The impact of human agents on spatial navigation and knowledge acquisition in a virtual environment

The impac o human agen s on spa ial na iga ion and knowledge acquisi ion in a i ual en i onmen T acy Sánchez Pacheco 1 *, Melissa Sa ia Mosque a 1 , Kaya Gä ne 1 , Vincen Schmid 1 , Debo a Nol e 1 , Sabine U. König 1† , Go don Pipa 1† and Pe e König 1 , 2† 1 Ins i u e o Cogni i e Science, Uni e si y o Osnab ück, Osnab ück, Ge many, 2 Depa men o Neu ophysiology and Pa hophysiology, Uni e si y Medical Cen e Hambu g-Eppendo , Hambu g, Ge many Concep s o spa ial na iga ion es on he idea o landma ks, which a e immobile ea u es o objec s in he en i onmen . Howe e , beha io ally ele an objec s o ellow humans a e o en mobile. This aises he ques ion o how he p esence o human agen s influences spa ial explo a ion and knowledge acquisi ion. He e, we in es iga e explo a ion and pe o mance in subsequen spa ial asks wi hin a i ual en i onmen con aining nume ous human a a a s. In he explo a ion phase, agen s had a locally limi ed e ec on na iga ion. They p omp ed pa icipan s o e isi loca ions wi h agen s du ing hei ini ial explo a ion wi hou significan ly al e ing o e all explo a ion pa e ns o he ex en o he a ea co e ed. Howe e , agen s and buildings compe ed o isual a en ion. When spa ial ecall was es ed, poin ing accu acy owa d buildings imp o ed when pa icipan s di ec ed hei a en ion o he buildings and nea by agen s. In con as , poin ing accu acy o agen s showed weake pe o mance and did no benefi om isual a en ion di ec ed owa d he adjacen building. Con ex ual agen s and incong uen agen -en i onmen pai ings u he enhanced poin ing accu acy, e ealing ha iola ions o expec a ions by agen s can significan ly shape na iga ional knowledge acquisi ion. O e all, agen s influenced spa ial explo a ion by di ec ing a en ion locally, wi h he in e ac ion be ween agen salience and en i onmen al ea u es playing a key ole in shaping na iga ional knowledge acquisi ion. KEYWORDS spa ial na iga ion, human agen s, i ual eali y, explo a ion-exploi a ion, social acili a ion 1 In oduc ion Spa ial na iga ion is essen ial o goal-o ien ed mo emen and ac i e en i onmen al in e ac ion (Eps ein e al., 2017;I o e al., 2015). In humans, egula engagemen in spa ial na iga ion, whe he s udied h ough na iga ion done in he con ex o p o essional ac i i ies (G iesbaue e al., 2022;Magui e e al., 2006;Woolle and Magui e, 2011), a ge ed aining (Choi e al., 2012), o i ual en i onmen s (Wes e al., 2017), is ela ed o he enhancemen o cogni i e unc ions, pa icula ly memo y and spa ial awa eness. This OPEN ACCESS EDITED BY Ma ia Pyasik, Uni e si y o Udine, I aly REVIEWED BY Ad iana Sala ino, Royal Mili a y Academy, Belgium Ru h Con oy Dal on, Lancas e Uni e si y, Uni ed Kingdom *CORRESPONDENCE T acy Sánchez Pacheco, acy.sanchez.pacheco@uni-osnab ueck.de † These au ho s con ibu ed equally and sha e senio au ho ship RECEIVED 16 Sep embe 2024 ACCEPTED 06 Ma ch 2025 PUBLISHED 02 Ap il 2025 CITATION Sánchez Pacheco T, Sa ia Mosque a M, Gä ne K, Schmid V, Nol e D, König SU, Pipa G and König P (2025) The impac o human agen s on spa ial na iga ion and knowledge acquisi ion in a i ual en i onmen . F on . Vi ual Real. 6:1497237. doi: 10.3389/ i .2025.1497237 COPYRIGHT © 2025 Sánchez Pacheco, Sa ia Mosque a, Gä ne , Schmid , Nol e, König, Pipa and König. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (CC BY). The use, dis ibu ion o ep oduc ion in o he o ums is pe mi ed, p o ided he o iginal au ho (s) and he copy igh owne (s) a e c edi ed and ha he o iginal publica ion in his jou nal is ci ed, in acco dance wi h accep ed academic p ac ice. No use, dis ibu ion o ep oduc ion is pe mi ed which does no comply wi h hese e ms. F on ie s in Vi ual Reali y on ie sin.o g01 TYPE O iginal Resea ch PUBLISHED 02 Ap il 2025 DOI 10.3389/ i .2025.1497237 in ica e link be ween spa ial na iga ion and cogni i e p ocesses highligh s he c ucial ole o spa ial na iga ion in de eloping and o ganizing spa ial knowledge. The ans o ma ion o na iga ional expe iences in o spa ial knowledge begins wi h ecognizing key en i onmen al elemen s, ollowed by hei g adual in eg a ion in o a cohesi e e e ence sys em (Eks om and Isham, 2017). When indi iduals explo e new su oundings eely, hey can iden i y elemen s ha aid hei o ien a ion and ha ness hem as building blocks o knowledge cons uc ion, such as emembe ing shops when isi ing a new own. This p ocess, known as landma k iden ifica ion, is essen ial o main aining posi ional awa eness and planning u u e pa hways (Janzen e al., 2006). Spa ial knowledge encompasses ecognizing and ecalling c i ical elemen s o he landscape, unde s anding hei spa ial ela ionships, and he ou es connec ing hem. Despi e ex ensi e esea ch on he ole o s a ic landma ks in spa ial cogni ion (Malanchini e al., 2020), he dynamic human aspec s influencing spa ial explo a ion and knowledge acquisi ion ha e ye o be ully explo ed. O en, humans a e iewed p ima ily as modifie s o na iga ional pa hs a he han as ac i e con ibu o s o spa ial knowledge acquisi ion (Bicanski and Bu gess, 2020;Eks om and Isham, 2017). This pe spec i e o e looks he significan ole o he indi iduals play in eal-wo ld spa ial cogni ion. Encoun e s wi h o he s can impac pedes ian dynamics (Dal on e al., 2019), influence isual explo a ion (Ge e al., 2020), p o ide i al in o ma ion abou he sa e y and usabili y o spaces (Bajo unai e e al., 2022), a ec he ecall o loca ions (Kuehn e al., 2018), and p omp a pa allel social mapping o he en i onmen (Scha e and Schille , 2018). Thus, inco po a ing o he humans in o na iga ion esea ch is essen ial o a deepe unde s anding o spa ial cogni ion. Howe e , s udying he influence o ellow humans on conc e e spa ial knowledge aces he di ficul ies o main aining a na iga ion scena io o eal-wo ld scale in con olled en i onmen s. When a ailable, esea che s o en conduc e ospec i e s udies using eal-wo ld da a, such as mobile phone and GPS da a, which can cha ac e ize human mobili y pa e ns bu lack he con olled a iables necessa y o isola ing specific ac o s media ing he esul s (Pappala do e al., 2015;Schläp e e al., 2021). Mo eo e , hese s udies do no link explo a ion pa e ns o he u u e acquisi ion o spa ial knowledge, hinde ing a comp ehensi e unde s anding o cogni i e p ocesses de i ed om na iga ion pa e ns. When mo ing o he o he side o he spec um, labo a o y se ups end o be bound o small-scale en i onmen s and simplified asks ha lack he challenge o spa ial scales o he eal wo ld (Wiene e al., 2020). Pa icipan s a e asked o use humans as ancho s inside a maze-like i ual eali y (VR) o as landma ks in isual flow expe imen s (Kuehn e al., 2018). Dal on e al. (2019) poin ou ha VR echnologies in wayfinding esea ch o en exclude he p esence o o he s, he eby unde es ima ing hei impo ance. They sugges ha he me e p esence o o he s (i.e., weak social cues), e en wi hou ac i e in e ac ion, can help in e he impo ance o a space wi hin i s en i onmen , and ac i ely call o mo e s udies o be done in his ealm o esea ch. Hence, i is acknowledged ha examining how he me e p esence o o he s migh impac na iga ion and knowledge acquisi ion when humans a e no p ima y a ge s is essen ial bu emains uns udied. Wi h his wo k, we aise he ques ion o whe he spa ially ele an in o ma ion can be ex ac ed om obse ing o he humans in he space a ound us. I is impo an o no e ha humans do no inhe en ly se e as fixed landma ks in na u al na iga ion, p ima ily due o hei mobili y, p e en ing hem om o ming in insic associa ions wi h specific places. Howe e , a he han conside ing o he humans as eliable e e ence poin s, we explo e he no ion ha hei ele ance wi hin a gi en con ex can be ha nessed as a sou ce o in o ma ion o enhance spa ial knowledge. Essen ially, i is no he p esence o o he s pe se ha aids in spa ial cogni ion bu a he how hey in e ac wi h he elemen s in hei su oundings ha could be ele an o spa ial explo a ion and knowledge acquisi ion. This p omp s us o in es iga e wha minimal change in a human agen ’s in e ac ion wi h he en i onmen can elici pa icipan s’di e en beha io al esponses. We seek o unde s and whe he hese human elemen s ac as dis ac ions, po en ially diminishing he saliency o he su ounding s imuli, o i , con e sely, hey con ibu e o imp o ed pe o mance by ancho ing indi iduals o a mo e i id men al ep esen a ion o specific loca ions. To add ess his complexi y, we in es iga ed how spa ial knowledge acquisi ion de eloped in a con olled one-squa e-kilome