Psychophysiological responses to eye contact in a live interaction and in video call

Psychophysiology. 2020;00:e13587. | 1 o 12 h ps://doi.o g/10.1111/psyp.13587 wileyonlinelib a y.com/jou nal/psyp 1 | INTRODUCTION Eye con ac has an impo an ole in social in e ac ion. Ano he pe son's gaze di ec ion indica es he di ec ion o hei a en ion (I ie & Ba y,2009), and, when i is di ec ed owa d onesel , i is usually pe cei ed as a posi i e social sig- nal, such as a sign o liking o communica i e in en ( o a e iew, see Kleinke,1986). People who make eye con ac a e pe cei ed as mo e pleasan , compe en , and a ac i e han hose who a oid di ec gaze. Eye con ac has also been ound o elici a ec - ela ed psychophysiological esponses in he pe cei e ( o a e iew, see Hie anen, 2018). In e es ingly, howe e , i is unclea wha gi es ise o hese e ec s. Some p e ious accoun s ha e ocused on he e ec s elici ed by he me e senso y inpu , ha is, he pe cep ion o a pe son's eyes di ec ed owa d he sel (e.g., Senju & Johnson,2009), whe eas o he s ha e s essed he impo ance o a psycholog- ical mechanism, an expe ience o being seen by ano he pe - son (e.g., Con y, Geo ge, & Hie anen, 2016; Myllyne a & Hie anen,2015; Teu el, Fle che , & Da is,2010). Recei ed: 5 Decembe 2019 | Re ised: 20 Ma ch 2020 | Accep ed: 23 Ma ch 2020 DOI: 10.1111/psyp.13587 ORIGINAL ARTICLE Psychophysiological esponses o eye con ac in a li e in e ac ion and in ideo call Jonne O.Hie anen | Mikko J.Pel ola | Ja i K.Hie anen This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. © 2020 The Au ho s. Psychophysiology published by Wiley Pe iodicals LLC on behal o Socie y o Psychophysiological Resea ch Human In o ma ion P ocessing Labo a o y, Facul y o Social Sciences/Psychology, Tampe e Uni e si y, Tampe e, Finland Co espondence Ja i K. Hie anen, Human In o ma ion P ocessing Labo a o y, Facul y o Social Sciences/Psychology, FI-33014 Tampe e Uni e si y, Tampe e, Finland. Email: [email p o ec ed]i Funding in o ma ion Al ed Ko delin Founda ion (g an o J. O. H.) Abs ac Ano he pe son's gaze di ec ed o onesel elici s au onomic a ousal and acial eac- ions indica ing posi i e a ec in i s obse e . These e ec s ha e only been ound o occu wi h mu ual, li e eye con ac and no in esponse o di ec gaze pic u es o when he obse e belie es ha he li e pe son canno see hem. The ques ion emains whe he he physical p esence o he o he pe son is necessa y o hese e ec s. We measu ed psychophysiological esponses o ano he pe son's di ec e - sus a e ed gaze in h ee condi ions: li e in e ac ion, bidi ec ional ideo call, and wa ching a me e ideo. Au onomic a ousal was measu ed wi h skin conduc ance esponses and acial eac ions wi h acial elec omyog aphy. In he li e and ideo call condi ions, bu no in he me e ideo condi ion, di ec gaze inc eased au onomic a ousal in compa ison o a e ed gaze. In all h ee condi ions, howe e , di ec gaze elici ed posi i e a ec i e acial eac ions. The e o e, an expe ience o being seen is essen ial o he au onomic eac ions bu no o he acial esponses ha a e elici ed by ano he pe son's di ec gaze. Mos impo an ly, he esul s sugges ha he physi- cal p esence o p oximi y o he o he pe son is no necessa y o hese psychophysi- ological esponses o eye con ac . KEYWORDS au onomic a ousal, di ec gaze, EMG, eye con ac , skin conduc ance, ideo call 2 o 12 | HIETANEN ET Al. In acco dance wi h he la e iew, many o he psycho- physiological esponses o eye con ac a e only elici ed by he gaze o a li e pe son. S udies ha ha e con as ed he pe - cep ion o a li e pe son wi h ha o a pic u e o a ideo ha e ound ha he e ec o eye con ac on au onomic ne ous sys em seems o be limi ed o a li e pe son's gaze (Hie anen, Leppänen, Pel ola, Linna-aho, & Ruuhiala,2008; Pönkänen, Pel ola, & Hie anen,2011; P insen & Alae s,2019). In hese s udies, ano he pe son's di ec gaze in compa ison o a e ed gaze elici ed g ea e skin conduc ance esponses (SCR) in- dica ing g ea e au onomic a ousal. This e ec occu ed only when pa icipan s we e acing a eal pe son whe eas, when hey we e looking a a me e pic u e o ideo, he