaction (TOCHI)
2, 3 (1995),201–219. [63] William Steptoe, Anthony Steed, and Mel Slater. 2013.Human tails: ownership and control of extendedhumanoid avatars.
IEEE Transactions on Visualization& Computer Graphics
Anthrozoös
18, 1 (2005), 18–27.[65] Unity Technologies. 2018. Unity. Website. (2018).Retrieved December 12, 2018 from https://unity3d.com/ .[66] Manos Tsakiris. 2010. My body in the brain: aneurocognitive model of body-ownership.
Neuropsychologia
48, 3 (2010), 703–712.[67] Manos Tsakiris and Patrick Haggard. 2005. The rubberhand illusion revisited: visuotactile integration andself-attribution.
Journal of Experimental Psychology:Human Perception and Performance
31, 1 (2005), 80.[68] Ubisoft Montreal. 2016.
Eagle Flight . Game [VR]. (18October 2016). Ubisoft, Montreuil, France. Last playedJanuary 2019.[69] Thomas Waltemate, Dominik Gall, Daniel Roth, MarioBotsch, and Marc Erich Latoschik. 2018. The Impact ofAvatar Personalization and Immersion on Virtual BodyOwnership, Presence, and Emotional Response.
IEEEtransactions on visualization and computer graphics
Game Based Learning
55 (2015).[71] Bob G. Witmer, Christian J. Jerome, and Michael J.Singer. 2005. The Factor Structure of the PresenceQuestionnaire.
Presence: Teleoperators and VirtualEnvironments
14, 3 (2005), 298–312.
DOI: http://dx.doi.org/10.1162/105474605323384654 [72] Bob G Witmer and Michael J Singer. 1998a. Measuringpresence in virtual environments: A presencequestionnaire.
Presence
7, 3 (1998), 225–240.[73] Bob G. Witmer and Michael J. Singer. 1998b.Measuring Presence in Virtual Environments: APresence Questionnaire.
Presence: Teleoperators andVirtual Environments
7, 3 (1998), 225–240.
DOI: http://dx.doi.org/10.1162/105474698565686 [74] Andrea Stevenson Won, Jeremy N Bailenson, and JaronLanier. 2015. Homuncular flexibility: the human abilityto inhabit nonhuman avatars.
Emerging Trends in theSocial and Behavioral Sciences: An Interdisciplinary,Searchable, and Linkable Resource (2015), 1–16.[75] W. Xu, A. Chatterjee, M. Zollhofer, H. Rhodin, P. Fua,H. Seidel, and C. Theobalt. 2019. Mo Cap : Real-timeMobile 3D Motion Capture with a Cap-mounted FisheyeCamera. IEEE Transactions on Visualization andComputer Graphics (2019), 1–1.
DOI: http://dx.doi.org/10.1109/TVCG.2019.2898650 [76] Nick Yee and Jeremy Bailenson. 2007. The Proteuseffect: The effect of transformed self-representation onbehavior.