e VR en i onmen wi h human agen s. The s udy inco po a ed human agen s a wo le els: one in which he agen in e ac ed wi h he en i onmen by holding an objec ele an o he con ex , such as a oolbox in on o a ha dwa e s o e (i.e., Con ex ual agen ), and ano he in which he agen simply s ood wi hou in e ac ing wi h any objec s a ound i (i.e., Acon ex ual agen ). These agen s we e placed in on o public buildings, such as s o es, baske ball cou s, es au an s, o esiden ial buildings. Fo ou fi s g oup o pa icipan s, all con ex ual agen s we e placed in on o public buildings ha ma ched hei objec in e ac ion. Fo he second g oup, we dis up ed his cong uency o s udy he sole influence o he ype o agen and hei building on he con ex in which hey a e si ua ed, unde he hypo hesis ha ha ing con ex -cong uen agen s will enhance he pa icipan ’s abili y o ecall. Specifically, we p opose h ee hypo heses: 1) he Visi Hypo hesis, which s a es ha con ex ual agen s will influence pa icipan s’explo a ion pa e ns by d awing hem back o p e iously isi ed loca ions; 2) he Dwell Hypo hesis, which posi s ha pa icipan s will alloca e mo e isual a en ion o con ex ually cong uen agen s compa ed o acon ex ual ones; and 3) he Pe o mance Hypo hesis, which sugges s ha his inc eased engagemen will lead o imp o ed pe o mance in spa ial knowledge asks, such as poin ing accu acy. The aim o ou s udy is o explo e he ole o hese human agen s and hei cong uency wi h con ex in spa ial explo a ion and he de elopmen o spa ial knowledge. 2 Resul s We examined he impac o human agen s on spa ial na iga ion and knowledge acquisi ion in a i ual ci y named Wes b ook, consis ing o 236 buildings. We iden ified 26 public buildings (e.g., shops, baske ball cou ) and 26 esiden ial buildings as ask- ele an , ma king hem wi h s ee a . Addi ionally, he e we e 180 buildings wi hou g a fi i and ou la ge buildings on he ci y’s ou ski s, which could se e as global landma ks gi en hei dimensions. We designed wo ca ego ies o human agen s: con ex ual agen s, who pe o med con ex - ele an ac ions (e.g., F on ie s in Vi ual Reali y on ie sin.o g02 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 holding a oolbox in on o a ha dwa e s o e), and acon ex ual agen s, who held a es ing posi ion wi hou in e ac ing wi h objec s. In he fi s expe imen , con ex ual agen s we e placed in public a eas, displaying ac ions cong uen wi h he buildings, while acon ex ual agen s we e posi ioned in on o esiden ial buildings wi hou in e ac ion. In he second expe imen , bo h agen ypes we e spli e enly ac oss public and esiden ial a eas, dis up ing he cong uen pai s. Pa icipan s in bo h expe imen s comple ed fi e 30-min explo a ion sessions, o aling 150 min. Addi ionally, o p o ide a baseline o compa ison o he explo a ion s a egies, we used he con ol g oup om Schmid e al. (2023) who explo ed he same VR ci y (Wes b ook) wi h he same session leng hs and numbe s bu no agen s p esen . We in es iga ed he explo a ion phase by analyzing pa icipan s’na iga ional co e age o he ci y, hei walking s a egies, agen -induced bias in hei explo a ion, and hei isualbeha io du ingexplo a ion. Finally, we es ed hei spa ial knowledge acquisi ion in a sepa a e session using VR poin ing asks. To es ablish compa abili y in spa ial o ien a ion abili ies be ween he pa icipan s o he expe imen s (expe imen 1, expe imen 2, and con ol), pa icipan s comple ed he FRS (F agebogen Räumliche S a egien) ques ionnai e, and hei sco es we e con as ed. The e we e no significan di e ences be ween he g oups a baseline on any o he h ee subscales (global, su ey, and ca dinal), χ2(2,N67)≤1.68,p≥0.43. The e o e, he g oups we e compa able in hei assessmen o hei use o spa ial s a egies be o e he s a o he expe imen s. 2.1 Assessmen o he explo a ion phase Du ing he VR ci y expe imen , we acked pa icipan s’ explo a ion, including walking beha io , na iga ional co e age, decision-poin s a egies, and isual beha io . This FIGURE 1 F ee explo a ion o Wes b ook (a) P esen s a hea map ha isualizes he spa ial na iga ion da a o all pa icipan s ac oss he explo a ion phase. I compiles mo emen ac oss all sessions, wi h pa icipan loca ions disc e ized o he nea es cen ime e and a e aged on a second-by-second basis a e emo ing he ini ial 2 s o each eco ding. We fi ed a hea map g id app oxima ing a 1-cm esolu ion and c ea ed a colo scale anging om 0 o 10 isi s, highligh ing a eas o a ied explo a ion in ensi y (b) Shows Wes b ook om abo e, o e laid wi h a g aph s uc u e we used o o malize na iga ional da a as decision uni s (c) Quan ifies explo a ion h ough a map co e age a io, calcula ed by di iding he numbe o unique nodes isi ed by he o al nodes in he ci y (d) Compiles hese a ios on a cumula i e basis, eflec ing he pa icipan s’expanding disco e y o he ci y o e ime. F on ie s in Vi ual Reali y on ie sin.o g03 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 comp ehensi e analysis e ealed hei na iga ional s a egies and engagemen , highligh ing he agen s’impac on hei explo a ion. 2.1.1 Na iga ional co e age o he ci y We quan ified pa icipan s’walking beha io in he i ual ci y using a p imal ci y g aph (Neal, 2013) o analyze pa icipan s’ ee explo a ion pa e ns. In his g aph, decision poin s (i.e., in e sec ions o walkable pa hs) we e ep esen ed as nodes and pa hs connec ing hem as edges. Pa icipan s’na iga ional coo dina es (see Figu e 1a) we e assigned o he nea es g aph elemen (see Figu e 1b), defining hei explo a ion as mo emen s om one g aph elemen o ano he . Ou o 159 nodes, pa icipan s isi ed be ween 45 and 113 unique nodes du ing each 30-min session (M= 87.06, SD = 16.49). We calcula ed he co e age a io by di iding he numbe o nodes isi ed a leas once by he o al numbe o nodes. To accoun o epea ed measu es, we implemen ed a linea mixed-e ec s model (LMM) ha conside ed in asubjec a iabili y and gene a ed indi idual in e cep s o each pa icipan . This model p edic ed he indi idual session co e age a io as a unc ion o he session, he expe imen al g oup (con ol e sus ci y wi h agen s), and hei in e ac ion. Using he fi s session as a baseline, we obse ed significan cumula i e inc eases in na iga ional co e age wi h each subsequen session. S a ing om an indi idual co e age o oughly hal he ci y (as shown by he mean o he blue cu es in Figu es 1c, d), he LMM analysis explained a subs an ial po ion o he a iance in na iga ional co e age, wi h ma ginal R2 m0.18 ( a iance explained by fixed e ec s) and condi ional R2 c0.70 ( a iance explained by bo h fixed and andom e ec s), emphasizing he ole o session-based lea ning while accoun ing o indi idual di e ences. The coe ficien s indica e changes ela i e o session 1(η2 p.40), wi h posi i e alues signi ying an inc ease: βSession 2 0.03 (SE = 0.01, 2.48, p0.01), βSession 3 0.07 (SE = 0.01, 6.54, p<.001), βSession 4 0.09 (SE = 0.01, 8.53, p<.001), and βSession 5 0.12 (SE = 0.01, 11.49, p<.001). The e ec o he expe imen g oup (agen s e sus con ol) on he co e age a io was no s a is ically significan (βExpe imen −0.02, SE = 0.02, −0.72, p0.48). Addi ionally, in e ac ions be ween he session and expe imen g oup we e also no significan : βSession i:Expe imen ≤−0.04, p≥.07. This demons a ed ha as sessions ad anced, pa icipan s co e ed mo e g ound wi hin he same ime ame, achie ing his equally in bo h a ci y wi h agen s and one de oid o hem. In o de o es ima e i he pa icipan s we e di e en ially accumula ing unique decision poin s in he ci y as he sessions p og essed, we calcula ed a cumula i e co e age a io, accumula ing he numbe o uniquely isi ed nodes as he sessions ad anced. A LMM wi h pa icipan s as andom e ec s was used o p edic he cumula i e co e age a io as a unc ion o session, expe imen g oup, and hei in e ac ion (R2 m0.74, R2 c0.91). This analysis showed a significan p og ession, wi h pa icipan s co e ing mo e o he ci y in each subsequen session (see Figu e 1d). S a ing om he same ini ial co e age a io mean (M= 0.49), significan cumula i e inc emen s in na iga ional co e age we e obse ed in each session (η2 p.91):βSession 2 0.19 (SE = 0.01, 24.71, p<.001), βSession 3  0.27 (SE = 0.01, 35.80, p<.001), βSession 4 0.31 (SE = 0.01, 40.75, p<.001), and βSession 5 0.33 (SE = 0.01, 43.48, p<.001). The e ec o he expe imen g oup on he co e age a io was no s a is ically significan (βExpe imen −0.02, p0.48). Addi ionally, in e ac ions be ween he session and expe imen g oup we e also no significan : βSession i:Expe imen ≤0.02, p≥.33. A e he fi h session, he end o he explo