pe - cei ed gaze di ec ion did no in luence au onomic a ousal. Acco dingly, s udies ha ha e used only pic o ial s imuli ha e o en ound no di e ence in SCRs be ween he pe cep- ion o di ec and a e ed gaze (Joseph, Eh man, McNally, & Keehn,2008; Lea i & Dono an,1979; Lyy a, Myllyne a, & Hie anen, 2018; Wiese , Pauli, Alpe s, & Mühlbe ge ,2009), hough some s udies ha e epo ed a g ea e a ousal e- sponse o di ec gaze in ce ain s udy condi ions o pa ic- ipan g oups also when showing pic u es o aces (Con y e  al., 2010; Kylliäinen & Hie anen, 2006; Soussignan e al.,2013). Simila ly, s udies ha ha e assessed a ousal by measu ing pupilla y esponses ha e no ound a consis en e- sponse o pic o ial p esen a ions o di ec e sus a e ed gaze (Kampe, F i h, & F i h, 2003; Po e , Hood, T oscianko, & Mac ae,2006). In acial elec omyog aphy (EMG) s udies, esponses associa ed wi h posi i e emo ion—ac i a ion o he zygoma icus majo muscle and elaxa ion o he co uga o supe cilii muscle—ha e been obse ed in esponse o li e, di ec gaze s imuli bea ing a neu al exp ession (Hie anen e al.,2018). S udies using pic o ial s imuli, howe e , ha e no ound a simila e ec (Mojzisch e al.,2006; Rychlowska, Zinne , Musca, & Nieden hal, 2012; Sch ammel, Pannasch, G aupne , Mojzisch, & Velichko sky, 2009; Soussignan e al.,2013). Mo eo e , s udies compa ing b ain esponses o di e en p esen a ion modes o di ec and a e ed gaze s imuli ha e ound ce ain e ec s only in esponse o a li e pe son's di ec gaze. Two s udies ha in es iga ed he asym- me ical on al b ain ac i i y wi h elec oencephalog aphic (EEG) eco dings ound ela i ely mo e le -sided b ain ac- i i y indica ing inc eased app oach mo i a ion in esponse o li e di ec e sus a e ed gaze (Hie anen e  al., 2008; Pönkänen, Pel ola, e al.,2011). This e ec was also obse ed only when acing a eal pe son and no while looking a a me e pic u e. Simila ly, s udies on e en - ela ed po en ials (ERP) and unc ional magne ic esonance imaging ( MRI) esponses o li e and pic o ial p esen a ions o di ec and a e ed gaze s imuli ha e ound e ec s associa ed wi h social and a ec i e cogni ion (enhanced N170 and EPN ampli udes and an e io medial p e on al co ex ac i i y) in esponse o a li e pe son's di ec e sus a e ed gaze, bu no o pic u es o di ec e sus a e ed gaze (Ca allo e al.,2015; Pönkänen, Alhoniemi, Leppänen, & Hie anen, 2011). Taken oge he , hese s udies s ongly imply ha an expe ience o being seen by ano he pe son may in luence he e ec s ha di ec gaze has on i s obse e . The e ec o being seen by ano he pe son was di ec ly in es iga ed by Myllyne a and Hie anen (2015). They ma- nipula ed he belie o being seen by p esen ing pa icipan s wi h a iew o ano he li e pe son's di ec and a e ed gaze ei he wi h o wi hou an alleged one-way mi o in be ween hem. Au onomic a ousal esponses (SCRs) as well as psy- chophysiological o ien ing esponses (hea a e decele a ion and on al P3 ERPs) we e enhanced o di ec gaze only i pa icipan s belie ed ha he o he pe son could see hem. Co obo a ing his inding, a ecen s udy showed ha p o- longed eye con ac ele a ed pa icipan s’ skin conduc ance le el only when i was bidi ec ional, ha is, when a pe son was bo h seeing and being seen by ano he pe son (Ja ick & Bencic,2019). These esul s p o ide conside able e idence ha a bidi ec ional iew be ween he wo people is an es- sen ial p e equisi e o many o he a ec i e and cogni i e e ec s o eye con ac . Impo an ly, howe e , all o he a o emen ioned s udies employed s udy designs whe e he o he pe son was p esen in he same oom. This wa an s he ques ion o whe he he physical p esence o he o he pe son is also equi ed o he e ec s. An MRI s udy by Redcay and colleagues (2010) sugges s ha social in e ac ion, and no he physical p esence o he o he pe son, is he c ucial a iable o he elici a ion o a leas some o he neu ocogni i e e ec s caused by he pe cep ion o di ec gaze. In hei s udy, pa icipan s ei he engaged in an in e ac ion wi h an expe imen e o e a ideo call o wa ched a eco ding o he same in e ac ion. Du ing he ideo call, bu no while wa ching he