a ion, he cumula i e numbe o unique nodes isi ed (M= 0.83, SD = 0.04) indica ed ha mos pa icipan s had seen he majo i y o he ci y. These findings sugges ha he p esence o an agen did no significan ly influence he na iga ional co e age o he ci y. 2.1.2 Explo a ion s a egies on decision poin s: Explo a o y s. conse a i e beha io To examine pa icipan s’walking s a egies, we analyzed whe he hey ended o choose pa hs hey had p e iously isi ed mo e equen ly (conse a i e) o i hey a o ed less- a eled ou es (explo a i e). We defined disc e e na iga ional decisions as mo emen s om one node o ano he and quan ified hem using a s a egy ma ix (see Figu es 2a,b). A linea mixed-e ec s model was fi ed o he da a o examine di e ences in decision numbe s based on session, s a egy (conse a i e s explo a i e), and expe imen (con ol s ci y wi h agen s), wi h andom in e cep s o pa icipan s o accoun o he nes ed da a s uc u e. The model’s fixed e ec s (R2 m0.49, R2 c0.83) indica ed ha he a e age numbe o decisions ac oss all ac o s was β073.14 (SE = 4.02, 18.20, p<.001). The analysis showed significan inc eases in decisions as pa icipan s gained expe ience in Wes b ook, wi h inc eases e iden om he fi s session onwa d (η2 p.69):βSession 2  30.17 (SE = 2.49, 12.11, p<.001), βSession 3 51.23 (SE = 2.49, 20.57, p<.001), βSession 4 67.49 (SE = 2.49, 27.09, p<.001), βSession 5 80.32 (SE = 2.49, 32.24, p<.001). The di e ence in decisions be ween conse a i e and explo a o y s a egies was significan , β75.52 (SE = 3.52, 21.44, p<.001, η2 p.14). The in e ac ion be ween session and s a egy was significan (η2 p.46), showing ha decision inc eases pe session we e lowe o he explo a o y s a egy compa ed o he conse a i e one: βSession 2:S a egy1 −44.70 (SE = 4.98, −8.97, p<.001), βSession 3:S a egy1 −67.66 (SE = 4.98, −13.58, p<.001), βSession 4:S a egy1 −86.02 (SE = 4.98, −17.26, p<.001), βSession 5:S a egy1 −102.37 (SE = 4.98, −20.55, p<.001). No significan di e ence in decision numbe s was obse ed be ween he wo expe imen s (βExpe imen −3.95, SE = 7.40, −0.53, p0.84), indica ing ha he p esence o absence o agen s did no ma kedly influence decision-making s a egies. Bo h beha io s inc eased as sessions p og essed. Ini ially, pa icipan s p io i ized explo a o y beha io , bu as hey gained expe ience, hey in eg a ed conse a i e beha io , e ec i ely combining bo h s a egies (see Figu e 2c), ega dless o agen p esence. This s a egy adap a ion, whe e explo a o y decisions emained s able while conse a i e decisions inc eased, occu ed wi hou significan influence om agen s, sugges ing ha he global explo a ion s a egy is una ec ed by human agen s. 2.1.3 Agen -induced bias on walking s a egies In o de o es ou Visi Hypo hesis, we e alua ed he impac o agen s on pa icipan s’explo a o y beha io ; we analyzed decision poin s whe e pa icipan s could choose be ween a pa h wi h an agen and one wi hou , assuming ha agen s migh elici isi s. Da a om hese poin s we e compa ed wi h a con ol g oup om Schmid e al. (2023), who explo ed he same VR ci y (Wes b ook) wi hou agen s F on ie s in Vi ual Reali y on ie sin.o g04 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 (see Figu e 3a). We used iden ical decision poin s in bo h scena ios. We applied a linea mixed-e ec s model o assess isi coun s, accoun ing o session and expe imen ype (con ol s ci y wi h agen s) wi h andom in e cep s o pa icipan s (R2 m0.41, R2 c0.82). Resul s indica ed a significan inc ease in isi coun s ac oss sessions (η2 p.71)compa ed o session 1. Specifically, isi coun s in session 2 we e, on a e age, 14.84 poin s highe (βSession 2 14.84, p<0.001), and his e ec con inued o g ow in subsequen sessions, culmina ing in a 38.55 poin inc ease by session 5(βSession 5 38.55, p<0.001). Howe e , he o e all di e ence be ween he expe imen ypes was no significan (βExpe imen −2.15, p0.50), sugges ing ha he p esence o agen s did no uni e sally a ec isi coun s ac oss all sessions. No ably, he in e ac ion be ween session and expe imen ype was significan in session 5 (βSession 5:Expe imen 5.47, p0.008, η2 p.01), showing a g ea e inc ease in isi s in he ci y wi h agen s g oup. To u he explo e he influence o agen s, we calcula ed he likelihood o pa icipan s adop ing explo a o y e sus conse a i e s a egies. This was done by di iding he numbe o choices o each s a egy cell (i.e., abo e and below he diagonal; see Figu e 3b) by he o al numbe o decisions made, as eco ded in he mi o ed cells o he s a egy ma ix. Fo ins ance, we summed he numbe o imes a pa icipan chose o mo e o a place hey had isi ed only once o e a place hey had no isi ed (cell [1,0], conse a i e beha io ) wi h he numbe o imes hey chose o go o a place hey had ne e been o e a place hey had isi ed once (cell [0,1], explo a i e beha io ), and di ided each coun by ha o al o decisions made in bo h cells (sum o he coun in bo h [1,0] and [0,1] cells). This indica o showed a clea dis inc ion be ween he beha io o pa icipan s in he ci y wi h agen s and ha o he con ol g oup. Wi hin he fi s explo a ion session, pa icipan s in he ci y wi h agen s had an a e age p opo ion o conse a i e beha io o M= 0.47 (SD = 0.16) compa ed o M= 0.20 (SD = 0.09) in he con ol da a, while he explo a i e beha io was M= 0.53 (SD = 0.16) o he ci y wi h agen s and M= 0.80 (SD = 0.09) o he con ol da a. This indica es ha agen s p omp mo e local conse a i e beha io , educing explo a o y ac ions du ing pa icipan s’ini ial ci y exposu e. These findings p o ide pa ial suppo o he Visi Hypo hesis, sugges ing ha while agen s influence local e u n pa e ns, hei e ec on o e all isi coun s is session-dependen a he han uni e sal. 2.1.4 Assessmen o isual beha io du ing explo a ion: In es iga ing dwell ime on agen s and buildings Acco ding o ou Dwell Hypo hesis, we expec ed pa icipan s o alloca e mo e isual a en ion o con ex ually meaning ul agen s wi hin he en i onmen . To cha ac e ize wha pa icipan s ocused on in he ci y, we quan ified hei isual beha io by summing he cumula i e ime spen gazing a each objec , e med dwell ime. We hypo hesized ha pa icipan s would ha e highe dwell imes o con ex ual agen s and public buildings compa ed o acon ex ual agen s and esiden ial buildings, assuming con ex ual agen s cong uen wi h hei su oundings would a ac he mos a en ion. The da a e ealed dis inc pa e ns in he a en ion pa icipan s alloca ed o di e en ypes o objec s in he ci y. No ably, gene al esiden ial houses ac oss he ci y we e obse ed o a sho e du a ion compa ed o ou expe imen al buildings, wi h FIGURE 2 Explo a ion s a egy quan ifica ion (a) This panel illus a es he ansla ion o con inuous mo emen in o quan ifiable beha io al uni s. Mo emen coo dina es, deno ed by beige do s, a e assigned o he closes node o edge cen oid. Nodes, ep esen ed by blue ci cles, a e posi ioned a decision poin s and a e in e connec ed by black lines (edges). A conse a i e decision occu s when pa icipan s e u n o a equen ly isi ed node, whe eas an explo a i e decision is made when a less equen ed node is chosen (b) This panel in oduces a s a egy ma ix designed o analyze na iga ional choices. The ma ix e alua es he pa h chosen by a pa icipan (displayed on he y-axis) agains all o he possible pa hs a a decision poin (x-axis). Fo ins ance, upon exi ing node 157 (as seen on Panel A), he pa icipan ’s choice o node 156 (wi h wo p e ious isi s) is compa ed agains nodes 155 and 158, which had 5 and 0 isi s, espec i ely. The ma ix is s uc u ed such ha ows indica e isi s o he selec ed node (in his case, 156) and columns o he unselec ed nodes (in his case, 155 and 158). En ies in he ma ix a posi ions [2,0] and [2,5] inc emen by one, eflec ing he selec ion o node 156 o e al e na i es. This ma ix isually encodes decision-making pa e ns: coun s abo e he diagonal sugges conse a i e decisions, and hose below indica e explo a o y ac ions (c) This panel syn hesizes decision-making ends by displaying he a e age sum o decisions, ca ego ized by s a egy, by all pa icipan s ac oss ou wo expe imen s and he con ol g oup in a single session. F on ie s in Vi ual Reali y on ie sin.o g05 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 pa icipan s spending on a e age app oxima ely 6.68 s (M= 6.68, SD = 6.69) gazing a hese buildings. This is in con as o he longe iewing imes o ou public ask buildings (M= 11.82, SD = 8.48) and esiden ial ask buildings (M= 11.66, SD = 8.29), which a ac ed mo e sus ained a en ion om pa icipan s. Fu he mo e, global landma ks wi hin he ci y ga ne ed significan ly longe dwell imes, wi h pa icipan s spending an a e age o 16.27 s (M= 16.27, SD = 10.69) ocusing on hese p ominen ea u es, nea ly h ee imes he a e age dwell ime o gene al buildings. In e ms o agen s, con ex ual agen s we e obse ed o an a e age o 3.63 s (M= 3.63, SD = 3.74), while acon ex ual agen s a ac ed sligh ly less a en ion, wi h an a e age dwell ime o 2.66 s (M= 2.66, SD = 2.78). These findings indica e ha ou expe imen al manipula ions e ec i ely cap u ed pa icipan s’gaze in he expec ed o de o bo h buildings and agen s. We u he analyzed how each expe imen al ac o influenced isual a en ion. On a e age, pa icipan s spen mo e ime looking a con ex ual agen s (M= 3.39, SD = 3.49) compa ed o acon ex ual agen s (M= 2.73, SD = 2.91, see Figu e 4a). Con ex ual agen s also seemed o dis ac pa icipan s om ocusing on he a ea behind hem, as buildings wi h con ex ual agen s had lowe dwell imes (M= 12.42, SD = 8.13, see Figu e 4e) compa ed o hose wi h acon ex ual agen s (M= 13.28, SD = 9.27) see Figu e 4b. These FIGURE 3 Agen -induced bias in explo a ion pa e ns (a) The figu e compa es decision-making in i ual en i onmen s wi h and wi hou agen p esence. I con as s ou expe imen al da a agains a con ol om Schmid e al. (2021). Panel A shows he a e age isi coun o agen loca ions in ou expe imen (black) and he isi s o hose same poin s in he con ol (g ey) (b) The p obabili y o pa icipan s op ing o pa hs whe e agen s a e loca ed, classi ying decisions as conse a i e when going owa ds he agen was he mo e amilia pa h, and explo a o y when choosing he agen pa h was he less equen ed ou e. We only show he fi s wo sessions as he di e ence disappea s a e wa d. F on ie s in Vi ual Reali y on ie sin.o g06 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 esul s unde sco e he ad e sa ial ela ionship be ween agen s and building a en ion, whe e inc eased ocus on con ex ual agen s co esponded wi h dec eased a en ion o he buildings behind hem. Rega ding building ype, dwell ime on he agen was no significan ly a ec ed by he su ounding a ea, as agen s in public a eas (M= 3.07, SD = 3.07) and esiden ial a eas (M= 3.03, SD = 3.37, see Figu e 4c) ecei ed simila a en ion. Howe e , pa icipan s spen mo e ime gazing a esiden ial buildings (M= 13.39, SD = 9.01) compa ed o public buildings (M= 12.32, SD = 8.42, see Figu e 4 ). These esul s imply ha incong uen agen s di e ocus om su oundings, as indica ed by he an i-co ela ion in dwell imes be ween agen s and buildings. Examining cong uency (see Figu e 4d), pa icipan s looked a incong uen agen s (M= 3.23, SD = 3.48, see Figu e 4d) o longe pe iods compa ed o cong uen agen s (M= 2.60, SD = 2.44). This pa e n was opposi e o building gazing ime, wi h pa icipan s spending mo e ime looking a buildings and su oundings when he agen ma ched he con ex in which i was placed (M= 12.92, SD = 7.65) compa ed o when he agen did no ma ch he su oundings (M= 12.83, SD = 9.15, see Figu e 4g). These esul s imply ha when aced wi h incong uen agen s, pa icipan s edi ec hei ocus away om he su oundings, as e idenced by he an i-co ela ion in dwell imes be ween agen s and buildings in he cong uency ac o . To accoun o he high in e -indi idual a iabili y and he nes ed s uc u e o he da a, we employed a linea mixed-e ec s model o p edic dwell ime on agen s, wi h subjec s as a andom e ec . The fixed e ec s included he con ex ( esiden ial s public), agen ype le el (acon ex ual s con ex ual), he cong uency o he agen wi h hei su oundings (no cong uen s cong uen ), and he in e ac ion be ween agen ype and con ex . The esul s (R2 m0.03, R2 c0.36) indica ed ha pa icipan s gazed a agen s o a significan ly sho e pe iod in esiden ial con ex s (βBuilding −0.41, SE 0.08, −5.30, p<0.001, η2 p.0026). Addi ionally, pa icipan s spen mo e ime gazing a con ex ual agen s compa ed o acon ex ual agen s (βAgen 1.32, SE 0.08, 16.94, p<0.001, η2 p.03). The cong uency be ween he agen and i s con ex also had a significan e ec , wi h pa icipan s gazing a cong uen agen s o a sho e du a ion (βCong uency −0.58, SE 0.13, −4.59, p<0.001, η2 p.0034). Mo eo e , he in e ac ion be ween con ex and agen ac ion le el was significan (β−0.76, SE 0.16, −4.88, p<0.001, η2 p.0022), sugges ing ha he longes dwell imes we e obse ed o con ex ual agen s in esiden ial se ings. The findings e eal ha esiden ial con ex ual agen s cap u e he mos a en ion, especially when hey clash wi h hei su oundings. We fi ed an analogous linea mixed-e ec s model o he dwell ime on buildings (R2 m0.014, R2 c0.089). We ound ha FIGURE 4 Dwell imes ac oss sessions: This figu e p esen s he dwell ime, defined as he cumula i e sum o fixa ions on specific objec s du ing he explo a ion phase, measu ed in seconds. The ba g aph shows he in e play o wo ac o s, building and agen ca ego y (a) Illus a es he a e age dwell imes o ou dis inc building and agen ype combina ions. Each ba ep esen s a unique combina ion: public buildings wi h con ex ual agen s, public buildings wi h acon ex ual agen s, esiden ial buildings wi h con ex ual agen s, and esiden ial buildings wi h acon ex ual agen s. The heigh o each ba indica es he a e age dwell ime, eflec ing he ela i e isual a en ion each combina ion ecei ed. To he igh , we examine one ac o a a ime o dwell ime on he agen on he op (b–d), and on buildings on he bo om (e–g). F on ie s in Vi ual Reali y on ie sin.o g07 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 pa icipan s gazed a buildings o a significan ly sho e pe iod in esiden ial con ex s (β−2.02, SE 0.25, −8.19, p<0.001, η2 p.0067). Addi ionally, pa icipan s spen less ime gazing a buildings wi h con ex ual agen s compa ed o hose wi h acon ex ual agen s (β−1.44, SE 0.25, −5.83, p<0.001, η2 p.0034). The cong uency be ween he agen and he building also had a significan e ec , wi h pa icipan s gazing a buildings wi h cong uen agen s o a longe du a ion (β3.07, SE 0.39, 7.77, p<0.001, η2 p.01). Mo eo e , he in e ac ion be ween con ex and agen ype was significan (β−1.61, SE 0.49, −3.26, p0.001, η2 p.0011), indica ing ha he sho es dwell imes we e obse ed o buildings wi h con ex ual agen s in esiden ial se ings. These findings suppo he Dwell Hypo hesis, demons a ing ha isual a en ion is modula ed by bo h agen con ex uali y and en i onmen al cong uency. Specifically, con ex ual agen s a ac ed mo e gaze ime compa ed o acon ex ual agen s, confi ming ha con ex ually meaning ul elemen s elici longe dwell imes. Howe e , he compe i i e ela ionship be ween agen s and hei su oundings sugges s ha when agen s ma ch hei en i onmen , a en ion shi s owa d he b oade spa ial con ex a he han he agen i sel . This highligh s he dynamic in e play be ween agen p esence and scene in eg a ion in guiding isual a en ion. 2.2 Tes ing o spa ial knowledge acquisi ion We assessed pa icipan s’spa ial knowledge acquisi ion using poin ing asks. In he Poin ing o Buildings ask, pa icipan s we e gi en sc eensho s o a ge buildings, while in he Poin ing o Agen s ask, hey ecei ed sc eensho s o agen s agains a g ey backg ound. Pa icipan s we e asked o poin owa d he a ge . Accu acy was measu ed by calcula ing he angula di e ence be ween hei poin ing di ec ion and he di ec ion o he cen e o he a ge building o agen . This angula e o se ed as he pe o mance indica o , wi h a pe ec sco e yielding ze o deg ees o e o and highe alues indica ing g ea e inaccu acy. 2.2.1 Poin ing o buildings Acco ding o ou Pe o mance Hypo hesis, pa icipan s would ha e mo e accu a e poin ing o public buildings, especially when a ge s had ac i e con ex ually cong uen agen -building pai s. We applied a linea mixed-e ec s model o p edic poin ing e o s, inco po a ing fixed and andom e ec s. The andom e ec s accoun ed o he es loca ion o he poin ing ask (28 dis inc loca ions wi h epea ed measu es). The fixed e ec s included he ype o building, ype o agen , cong uency pai s, and he in e ac ion be ween agen and building (R2 m0.02, R2 c0.56). The analysis e ealed ha poin ing accu acy was significan ly be e , wi h lowe e o s in public buildings (β−5.51, SE 1.69, −3.28, p<.001, η2 p.0.0088). Con ex ual agen s significan ly imp o ed poin ing accu acy compa ed o acon ex ual agen s (β−7.17, SE 1.72, −4.17, p<.001, η2 p.0179). The cong uency o agen ac ions also played a c ucial ole, wi h incong uen pai s (whe e agen ac ions did no ma ch he con ex ) leading o be e pe o mance han cong uen pai s (β6.88, SE 1.97, 3.50, p<.001, η2 p.0122). Addi ionally, he in e ac ion be ween agen and building ype was non-significan , indica ing ha he main