ideo, enhanced ac i a ion was obse ed in b ain a eas ela ed o social cog- ni ion ( igh empo opa ie al junc ion and igh supe io em- po al sulcus) and ewa d p ocessing ( en al s ia um and igh amygdala). Howe e , because hese condi ions we e no compa ed o a li e in e ac ion wi h ano he pe son p esen in he same physical space, i is unclea whe he he e ec s o a ideo call a e equal o hose elici ed by a na u al, li e en- coun e wi h he o he pe son. To ou knowledge, no p e ious s udy has di ec ly in es iga ed wha ole he o he pe son's physical p esence plays on he e ec s ha eye con ac e okes. In he p esen s udy, we in es iga ed whe he he physical p esence o he o he pe son is necessa y o he psychophysi- ological esponses o eye con ac . Fo his pu pose, he e ec s o li e eye con ac , eye con ac o e a bidi ec ional ideo call, and wa ching a me e ideo o he o he pe son we e com- pa ed. In each o he h ee condi ions, esponses o di ec gaze we e compa ed o hose elici ed by gaze a e ed o he side. We measu ed pa icipan s’ au onomic a ousal wi h skin conduc ance and a ec i e acial esponses wi h acial EMG. | 3 o 12 HIETANEN ET Al. I he physical p esence o he o he pe son is necessa y o he psychophysiological e ec s, only li e eye con ac , and no eye con ac o e a ideo call, will elici g ea e au onomic a ousal esponses and acial muscle esponses associa ed wi h posi i e a ec han a e ed gaze wi hin he same condi ion. I , howe e , he physical p esence o he o he pe son is no equi ed and being seen by he o he pe son is he only essen- ial p e equisi e o he wo, hese esponses will be elici ed by bo h li e eye con ac and eye con ac o e a ideo call. In ei- he case he pe cei ed gaze di ec ion is no expec ed o ha e an in luence on hese esponses when wa ching a me e ideo o he o he pe son. We also measu ed pa icipan s’ subjec- i e eelings o a ousal and alence o in es iga e whe he awa eness o being obse ed o he o he pe son's p esence a ec hese e alua ions. 2 | METHOD 2.1 | Pa icipan s The pa icipan s we e 32 adul s (16 women, 16 men, Mage=27.8yea s, SDage=5.3, age ange=20–42yea s) e- c ui ed om email lis s o Uni e si y o Tampe e and a local Facebook g oup. Sensi i i y analysis (1–β = .80, α = .05, d n = 2, d dn = 62, es ima ed co ela ion = .5, ε = 1) wi h G*Powe (Faul, E d elde , Lang, & Buchne , 2007) indica ed ha , wi h his sample size, a 2 × 3 ANOVA can de ec medium in e - ac ions (η2 = .05, e ec size speci ica ion as in GPowe 3.0) be ween ac o s. This is he c ucial s a is ical es o de e min- ing whe he esponses o he gaze di ec ions di e be ween condi ions. The pa icipan s we e all na i e speake s o Finnish wi h no epo ed his o y o neu ological o psychia ic diso de s. They all had no mal o co ec ed- o-no mal ision. Pa icipan s we e ewa ded wi h a mo ie icke o cou se c edi . The s udy was ca ied ou in acco dance wi h The Code o E hics o he Wo ld Medical Associa ion (Decla a ion o Helsinki) and wi h a p o ocol e iewed by he E hics Commi ee o he Tampe e egion. All pa icipan s ga e hei w i en in o med consen . 2.2 | S imuli The expe imen consis ed o h ee condi ions: a li e in e ac- ion, a ideo call on a compu e , and wa ching a me e ideo on a compu e sc een. In all condi ions, he s imulus was he ace o a model pe son o he same sex as he pa icipan p e- sen ed agains a black backg ound. Two esea ch assis an s (one o each sex) se ed as models. The models we e bo h na i e Finnish speake s. The emale model was 26yea s old ha ing Nica aguan and Finnish ances y, and he male model was 22yea s old and o Zambian and Finnish ances y. The models had a neu al exp ession and hey main ained hei aces as mo ionless as possible. Depending on he ial, hey di ec ed hei gaze ei he s aigh ahead o app oxima ely 30° o he le o o he igh . The e we e no o he expe i- men al condi ions and no o he s imuli we e p esen ed o he pa icipan s. In all condi ions, he s imuli (a li e ace o he same ace appea ing on a compu e sc een) we e p esen ed h ough a ol age-sensi i e, liquid-c ys al shu e window (NSG UMU P oduc s Co., L d.) measu ing 21.5cm×38cm. The shu e window