ac o cap u ed he ele an in o ma ion ega ding pe o mance (β4.03, SE 2.45, 1.64, p.101). These esul s p o ide suppo o he Pe o mance Hypo hesis, sugges ing ha con ex ual agen s aid spa ial knowledge acquisi ion. A e he linea mixed-e ec s analysis, we examined he es ima ed ma ginal means (EMM) o cla i y how di e en ac o s influenced poin ing accu acy. Public buildings (EMM 47.4, SE 2.81) esul ed in significan ly lowe e o s han esiden ial buildings (EMM 50.9, SE 2.94). acon ex ual agen s we e associa ed wi h highe e o s (EMM 51.7, SE 2.93) compa ed o con ex ual agen s (EMM 46.6, SE 2.82). The in e ac ion be ween agen and building ype e ealed ha in esiden ial con ex s, acon ex ual agen s esul ed in he highes e o s (EMM 54.5, SE 3.03), while con ex ual agen s in esiden ial con ex s showed lowe e o s (EMM 46.8, SE 2.77). In public con ex s, acon ex ual agen s had highe e o s (EMM 49.2, SE 2.75), while con ex ual agen s in public buildings showed he lowes e o s (EMM 45.8, SE 2.81). The cong uency o agen ac ions also played a key ole, wi h incong uen pai s pe o ming be e han cong uen pai s. Specifically, incong uen pai s had lowe e o s (EMM 45.7, SE 2.81) compa ed o cong uen pai s (EMM 52.6, SE 3.14). The findings emphasize ha con ex ually incong uen agen s imp o e poin ing accu acy, educing e o s and equalizing pe o mance ac oss building ypes (see Figu e 5a). 2.2.2 Poin ing o agen s Acco ding o ou Pe o mance Hypo hesis, pa icipan s would demons a e lowe poin ing e o s o he loca ions o con ex ual agen s, pa icula ly hose posi ioned in on o public buildings. To es his hypo hesis, we employed a linea mixed-e ec s model simila o he one used o analyzing poin ing- o-building pe o mance. The model included c ossed andom e ec s o subjec s and he s a ing loca ions o he poin ing asks, co e ing 28 dis inc loca ions. The analysis e ealed (R2 m0.01, R2 c0.49) a significan main e ec o he building ype (β4.41, SE 2.12, 2.08, p.038, η2.0022). Howe e , he in e ac ion be ween con ex and agen ac ion was non-significan (β−0.21, SE 3.22, −0.06, p.949). Consis en wi h he Pe o mance Hypo hesis, pa icipan s exhibi ed lowe poin ing e o s o con ex ual agen s (EMM 53.7, SE 3.06) compa ed o acon ex ual agen s (EMM 57.2, SE 3.02), indica ing ha con ex ual agen s we e be e emembe ed. Howe e , con a y o expec a ions, poin ing e o s we e g ea e o public buildings (EMM 56.6, SE 3.04) compa ed o esiden ial buildings (EMM 54.3, SE 3.03). These findings sugges ha while agen con ex uali y played a ole in educing poin ing e o s, he expec ed acili a i e e ec o public buildings did no eme ge. Ins ead, pa icipan s demons a ed be e ecall o agen s a esiden ial loca ions, sugges ing ha memo y encoding may ha e been influenced by o he en i onmen al o a en ional ac o s beyond public-p i a e dis inc ions see Figu e 5b. 2.2.3 Accu acy di e ences be ween poin ing o buildings and poin ing o agen s We expec ed pa icipan s o be less p ecise when poin ing o agen s imuli compa ed o building s imuli. To in es iga e his, we fi s assessed whe he pa icipan s exhibi ed significan ly lowe F on ie s in Vi ual Reali y on ie sin.o g08 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 p ecision when poin ing o agen s. This was achie ed by fi ing a wo-c ossed andom e ec s model (i.e., ID and poin ing loca ion) and p edic ing poin ing e o wi h he ype o s imuli (agen s building) as he sole p edic o . The analysis e ealed (R2 m0.008, R2 c0.51) ha pa icipan s we e indeed significan ly less p ecise when poin ing o agen s imuli (β−7.74, SE 0.98, p<.001, η2.0079) compa ed o hei pe o mance in he poin ing- o- building ask. The es ima ed ma ginal means showed ha pa icipan s had highe poin ing e o s o agen s imuli (EMM 54.4, SE 2.55) compa ed o building s imuli (EMM 46.7, SE 2.53). This clea di e ence is illus a ed in Figu e 5c, whe e he a e age pe o mance indica es ha pa icipan s could ecall he p ecise loca ions o buildings mo e accu a ely han human agen s. The da a e eal ha while building loca ions we e ecalled mo e accu a ely, agen s may ha e se ed as a salien ea u e, enhancing he o e all spa ial ecall e en i hei p ecise posi ions we e less accu a ely emembe ed. 2.2.4 Inclusion o gaze as a p edic o o pe o mance To assess whe he he ime spen looking a bo h agen s and buildings would significan ly p edic pa icipan s’pe o mance, we inco po a ed he dwell ime in seconds each pa icipan spen gazing a he agen s and buildings as fixed e ec s in he poin ing- o- building and poin ing- o-agen asks (see Figu e 6). Fo he poin ing- o-building ask, he esul s showed (R2 m0.007, R2 c0.144) ha only he dwell ime on buildings significan ly p edic ed pe o mance (β−0.18, SE 0.05, −3.62, p<.001, η2.0012), while he dwell ime on agen s did no ha e a significan e ec (β−0.27, SE 0.15, −1.79, p.073). To assess he obus ness o ou findings, we conduc ed a sensi i i y analysis using boo s ap esampling (1,000 i e a ions). The esul s confi med ha all significan p edic o s in he model e ained hei e ec s, wi h boo s ap confidence in e als no c ossing ze o. Con e sely, p edic o s ha we e non-significan in he o iginal model had confidence in e als ha included ze o, indica ing g ea e unce ain y in hei e ec s. This alignmen be ween he boo s ap and model-based esul s sugges s ha ou findingsa es ableandno undulyinfluenced by sample a iabili y as can be seen on Table 1.The ypeo agen (con ex ual s acon ex ual) emained a significan p edic o , wi h con ex ual agen s leading o be e pe o mance (β−6.79, SE 1.74, −3.91, p<.001, η2.0016). The ype o building also emained significan (β−5.71, SE 1.68, −3.41, p<.001, η2.0010). Addi ionally, he cong uency be ween he agen s and he building showed a significan e ec (R2 m0.007, R2 c0.144) e ealed ha nei he he dwell ime on agen s (β−0.47, SE 0.24, −1.72, p.086) no he dwell ime on buildings (β−0.002, SE 0.09, −0.02, p.981) we e significan p edic o s o pe o mance. Howe e , building ype, wi h he opposi e pa e n as in poin ing o buildings (i.e., esiden ial loca ions being be e emembe ed, β4.35, SE 1.53, 2.60, p.009, η2 p0.0012) wi h no o he significan e ec s. Compa ing he esul s om he wo asks, i becomes e iden ha he e is an in e se ela ionship be ween he abili y o loca e agen s and buildings, as bo h compe e o a en ion. Con ex ual agen s imp o ed he ecall o hei loca ions, ye he p esence o public buildings appea ed o de ac om he abili y o emembe he agen . The e o e, while agen s may se e as use ul p oxies o FIGURE 5 Analysis o poin ing accu acy ac oss expe imen al condi ions. This figu e illus a es he absolu e poin ing angula e o , a measu e o spa ial ecall accu acy whose lowe alues indica e be e pe o mance, wi h da a poin s ep esen ing he mean e o a es and e o ba s indica ing s anda d e o s (a) poin ing o buildings: This plo delinea es pa icipan s’poin ing accu acy o buildings pai ed wi h con ex ual and acon ex ual agen s (b) Poin ing o Agen s: He e, pa icipan s’accu acy in poin ing o agen s is compa ed, wi h a ocus on he agen ’s ac i i y le el and he ype o building (c) Agg ega ed Task Pe o mance: This plo displays o e all pe o mance ac oss bo h building and agen poin ing asks. F on ie s in Vi ual Reali y on ie sin.o g09 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 he a ge . In expe imen 1, he a ge s we e solely buildings (i.e., poin ing- o-building). In con as , in expe imen 2, we included a es ha displayedonly heagen agains ag aybackg ound (i.e., poin ing- o-agen ). To accoun o he sequence e ec , he ask o de was s a egically andomized, wi h one-hal o he pa icipan s s a ing wi h he poin ing- o-agen ask ollowed by he poin ing- o- building ask and he o he comple ing he asks in e e se o de . Finally, all pa icipan s comple ed h ee sel -assessmen ques ionnai es ha inqui ed abou hei pe cep ion o each agen , hei social endencies in eal li e, and he ealism le el o he VR. The ques ionnai e da a has no been analyzed in his wo k. 4.5 Expe imen al se up The explo a ion and ask assessmen sessions we e conduc ed wi h a desk op compu e wi h an In el (R) Xeon (R) W-2133 CPU, 16 GB RAM, and a N idia RTX 2080 Ti g aphics ca d. The VR en i onmen was ende ed wi h a HTC Vi e P o Eye head-moun ed display (HMD) se up wi h a e esh a e o 90 Hz and ho izon al and e ical field o iew o 106°and 110°, espec i ely. We used ou S eamVR Base S a ions 2.0, an HTC VIVE body acke 2.0, and Val e Index con olle s o moni o pa icipan s’posi ions wi hin he en i onmen . This combined se up achie ed sub-millime e p ecision in cap u ing he head, body, and eye posi ions, as well as o a ion and o ien a ion. 4.6 Spa ial ask Pa icipan s pe o med poin ing asks in VR om a fi s -pe son pe spec i e. In he poin ing- o-building ask, hey we e elepo ed o 28 unique loca ions wi hin he ci y, each se ing as a dis inc e e ence poin . The sequence o hese loca ions was andomized o each pa icipan o p e en o de e ec s. To u he minimize sys ema ic biases, pa icipan s’o ien a ions a e elepo a ion we e also andomized, ensu ing ha hey did no always begin acing he same di ec ion. A each loca ion, pa icipan s poin ed epea edly owa d one o 56 po en ial a ge s, ep esen ed by s a ic images o buildings, wi h agen s posi ioned in on o hem. These images we e cap u ed pe pendicula ly a a heigh o 1.80 m o ensu e consis ency in isual p esen a ion. The ials began wi h a isual and audi o y cue: a 25 ms g een ci cula loading ba a he sc een’s cen e accompanied by a beep. The a ge image appea ed a he uppe cen e o he sc een, wi h a g een dashed lase beam p o iding a isual guide o poin ing. Pa icipan s indica ed hei di ec ion by p essing a igge bu on on hei con olle s. Each ial was imed o 30 s, au oma ically concluding i a di ec ion was no indica ed wi hin his pe iod. Pe o mance was assessed by measu ing he angula di e ence be ween he pa icipan ’s poin ing di ec ion and he p ecise cen e ec o o he a ge loca ion. In he ini ial expe imen , pa icipan s comple ed 336 ials, poin ing a 12 unique a ge s om each o he 28 e e ence loca ions. The educ ion in he numbe o ials in he second expe imen , whe e pa icipan s comple ed only 224 ials, poin ing a eigh dis inc a ge s om each loca ion, was necessi a ed by he addi ional poin ing- o-agen s ask. In bo h expe imen s, he ials we e balanced o ensu e an e en dis ibu ion o a ge ypes: 50% di ec ed pa icipan s o public loca ions and 50% o esiden ial a eas. 4.7 Poin ing o agen s ask This ask was undamen ally equi alen o he poin ing- o- building bu specifically ocused on agen s as he a ge . The a ge s ea u ed cen e ed sc eensho s o indi idual agen s se agains a g ey backd op, cap u ed pe pendicula ly a a heigh o 1.80 m. Pa icipan s unde ook 224 ials and we e ins uc ed o poin a eigh unique agen a ge s om each o he 28 e e ence loca ions. The ask mi o ed he s uc u e o he building ask in e ms o a ge placemen , isual and audi o y cues, and ime cons ain s o main ain consis ency in he es ing condi ions. Addi ionally, he sequence o hese poin ing loca ions was andomized o each pa icipan , wi h he goal o a oiding o de e ec s and p ese ing he in eg i y o he expe imen al da a. 4.8 “F agebogen Räumliche S a egien” (FRS) ques ionnai e The “F agebogen Räumliche S a egien”[FRS; Münze and Hölsche (2011)] ques ionnai e is a 7-poin Like scale ha asks he pa icipan s o es ima e hei spa ial o ien a ion abili ies in h ee a eas o spa ial knowledge in eal-wo ld scena ios. Fi s , he global sub-scale consis ed o 10 i ems (α0.89)inqui ing abou he subjec ’s abili y o na iga e ou es om an egocen ic pe spec i e. The su ey sub-scale inco po a es se en i ems (α0.87) ocused on he subjec ’sabili y o men al mapping om an allocen ic pe spec i e. The ca dinal sub-scale comp ises wo i ems (α0.80) ha que y he abili y o poin owa d ca dinal poin s. The FRS measu es pa icipan s’likelihood o apply spa ial s a egies ela ed o egocen ic/global knowledge, su ey knowledge, o ca dinal di ec ions, espec i ely. 4.9 Mo emen acking in he ci y 4.9.1 Na iga ional acking To analyze pa icipan ajec o ies and hei explo a ion decisions, we c ea ed a da a-d i en g aph based on hei ac ual ajec o ies while explo ing he i ual ci y. Thus, his g aph eflec s only he pa hs and a eas ha pa icipan s walked h ough, cons i u ed by he s ee s (edges) and c ossings o decision poin s (nodes) ha pa icipan s walked h ough. We fi s gene a ed a hea map o pa icipan s’mo emen o ansi ion om aw sequen ial coo dina es o spa ial da a. This hea map accoun ed o he numbe o imes a pa icipan s ood a a specific cell wi hin a defined 4 m × 4 m g id on op o he ci y map. This hea map was hen u ned in o a bina y image, whe e cells isi ed a leas once we e assigned a alue o one and cells wi hou isi s a alue o ze o. The esul ing image clea ly ou lined he walkable pa hs and connec ions wi hin he ci y. We filled isola ed holes wi hin he s ee s o ensu e he algo i hm gene a ing he g aph did no c ea e ex aneous nodes in hese a eas. We hen gene a ed a skele on om he bina y image, educing he ci y’s ep esen a ion o a one-pixel F on ie s in Vi ual Reali y on ie sin.o g16 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 wid h while p ese ing i s opog aphy and connec i i y. This skele on was he ounda ion o iden i ying he g aph’s nodes and edges. Using he ex e nal Py hon lib a y “sknw,”we gene a ed a Ne wo kX g aph om he ci y skele on image. Nodes we e numbe ed sequen ially, and edges we e named based on he nodes hey connec ed (see Figu e 2a). Fo example, edge [76,60] connec ed nodes 76 and 60. Some manual adjus men s, such as adding an edge, we e necessa y o pe ec he g aph. Each pixel in he skele on was ecognized as belonging o a node o an edge. This in o ma ion was la e used o plo he g aph and calcula e dis ances o iden i y he g aph elemen s isi ed by pa icipan s du ing hei sessions. To analyze explo a ion s a egies, we de eloped an algo i hm ha con e s pa icipan s’coo dina es da a in o a sequence o isi ed nodes and edges. This p ocess in ol ed mapping he coo dina es o he skele on hea map’s 4 m × 4 m cells and de e mining he closes g aph elemen using he Euclidean dis ance o mula. The algo i hm acked ansi ions be ween edges and nodes, eco ding each isi ed elemen . This eco d included he ime o en y, g aph elemen ype (node/edge), and numbe o isi s o he g aph wi h a, well as he a ailable pa hs om he node o o igin and hei espec i e numbe o p e ious isi s o he possible elemen s a ailable om ha posi ion. We adjus ed he nodes’ adii o ma ch he wid h o he co esponding s ee s. This adjus men ensu ed accu a e de ec ion o pa icipan p esence a nodes, p e en ing unno iced ansi ions be ween edges. O e lapping node adii in a eas wi h se e al sho s ee s we e esol ed by assigning pa icipan posi ions o he nea es node cen oid. 4.9.2 Na iga ional pa e n classifica ion: S a egy ma ix We condensed pa icipan s’decisions a nodes in o a s a egy ma ix o analyze hei explo a ion pa e ns. The s a egy ma ix was o ganized wi h he numbe o isi s o he chosen node on he ows and he numbe o isi s o he no -chosen nodes on he columns. Each decision was eco ded in he ow co esponding o he numbe o p e ious isi s o he selec ed pa h. We added one o each column co esponding o he numbe o p io isi s on he o he a ailable pa hs om ha decision poin ha we e no selec ed. Fo ins ance, i a pa icipan was a a node ha had h ee di ec neighbo s, isi ed ze o, fi e, and wo imes, and chose o mo e in he di ec ion o he node isi ed wo imes, we would add a coun o one o he posi ions [0,2] and [5,2]. This ep esen s ha he pa icipan mo ed o a node wi h wo isi s o e he op ions ha had been isi ed ze o and fi e imes (see Figu e 2b). Decisions abo e he diagonal line in he s a egy ma ix we e conside ed conse a i e, indica ing ha pa icipan s p e e ed nodes hey had isi ed mo e equen ly in he p e ious example, he one added a [0,2]. In con as , decisions below he diagonal line we e conside ed explo a o y, as pa icipan s chose nodes wi h ewe isi s compa ed o hei neighbo ing nodes, in he example abo e he one added a [5,2]. Decisions exac ly on he diagonal we e neu al, as bo h he chosen and he adjacen nodes had he same numbe o p e ious isi s. This me hod allowed us o quan i y and compa e pa icipan s’explo a o y and conse a i e endencies as hey na iga ed he i ual ci y, p o iding insigh in o hei spa ial decision-making p ocesses as hey gained expe ience inside Wes b ook. 4.9.3 Eye- acking p ep ocessing and classifica ion We applied a eloci y-based algo i hm ha classified con inuous eye mo emen s in o gazes and saccades and co ec ed he esul ing gazes o he pa icipan s’mo emen in he 3D en i onmen , as de eloped by (Nol e e al., 2024). In p epa a ion o his algo i hm, we p ep ocessed he da a by excluding po ions de ec ed as in alid (e.g., blinks). In cases whe e mo e han one collide hi was de ec ed wi hin he same sample, we e ained he closes hi o he pa icipan , excep o backg ound collide s, such as lea es o ences, in which case we kep he second closes hi . As a las s ep, we d opped duplica ed samples and applied a 5-poin median fil e o he gaze coo dina es. The algo i hm calcula es he median eloci y o eye mo emen s wi hin a ime window (in ou case, a 10-s window). Samples exceeding his eloci y h eshold a e iden ified as saccades, while hose alling below he h eshold a e classified as gazes. To ensu e accu acy, we applied ou lie de ec ion based on median absolu e de ia ion o co ec o gaze e en s wi h anomalous du a ions (i.e., exceeding h ee median absolu e de ia ions). Dwell ime was defined as he cumula i e ime a subjec spen gazing a a specific objec wi hin he ci y ac oss all sessions. We compu ed he dwell ime o each subjec -objec pai du ing hei en i e explo a ion pe iod wi hin he Wes b ook en i onmen . 