was a ached o a black panel ha was placed on a black able. The e was also a compu e sc een and a keyboa d on pa icipan s’ side o he able o espond- ing o ques ionnai es. Pa icipan s we e sea ed a a dis ance o 80cm om he shu e window. The s a e o he shu e window ( anspa en o opaque) was con olled wi h E-P ime 2.0 so wa e (Psychology So wa e Tools, Pi sbu gh, PA) on a desk op compu e . The shu e window swi ched be ween opaque and anspa en s a es in 3 milliseconds. Fo an illus- a ion o pa icipan s’ iew in li e and ideo call condi ions, see Figu e1. The s imulus in he li e condi ion was he ace o he model pe son on he o he side o he shu e window a a dis ance o 60cm om i . In he ideo call and ideo condi ions, he s imuli we e p esen ed on a 19-inch compu e sc een placed a a dis ance o 28 cm behind he shu e window. Du ing hese condi ions, he model pe son was in ano he oom si - ing in on o a compu e sc een and a web came a ha ansmi ed his o he image in eal ime o he pa icipan s’ FIGURE 1 Pa icipan s’ iew o he model pe son's di ec gaze in he li e (le ) and ideo call ( igh ) condi ions. In he me e ideo condi ion, he iew on he sc een was iden ical o ha in he ideo call condi ion. The compu e sc een ( o ques ionnai es) is isible in he o e on 4 o 12 | HIETANEN ET Al. sc een. See Figu e2 o an illus a ion o he se up in each condi ion. In all condi ions, pa icipan 's iew o he size and loca ion o he s imuli was con olled. The wid h o he ace s imulus in pa icipan 's iew was app oxima ely 7°. In he beginning o each condi ion, he model pe son's sea was adjus ed so ha his o he ace was di ec ly in on o he pa icipan and a he same heigh as he pa icipan 's ace. In he ideo call and ideo condi ions, an imp ession o eye con- ac was achie ed by placing he web came a and compu e sc een on he model pe son's side a enough om he model pe son so ha i appea ed as i he o she was looking di ec ly a he came a e en hough, du ing sea adjus men , hemodel pe son was looking a he sc een benea h i . Du ing s imulus p esen a ions, he model pe son was always looking ei he di ec ly in he eye egion o he pa icipan (li e condi ion) o di ec ly in o he came a ( ideocall and ideo condi ions). G ea ca e was aken o ensu e ha he s imuli we e isu- ally compa able in he h ee condi ions and ha he di ec gaze looked di ec in all condi ions. In he ideo call condi- ion, a web came a was on op o he compu e sc een so ha he model pe son was able o see he pa icipan , which he pa icipan was awa e o . In he beginning o he me e ideo condi ion, he web came a was on op o he sc een o sea adjus men , a e which i was emo ed, and he disconnec ed cable was shown o he pa icipan . In hese wo condi ions, audio in e ac ion be ween he pa icipan and he model pe - son was enabled wi h he web came a mic ophone, and hus, audio in e ac ion was only possible when he web came a was connec ed. Impo an ly, a e sea adjus men , no audio in e ac ion ook place in any condi ions. The shu e window was opaque a all imes excep du ing sea adjus men and s imulus p esen a ions. 2.3 | P ocedu e A e a i ing a he labo a o y, he pa icipan was in o med ha in he expe imen physiological measu emen s and ques- ionnai e da a would be collec ed du ing a simple in e ac ion si ua ion. A e his, he expe imen e ob ained he pa - icipan 's in o med consen . The model pe son was al eady p esen in he labo a o y, bu he o she did no gi e any in- s uc ions. The measu emen senso s we e hen a ached o he pa icipan 's le hand and ace. To conceal he pu pose o he acial senso s, he expe imen e old ha he senso s we e used o measu e skin empe a u e. EEG was also measu ed o in es iga e esea ch ques ions no ela ed o his a icle. The expe imen consis ed o h ee condi ions ha we e p esen ed in a coun e balanced o de . In he li e condi ion, he s imulus was a li e model pe son's ace on he o he side o he shu e window. In he ideo call condi ion, a compu e sc een wi h a simila iew o he model pe son's ace was on he o he side o he shu e window. A web came a was on op o he sc een o enable a bidi ec ional iew be ween he wo. The ideo condi