4.10 Da a analysis Gi en he hie a chical s uc u e o ou da a, we employed Linea Mixed-E ec s Models o ou analysis. The modeling was done using R 4.3.2 wi h he lme () unc ion om he lme4 package. We used Res ic ed Maximum Likelihood o es ima ion and he nlop w ap op imize (Ba es e al., 2015). This app oach allows us o handle he nes ed s uc u e o ou da a, wi h andom e ec s o accoun o wi hin-subjec a iabili y. Fixed e ec s wi h wo le els we e e ec -coded o ensu e he be as eflec he di e ences be ween hese le els, using he fi s le el o each pai as he base o compa ison. 4.10.1 Explo a ion phase analysis: Na iga ional co e age o he ci y To analyze pa icipan s’ ee explo a ion pa e ns, we quan ified hei walking beha io h ough he i ual ci y using a p imal ci y g aph Neal (2013). The co e age a io, defined as he p opo ion o unique nodes isi ed du ing each session, was modeled using a linea mixed-e ec s app oach. The model included fixed e ec s o he session, expe imen , and in e ac ions, wi h planned con as es ing o each session agains he fi s . Random in e cep s o pa icipan s we e included o accoun o epea ed measu es wi hin subjec s. The model o mula o he indi idual session co e age a io was (see Equa ion 1): Indi idual Ra io ~ Session × Expe imen +1|pa icipan  (1) The same s uc u e was used o es o he cumula i e a io o isi ed nodes, in which we kep ack o how many unique decision poin s each pa icipan had isi ed, accumula ing hem be ween sessions. The model o mula o he cumula i e co e age a io was (see Equa ion 2): Cumula i e Ra io ~ Session × Expe imen +1|pa icipan  (2) F on ie s in Vi ual Reali y on ie sin.o g17 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 4.10.2 Walking s a egies: Explo a o y s conse a i e na iga ional beha io To in es iga e he ex en o which pa icipan s used conse a i e o explo a o y walking s a egies, we modeled he numbe o decisions a each node. A linea mixed-e ec s model (see Equa ion 3) was used wi h he session, s a egy (conse a i e s explo a o y), and expe imen as fixed e ec s and andom in e cep s o pa icipan s. Numbe o decisions ~ Session × S a egy +Expe imen +1|pa icipan  (3) 4.10.3 Visual beha io du ing explo a ion We quan ified pa icipan s’ isual beha io by summing he cumula i e ime spen gazing a objec s (dwell ime) du ing he en i e explo a ion phase. Sepa a e models p edic ed dwell ime on agen s (see Equa ion 4) and buildings (see Equa ion 5), wi h fixed e ec s o agen ype (acon ex ual s con ex ual), con ex e ec ( esiden ial s public), he cong uency o he agen wi h hei su oundings (no cong uen s cong uen ), and hei in e ac ions, including andom e ec s o pa icipan s wi h he ollowing o mulas: Dwell Timeagen ~1+Building Ca ego y × Agen Ca ego y +1|pa icipan  +1|poin ing loca ion  (4) Dwell Timebuilding ~1+Building × Agen Ca ego y +Cong uence +1|pa icipan  +1|poin ing loca ion  (5) 4.10.4 Poin ing ask: Poin ing o buildings We assessed spa ial knowledge using poin ing asks, calcula ing he angula e o as he pe o mance indica o . Two sepa a e linea mixed-e ec s models (see Equa ions 6,7) p edic ed poin ing e o based on building ype ( esiden ial s public), agen ype (acon ex ual s con ex ual), he cong uency o he agen s wi h hei su oundings, dwell ime on agen s, dwell ime on buildings, and he in e ac ion o agen and building ype. Random e ec s o pa icipan s and poin ing loca ions we e included in each model: Poin ing E o building ~1+Building Ca ego y × Agen Ca ego y +Cong uency +Dwell Timebuilding +Dwell Timeagen +1|pa icipan  +1|poin ing loca ion  (6) Poin ing E o agen ~1+Building Ca ego y × Agen Ca ego y +Dwell Timeagen +Dwell Timebuilding +1|pa icipan  +1|poin ing loca ion  (7) 4.10.5 Compa ing poin ing accu acy be ween asks To compa e accu acy be ween poin ing o buildings and poin ing o agen s, we used a model wi h he ype o s imuli (agen s building) as he fixed e ec and pa icipan s and poin ing loca ions as andom e ec s (see Equa ion 8): Poin ing E o ~ 1 +Tes +1|pa icipan  +1|poin ing loca ion  (8) Following model fi ing, we pe o med likelihood a io es s o compa e each model agains a null model con aining only he in e cep , e alua ing he added p edic i e powe o ou ac o s. 4.10.6 Pe mission o euse and copy igh Figu es, ables, and images will be published unde a C ea i e Commons CC-BY licence and pe mission mus be ob ained o use o copy igh ed ma e ial om o he sou ces (including e-published/adap ed/modified/pa ial figu es and images om he in e ne ). I is he esponsibili y o he au ho s o acqui e he licenses, o ollow any ci a ion ins uc ions eques ed by hi d-pa y igh s holde s, and co e any supplemen a y cha ges. Da a a ailabili y s a emen The aw da a suppo ing he conclusions o his a icle will be made a ailable by he au ho s, wi hou undue ese a ion. E hics s a emen The s udies in ol ing humans we e app o ed by Uni e si y o Osnab ück E hics Commi ee. The s udies we e conduc ed in acco dance wi h he local legisla ion and ins i u ional equi emen s. The pa icipan s p o ided hei w i en in o med consen o pa icipa e in his s udy. Au ho con ibu ions TS: Concep ualiza ion, Fo mal Analysis, In es iga ion, W i ing –o iginal d a , W i ing – e iew and edi ing. MS: Fo mal Analysis, In es iga ion, Resou ces, W i ing – e iew and edi ing. KG: Fo mal Analysis, In es iga ion, Resou ces, W i ing – e iew and edi ing. VS: Concep ualiza ion, Resou ces, W i ing – e iew and edi ing. DN: Fo mal Analysis, In es iga ion, Resou ces, W i ing – e iew and edi ing. SK: Concep ualiza ion, Supe ision, W i ing –o iginal d a , W i ing – e iew and edi ing. GP: Concep ualiza ion, Funding acquisi ion, Supe ision, W i ing – e iew and edi ing. PK: Concep ualiza ion Funding acquisi ion, Supe ision, W i ing –o iginal d a , W i ing – e iew and edi ing. Funding The au ho (s) decla e ha financial suppo was ecei ed o he esea ch and/o publica ion o his a icle. The Uni e si y o Osnab ück suppo ed his wo k in coope a ion wi h he Deu sche Akademische Aus auschdiens (DAAD), G an No. 57440921 and he Deu sche Fo schungsgemeinscha (DFG, Ge man Resea ch Founda ion)—GRK 2340. F on ie s in Vi ual Reali y on ie sin.o g18 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 Acknowledgmen s The au ho s exp ess hei g a i ude o e e yone who con ibu ed o his p ojec . They would like o hank No a Maleki and Linus Tiemann o hei help in de eloping he VR ci y and he Poin ing Tasks equi ed o hese expe imen s, and Philipp Spaniol o his 3- dimensional a and implemen a ion o di e en ial loading o le els o de ails wi h he agen s on his scene. Conflic o in e es The au ho s decla e ha he esea ch was conduc ed in he absence o any comme cial o financial ela ionships ha could be cons ued as a po en ial conflic o in e es . The au ho (s) decla ed ha hey we e an edi o ial boa d membe o F on ie s, a he ime o submission. This had no impac on he pee e iew p ocess and he final decision. Publishe ’s no e All claims exp essed in his a icle a e solely hose o he au ho s and do no necessa ily ep esen hose o hei a filia ed o ganiza ions, o hose o he publishe , he edi o s and he e iewe s. Any p oduc ha may be e alua ed in his a icle, o claim ha may be made by i s manu ac u e , is no gua an eed o endo sed by he publishe . Re e ences Bach, P., Nicholson, T., and Hudson, M. (2014). The a o dance-ma ching hypo hesis: how objec s guide ac ion unde s anding and p edic ion. F on . Hum. Neu osci. 8, 254. doi:10.3389/ nhum.2014.00254 Bajo unai e, L., B ews e , S., and R. Williamson, J. (2022). ““Reali y ancho s”: b inging cues om eali y in o VR on public anspo o alle ia e sa e y and com o conce ns,”in CHI con e ence on human ac o s in compu ing sys ems ex ended abs ac s (New O leans LA USA: ACM), 1–6. doi:10.1145/3491101.3519696 Ba es, D., Mächle , M., Bolke , B., and Walke , S. (2015). Fi ing linea mixed-e ec s models using lme4.J. S a . So w. 67. doi:10.18637/jss. 067.i01 Bicanski, A., and Bu gess, N. (2020). Neu onal ec o coding in spa ial cogni ion. Na . Re . Neu osci. 21, 453–470. doi:10.1038/s41583-020-0336-9 Bonne , M. F., and Eps ein, R. A. (2017). Coding o na iga ional a o dances in he human isual sys em. P oc. Na l. Acad. Sci. 114, 4793–4798. doi:10.1073/pnas. 1618228114 Bonsch, A., Radke, S., O e a h, H., Asche, L. M., Wend , J., Vie jahn, T., e al. (2018). “Social VR: how pe sonal space is a ec ed by i ual agen s’emo ions,”in 2018 IEEE con e ence on i ual eali y and 3D use in e aces (VR) (Reu lingen: IEEE), 199–206. doi:10.1109/VR.2018.8446480 Bu goon, J. K. (2015). “Expec ancy iola ions heo y,”in The in e na ional encyclopedia o in e pe sonal communica ion. Edi o s C. R. Be ge , M. E. Rolo , S. R. Wilson, J. P. Dilla d, J. Caughlin, and D. Solomon 1 edn. (Wiley), 1–9. doi:10. 