ion was o he wise simila o he ideo call condi- ion, excep ha he e was no web came a, and hus, he pa ici- pan s me ely wa ched he ace s imuli wi hou being hemsel es seen by he model pe son. Each condi ion began wi h he model pe son adjus ing his o he sea and con i ming ha he pa ici- pan also pe cei ed hei eyes o be on he same le el. FIGURE 2 An illus a ion o he expe imen al se up in he h ee condi ions. In he ideo call and ideo condi ions, he model pe son was in ano he oom and his o he image was po ayed in eal ime wi h a zoomed-in web came a. The dis ances a e no in exac scale Li e Video Call Video Pa icipan Pa icipan Pa icipan Model Shu e window Model Model | 5 o 12 HIETANEN ET Al. Each condi ion began wi h 16 ials o s imulus p esen a- ion o 5s a a ime du ing which he physiological esponses we e measu ed. On eigh ials, he model pe son di ec ed his o he gaze s aigh ahead (di ec gaze), and on eigh ials, o ei he side (a e ed gaze, ou ials o each side). The ials we e p esen ed in a pseudo andomized o de wi h no mo e han wo consecu i e ials wi h he same gaze di ec ion. A e each ial, he e was an in e s imulus in e al (ISI) o a leas 14.5s du ing which he shu e window was opaque. A e he expe imen e had con i med ha he pa icipan 's skin conduc ance had e u ned o he p es imulus le el, he nex ial was s a ed. In each condi ion, physiological measu emen s we e ollowed by ques ionnai es. Pa icipan s comple ed wo ques ionnai e asks ega ding sel - e e en ial p ocessing and sel -awa eness. Toge he hese wo asks consis ed o 13 ials pe condi ion, each wi h a 5-s p esen a ion o di- ec gaze s imulus. Fo conciseness and because hese mea- su es we e no di ec ly ela ed o he esea ch ques ion o he p esen s udy, hey a e epo ed in Appendix A in he online Suppo ing In o ma ion. Subjec i e alence and a ousal e alua ions we e assessed wi h he Sel -Assessmen Manikin (B adley & Lang, 1994) in esponse o di ec and a e ed gaze in all condi ions. Pa icipan s i s ead ins uc ions on he compu e sc een elling hem ha soon hey will see he p esen a ions again and, du ing each p esen a ion, he ask is o e alua e hei le el o a ousal and alence. The ins uc ions in o med hem abou he meaning o a ousal and alence and abou he used scale. They we e p esen ed wi h ei he di ec o a e ed gaze o 5s, and hen, asked o e alua e hei subjec i e eelings on he wo dimensions on a 9-poin scale (1=calm/unpleasan , 9=a ousing/pleasan ). They we e hen simila ly p esen ed wi h he o he gaze di ec ion. The o de o he gaze di ec ions was pseudo andomized. A e he h ee condi ions, he measu emen senso s we e emo ed. Pa icipan s we e hen adminis e ed a ques ionnai e un ela ed o he p esen s udy, deb ie ed, hanked, and gi en he pa icipa ion ewa ds. No o he da a we e collec ed in he expe imen . 2.4 | Acquisi ion and analysis o physiological da a 2.4.1 | Skin conduc ance Skin conduc ance was measu ed wi h wo elec odes (Ag/ AgCl) coa ed wi h iso onic pas e and a ached o he pal- ma su ace o he medial phalanges o he index and mid- dle inge s on he pa icipan 's le hand. SCRs and EMG we e eco ded using a B ainVision QuickAmp ampli ie and B ainVision Reco de so wa e (B ain P oduc s GmbH, Munich, Ge many) a a sampling a e o 1,000Hz. O line, he SCR da a we e esampled o 100 Hz and il e ed wi h a 10Hz low-pass il e . Response was de ined as he maximum, con inuous inc ease in ampli ude wi hin a ime ame o 0.9–6.5s a e s imulus onse . The esponse was calcula ed by sub ac ing he p eceding minimum am- pli ude om he hus ound maximum ampli ude. The 0.9– 6.5-s analysis window was chosen based on he ime cou se o SCRs (a 0.9–3.5 s la ency o ini ia ion [Sjouwe man & Lonsdo , 2019] and a 1–3 s ise ime o peak [Dawson, Schell, & Filion,2000]). A ial was coded as a ze o esponse i he maximum inc ease in ampli ude was less han 0.01 µS, i he inc ease ini ia ed la e han 3.5 s a e he s im- ulus onse , o i he inc ease was s eady indica ing a onic change in skin conduc ance le el. I he e was an ampli ude inc ease o 0.01 µS o mo e wi hin a ime ame o −1 o 0.9s om he s imulus onse indica ing a p ema u e esponse un ela ed o he s imulus, he ial was ejec ed (4.8% o all ials). Based on isual inspec ion, ials wi h excessi e a i- ac s we e ejec ed (1.4% o all ials). Fo each pa icipan and o each condi ion, mean SCR magni ude was calcula ed by a e aging he da a om all accep ed ials including hose wi h ze o esponses (mean numbe o ials in each condi ion: Mli e=15.1, M ideocall=14.8, M ideo=15.1). The mean SCR magni udes we e used in he s a is ical analyses. 2.4.2 | Facial muscle ac i i y Facial muscle ac i i y was measu ed o e he zygoma icus majo and co uga o supe cilii muscle egions. The skin o e he eco ding si es was cleaned and sligh ly ab aded wi h al- cohol. Fo bipola measu emen , wo 4mm elec odes (Ag/ AgCl) coa ed wi h elec ode pas e we e a ached 1cm apa o he eco ding si es acco ding o he guidelines by F idlund and Cacioppo (1986). O line, he signal was il e ed wi h a 10–499-Hz band- pass il e and a 50-Hz no ch il e and ec i ied. Based on isual inspec ion, ials wi h a i ac s due o excessi e mus- cle mo emen s and blinks we e ejec ed (6.9% o all ials). One pa icipan had less han 50% a i ac - ee epochs in one condi ion and she was excluded om he analysis. The signal was hen segmen ed in o 500-ms epochs om 500ms p io o s imulus onse (baseline) o 5,000ms a e s imulus onse . The alues we e s anda dized wi hin each pa icipan and muscle si e. Change sco es we e calcula ed by sub ac ing he baseline muscle ac i i y om he a e age alue o each 500- ms epoch and a e aged ac oss all accep ed ials wi hin each expe imen al condi ion (mean numbe o ials in each con- di ion: Mli e=14.8, M ideocall=15.1, M ideo=14.8). These change sco es we e used in he s a is ical analyses. 6 o 12 | HIETANEN ET Al. 2.5 | S a is ical analyses Wi hin-subjec s analyses o a iance (ANOVA) wi h condi- ion (li e s. ideo call s. ideo) and gaze di ec ion (di ec s. a e ed) as wi hin-subjec s ac o s we e used o compa e he psychophysiological and subjec i e esponses. In he EMG analysis, ime (10 epochs, each las ing 500ms) was used as a hi d wi hin-subjec s ac o , because he ime cou se o acial EMG esponses can ange om b ie spikes o many seconds depending on he unde lying emo ional p ocesses (Cacioppo, Ma zke, Pe y, & Tassina y,1988). When he as- sump ion o sphe ici y was iola ed, deg ees o eedom we e co ec ed using G eenhouse–Geisse es ima es o sphe ici y. When a signi ican in e ac ion be ween condi ion and gaze di ec ion was obse ed, pai ed samples es s we e used o compa e he esponses o di ec and a e ed gaze wi hin each condi ion. Nonsigni ican esul s om he es s we e u he explo ed wi h he Two One-Sided Tes (TOST) p ocedu e wi h equi alence bounds se a d = ±0.30 (Lakens,2017). Fo conciseness, addi ional analyses ela ed o nonsigni ican indings, nonpa ame ic es s, and subjec i e a ings a e e- po ed in Appendix B in he online Suppo ing In o ma ion. 3 | RESULTS 3.1 | Skin conduc ance A 2×3 (Gaze Di ec ion×Condi ion) wi hin-subjec s ANOVA indica ed a main e ec o gaze di ec ion, F(1, 31) = 4.73, p = .037, 𝜂2 p = .13. The SCR esponses we e g ea e in esponse o di ec gaze (M=0.034 µS, SD=0.045, 95% CI [0.018, 0.050]) han o a e ed gaze (M=0.021 µS, SD=0.027, 95% CI [0.012, 0.031]). The main e ec o condi ion, F(2, 62) = 0.93, p = .399, 𝜂2 p = .03, was no signi ican . Impo an ly, he e was a s a is ically signi ican in e ac ion be ween gaze di ec ion and condi ion, F(2, 62) = 4.49, p = .015, 𝜂2 p = .13. When analyzing he h ee condi ions sepa a ely, pai ed es s indica ed g ea e SCRs o di ec e sus a e ed gaze in he li e condi ion, (31) = 2.57, p = .015, d=0.45, and in he ideo call condi ion, (31) = 2.15, p = .040, d=0.38, bu no in he me e ideo condi ion, (31) = −0.69, p = .494, d = −0.12. A TOST p ocedu e indica ed ha he obse ed e - ec size was signi ican ly wi hin he uppe bound o d=0.30, (31) = 2.36, p = .012, bu no wi hin he lowe bound o d = −0.30, (31) = 1.03, p = .154. The e o e, in he ideo condi ion, we can conclude ha he pe cep ion o di ec gaze did no elici a meaning ul inc ease in SCRs in compa ison o a e ed gaze, bu i is possible ha i may ha e dec