1002/9781118540190.wbeic102 Choi, S. H., Jung, T. M., Lee, J. E., Lee, S.-K., Sohn, Y. H., and Lee, P. H. (2012). Volume ic analysis o he subs an ia innomina a in pa ien s wi h Pa kinson’s disease acco ding o cogni i e s a us. Neu obiol. Aging 33, 1265–1272. doi:10.1016/j. neu obiolaging.2010.11.015 Dal on, R. C., Hölsche , C., and Mon ello, D. R. (2019). Wayfinding as a social ac i i y. F on . Psychol. 10, 142. doi:10.3389/ psyg.2019.00142 Dickinson, P., Ge ling, K., Hicks, K., Mu ay, J., Shea e , J., and G eenwood, J. (2019). Vi ual eali y c owd simula ion: e ec s o agen densi y on use expe ience and beha iou . Vi ual Real. 23, 19–32. doi:10.1007/s10055-018-0365-0 Eks om, A. D., and Isham, E. A. (2017). Human spa ial na iga ion: ep esen a ions ac oss dimensions and scales. Cu . Opin. Beha . Sci. 17, 84–89. doi:10.1016/j.cobeha. 2017.06.005 Eps ein, R. A., Pa ai, E. Z., Julian, J. B., and Spie s, H. J. (2017). The cogni i e map in humans: spa ial na iga ion and beyond. Na . Neu osci. 20, 1504–1513. doi:10.1038/nn. 4656 Fa an, E. K., Fo mby, S., Daniyal, F., Holmes, T., and Van He wegen, J. (2016). Rou e-lea ning s a egies in ypical and a ypical de elopmen ; eye acking e eals a ypical landma k selec ion in Williams synd ome. J. In ellec . Disabil. Res. 60, 933–944. doi:10.1111/ji .12331 Fa zan a , D., Spie s, H. J., Mosco i ch, M., and Rosenbaum, R. S. (2023). F om cogni i e maps o spa ial schemas. Na . Re . Neu osci. 24, 63–79. doi:10.1038/s41583- 022-00655-9 F anke, C., and Schweika , J. (2017). Men al ep esen a ion o landma ks on maps: in es iga ing ca og aphic isualiza ion me hods wi h eye acking echnology. Spa ial Cogni ion and Compu . 17, 20–38. doi:10.1080/13875868.2016.1219912 Ge , A. L., Ehinge , B. V., Kie zmann, T. C., and Konig, P. (2020). “Faces s ongly a ac ea ly fixa ions in na u ally sampled eal-wo ld s imulus ma e ials,”in ACM symposium on eye acking esea ch and applica ions (s u ga Ge many: acm),1–5. doi:10.1145/3379156.3391377 G iesbaue , E.-M., Manley, E., Wiene , J. M., and Spie s, H. J. (2022). London axi d i e s: a e iew o neu ocogni i e s udies and an explo a ion o how hey build hei cogni i e map o London. Hippocampus 32, 3–20. doi:10.1002/hipo.23395 Gunalp, P., Moossaian, T., and Hega y, M. (2019). Spa ial pe spec i e aking: e ec s o social, di ec ional, and in e ac i e cues. Mem. and Cogni ion 47, 1031–1043. doi:10. 3758/s13421-019-00910-y Ha el, A., K a i z, D. J., and Bake , C. I. (2014). Task con ex impac s isual objec p ocessing di e en ially ac oss he co ex. P oc. Na l. Acad. Sci. 111, E962–E971. doi:10. 1073/pnas.1312567111 Hende son, J. M. (2007). Rega ding scenes. Cu . Di . Psychol. Sci. 16, 219–222. doi:10.1111/j.1467-8721.2007.00507.x I o, H. T., Zhang, S.-J., Wi e , M. P., Mose , E. I., and Mose , M.-B. (2015). A p e on al– halamo–hippocampal ci cui o goal-di ec ed spa ial na iga ion. Na u e 522, 50–55. doi:10.1038/na u e14396 Janzen, G., Wagens eld, B., and Van Tu ennou , M. (2006). Neu al ep esen a ion o na iga ional ele ance is apidly induced and long las ing. Ce eb. Co ex 17, 975–981. doi:10.1093/ce co /bhl008 Kuehn, E., Chen, X., Geise, P., Ol me , J., and Wolbe s, T. (2018). Social a ge s imp o e body-based and en i onmen -based s a egies du ing spa ial na iga ion. Exp. B ain Res. 236, 755–764. doi:10.1007/s00221-018-5169-7 Li, H., Th ash, T., Hölsche , C., and Schinazi, V. R. (2019). The e ec o c owdedness on human wayfinding and locomo ion in a mul i-le el i ual shopping mall. J. En i on. Psychol. 65, 101320. doi:10.1016/j.jen p.2019.101320 Magui e, E. A., Woolle , K., and Spie s, H. J. (2006). London axi d i e s and bus d i e s: a s uc u al MRI and neu opsychological analysis. Hippocampus 16, 1091–1101. doi:10.1002/hipo.20233 Malanchini, M., Rim eld, K., Shakesha , N. G., McMillan, A., Schofield, K. L., Rodic, M., e al. (2020). E idence o a uni a y s uc u e o spa ial cogni ion beyond gene al in elligence. npj Sci. Lea n. 5, 9. doi:10.1038/s41539-020-0067-8 Mille , A. M. P., Vedde , L. C., Law, L. M., and Smi h, D. M. (2014). Cues, con ex , and long- e m memo y: he ole o he e osplenial co ex in spa ial cogni ion. F on . Hum. Neu osci. 8, 586. doi:10.3389/ nhum.2014.00586 Münze , S., and Hölsche , C. (2011). En wicklung und Validie ung eines F agebogens zu äumlichen S a egien. Diagnos ica 57, 111–125. doi:10.1026/0012-1924/a000040 Neal, Z. P. (2013). “The connec ed ci y: how ne wo ks a e shaping he mode n me opolis,”in Me opolis and mode n li e. 1s ed edn (New Yo k, NY: Rou ledge). Nol e, D., Vidal De Palol, M., Kesha a, A., Mad id-Ca ajal, J., Ge , A. L., Von Bu le , E.-M., e al. (2024). Combining EEG and eye- acking in i ual eali y: ob aining fixa ion-onse e en - ela ed po en ials and e en - ela ed spec al pe u ba ions. A en. Pe cep . and Psychophys. 87, 207–227. doi:10.3758/s13414-024-02917-3 Ohm, C., Mülle , M., Ludwig, B., and Bienk, S. (2014). Whe e is he Landma k? Eye T ack. S ud. La ge-Scale Indoo En i on. Publishe : Uni e si ä Regensbu g. doi:10. 5283/EPUB.31436 Pappala do, L., Simini, F., Rinzi illo, S., Ped eschi, D., Gianno i, F., and Ba abási, A.- L. (2015). Re u ne s and explo e s dicho omy in human mobili y. Na . Commun. 6, 8166. doi:10.1038/ncomms9166 Rounds, J. D., C uz-Ga za, J. G., and Kalan a i, S. (2020). Using pos e io EEG he a band o assess he e ec s o a chi ec u al designs on landma k ecogni ion in an u ban se ing. F on . Hum. Neu osci. 14, 584385. doi:10.3389/ nhum.2020. 584385 F on ie s in Vi ual Reali y on ie sin.o g19 Sánchez Pacheco e al. 10.3389/ i .2025.1497237 Scha e , M., and Schille , D. (2018). Na iga ing social space. Neu on 100, 476–489. doi:10.1016/j.neu on.2018.10.006 Schläp e , M., Dong, L., O’Kee e, K., San i, P., Szell, M., Sala , H., e al. (2021). The uni e sal isi a ion law o human mobili y. Na u e 593, 522–527. doi:10.1038/s41586- 021-03480-9 Schmid , V., König, S. U., Dilawa , R., Sánchez Pacheco, T., and König, P. (2023). Imp o ed spa ial knowledge acquisi ion h ough senso y augmen a ion. B ain Sci. 13, 720. doi:10.3390/b ainsci13050720 Sla e , M. (2009). Place illusion and plausibili y can lead o ealis ic beha iou in imme si e i ual en i onmen s. Philosophical T ans. R. Soc. B Biol. Sci. 364, 3549–3557. doi:10.1098/ s b.2009.0138 Summe field, C., and Egne , T. (2009). Expec a ion (and a en ion) in isual cogni ion. T ends Cogni i e Sci. 13, 403–409. doi:10.1016/j. ics.2009.06.003 T e sky, B., and Ha d, B. M. (2009). Embodied and disembodied cogni ion: spa ial pe spec i e- aking. Cogni ion 110, 124–129. doi:10.1016/j.cogni ion.2008.10.008 Wal e , J. L., Essmann, L., König, S. U., and König, P. (2022). Finding landma ks - an in es iga ion o iewing beha io du ing spa ial na iga ion in VR using a g aph- heo e ical analysis app oach. PLOS Compu . Biol. 18, e1009485. doi:10.1371/jou nal. pcbi.1009485 Wes , G. L., Zendel, B. R., Konishi, K., Benady-Cho ney, J., Bohbo , V. D., Pe e z, I., e al. (2017). Playing Supe Ma io 64 inc eases hippocampal g ey ma e in olde adul s. PLOS ONE 12, e0187779. doi:10.1371/jou nal.pone.0187779 Wiene , J. M., Ca oll, D., Moelle , S., Bibi, I., I ano a, D., Allen, P., e al. (2020). A no el i ual- eali y-based ou e-lea ning es sui e: assessing he e ec s o cogni i e aging on na iga ion. Beha . Res. Me hods 52, 630–640. doi:10.3758/s13428-019-01264-8 Wol e, J. M. (2020). Visual sea ch: how do we find wha we a e looking o ? Annu. Re . Vis. Sci. 6, 539–562. doi:10.1146/annu e - ision-091718-015048 Woolle , K., and Magui e, E. (2011). Acqui ing “ he knowledge”o london’s layou d i es s uc u al b ain changes. Cu . Biol. 21, 2109–2114. doi:10.1016/j.cub.2011. 11.018 F on ie s in Vi ual Reali y on ie sin.o g20 Sánchez Pacheco e al. 10.3389/ i .2025.1497237