eased he esponses. See Figu e3 o SCR magni udes in esponse o bo h gaze di ec ions wi hin each condi ion. In o de o compa e he magni ude o SCR inc ease by di- ec gaze in he li e and ideo call condi ions, a 2×2 (Gaze Di ec ion × Condi ion) wi hin-subjec s ANOVA was con- duc ed. The main e ec o gaze di ec ion, F(1, 31) = 8.22, p = .007, 𝜂2 p = .21, was signi ican , whe eas he main e ec o condi ion, F(1, 31) = 1.15, p = .292, 𝜂2 p = .04, o he in e ac- ion be ween he wo, F(1, 31) = 1.35, p = .255, 𝜂2 p = .04, we e no . Di e en ial esponses we e calcula ed by sub ac ing he SCRs in esponse o a e ed gaze om hose o di ec gaze wi hin hese condi ions. Al hough no di e ence was obse ed in he magni ude o SCR inc ease by di ec gaze be ween he li e and ideo call condi ions, (31) = 1.16, p = .255, d=0.21, a TOST p ocedu e wi h equi alence bounds o d = ±0.30 did no suppo he absence o a meaning ul di e ence be ween he condi ions, (31) = 0.54, p = .298. 3.2 | Facial muscle ac i i y Zygoma ic EMG esponses we e analyzed wi h a 2× 3 × 10 (Gaze Di ec ion × Condi ion × Time) wi hin-subjec s ANOVA. S a is ically signi ican main e ec s we e ound o gaze di ec ion, F(1, 30) = 18.34, p < .001, 𝜂2 p = .38, and FIGURE 3 SCR magni udes and s anda d e o s o he mean (SEM) in esponse o di ec and a e ed gaze wi hin each condi ion. *p < .05, wo- ailed es | 7 o 12 HIETANEN ET Al. ime, F(3.99, 119.73) = 4.73, p = .001, 𝜂2 p = .14 (G eenhouse– Geisse co ec ed, ε=0.44), bu no o condi ion, F(2, 60) = 0.44, p = .649, 𝜂2 p = .01. The zygoma ic esponses we e g ea e in esponse o di ec gaze (M=0.54, SD=0.56, 95% CI [0.34, 0.74]) han o a e ed gaze (M=0.21, SD=0.40, 95% CI [0.07, 0.35]), and he zygoma ic ac i i y inc eased as a unc ion o ime. Impo an ly, howe e , con a y o wha was expec ed, he e was no signi ican in e ac ion be ween condi ion and gaze di ec ion, F(2, 60) = 0.30, p = .739, 𝜂2 p = .01. No o he signi ican in e ac ion e ec s we e ound ei he (all ps > .10). Co uga o egion EMG esponses we e analyzed like- wise wi h a 2×3 × 10 wi hin-subjec s ANOVA. The main e ec o gaze di ec ion was signi ican , F(1, 30) = 8.00, p = .008, 𝜂2 p = .21; co uga o esponses we e smalle in e- sponse o di ec gaze (M = −0.10, SD=0.71, 95% CI [−0.35, 0.15]) han o a e ed gaze (M=0.15, SD=0.48, 95% CI [−0.02, 0.31]). The main e ec s o condi ion, F(2, 60) = 1.15, p = .324, 𝜂2 p = .04, and ime, F(1.76, 52.69) = 1.22, p = .300, 𝜂2 p = .04 (G eenhouse–Geisse co ec ed, ε=0.20), we e no signi ican . The e was a signi ican wo-way in e ac- ion be ween gaze di ec ion and ime, F(2.90, 87.03) = 4.08, p = .010, 𝜂2 p = .12 (G eenhouse–Geisse co ec ed, ε=0.32). A close look a he in e ac ion e ealed ha he e ec o ime was no signi ican in he di ec gaze ials, F(1.73, 51.79) = 2.27, p = .120, 𝜂2 p = .07 (G eenhouse–Geisse co ec ed, ε=0.19), o in he a e ed gaze ials, F(3.36, 100.85) = 2.12, p = .096, 𝜂2 p = .07 (G eenhouse–Geisse co ec ed, ε=0.37). A ma ginal wo-way in e ac ion was ound be ween gaze di- ec ion and condi ion, F(2, 60) = 3.05, p = .055, 𝜂2 p = .09. No o he signi ican in e ac ion e ec s we e ound (all ps > .10). See Figu e4 o he EMG esponses o bo h gaze di ec ions wi hin each condi ion. 3.3 | Subjec i e alence and a ousal Subjec i e a ings o a ec i e alence and a ousal we e measu ed wi h he Sel -Assessmen Manikin in esponse o bo h gaze di ec ions in all h ee condi ions. 2 × 3 wi hin- subjec s ANOVAs indica ed no signi ican main e ec s o in e ac ion e ec s o ei he scale (all ps > .10). See Table1 o he alence and a ousal a ings in esponse o bo h gaze di ec ions wi hin each condi ion, and Appendix B o mo e de ailed s a is ical in o ma ion. 4 | DISCUSSION The p esen s udy in es iga ed he oles o being seen by an- o he pe son and his o he physical p esence in he psycho- physiological esponses o eye con ac . Fo his pu pose, we measu ed au onomic a ousal wi h SCRs and acial EMG e- sponses associa ed wi h posi i e a ec o di ec e sus a e ed gaze in h ee di e en condi ions: one wi h a li e in e ac ion, ano he wi h a ideo call in e ac ion, and a hi d condi ion o me ely wa ching a ideo o ano he pe son. Di ec gaze was ound o elici g ea e SCRs indica ing g ea e au onomic a ousal han a e ed gaze in li e in e ac ion and in ideo call, bu no such e ec was obse ed in he me e ideo condi ion. The e ec was o simila magni ude in he li e and ideo call condi ions. Rega ding acial eac ions, howe e , di ec gaze elici ed mo e posi i e a ec i e acial eac ions—g ea e zy- goma ic and smalle co uga o EMG ac i i y— han a e ed gaze in all h ee condi ions. No signi ican di e ences in sub- jec i e a ings o alence and a ousal we e ound be ween gaze di ec ions in hese h ee condi ions. Pa icipan s’ a ousal esponses o di ec gaze we e el- e a ed only in he li e and ideo call condi ions. While me ely wa ching a ideo o ano he pe son, he pe cei ed gaze di ec ion did no a ec he pa icipan s’ au onomic a ousal. We, he e o e, eplica ed p e ious indings o heigh ened au onomic a ousal in li e eye con ac (e.g., Hie anen e al.,2008; Ja ick & Bencic,2019; Nichols & Champness,1971; Pönkänen, Pel ola, e al.,2011; P insen & Alae s, 2019) and no inc ease in a ousal in esponse o di ec e sus a e ed gaze p esen ed in ideo (Lea i & Dono an,1979; Lyy ae al., 2018; Wiese e al.,2009). As being seen by he o he pe son was he common de- nomina o be ween he li e and ideo call condi ions, he esul s also con e ge wi h he indings o Myllyne a and Hie anen (2015) whe ein di ec gaze was shown o elici g ea e a ousal han a e ed gaze only when pa icipan s belie ed o be seen by he o he pe son. The li e and ideo call condi ions also di e om he me e ideo condi ion in he possibili y o sub le non e - bal in e ac ion. Al hough he models we e ins uc ed o s ay mo ionless and exp essionless and no o ecip oca e wi h he pa icipan s e bally o non e bally, in hese wo condi- ions, limi ed in e ac ional con ingencies, such as he model blinking as a eac ion o some hing he pa icipan did, we e possible. Thus, i is possible ha sub le non e bal eac ions could also explain he au onomic a ousal esponses o di ec e sus a e ed gaze in he li e and ideo call condi ions. Howe e , his does no seem likely, because in he s udy by Myllyne a and Hie anen (2015) only pa icipan s’ be- lie o being seen was manipula ed and he models we e, in ac , always able o see and, hus, eac o, he pa icipan s’ ac ions, and ye , he au onomic esponses we e g ea e o di ec han a e ed gaze only when pa icipan s belie ed o be seen. The e o e, he mos p obable explana ion o he p esen and p e ious esul s is ha he au onomic a ousal esponse o di ec gaze is elici ed by an expe ience o being seen by ano he indi idual. 8 o 12 | HIETANEN ET Al. FIGURE 4 S anda dized mean zygoma ic and co uga o EMG esponses (and SEM) in esponse o di ec and a e ed gaze wi hin each condi ion | 9 o 12 HIETANEN ET Al. Mos impo an ly, he p esen s udy ex ends he knowl- edge o wha unde lies he eye con ac e ec s. P e ious s ud- ies ha e obse ed his inc ease in a ousal only in esponse o he gaze o ano he pe son p esen in he same oom, and he e o e, i has been unclea whe he o no he o he pe son's physical p esence is also equi ed o he e ec (Helminen, Kaasinen, & Hie anen, 2011; Hie anen e  al., 2008, 2018; Ja ick & Bencic, 2019; Myllyne a & Hie anen, 2015, 2016; Myllyne a, Ran a, & Hie anen, 2015; Nichols & Champness,1971; Pönkänen, Pel ola, e al.,2011). By show- ing ha he pe cep ion o di ec gaze elici s a g ea e a ousal esponse han a e ed gaze also in a ideo call in e ac ion, he p esen s udy is he i s o demons a e ha ano he pe son's physical p esence is no equi ed o he e ec . The p esen esul s ha e impo ance also o social cogni- ion esea ch mo e gene ally. In s udies o social cogni ion, pa icipan s a e o en p esen ed wi h pic u es o ideos o o he people in o de o achie e high con ollabili y o s im- ulus p esen a ion. Impo an ly, howe e , pic o ial s imuli do no allow any ac ual in e ac ion be ween he obse e and he s imulus pe sons. In ecen yea s, he use o such spec a o ial se ings has been c i icized o his limi a ion (Risko, Laidlaw, F ee h, Foulsham, & Kings one,2012; Risko, Richa dson, & Kings one,2016). Schilbach and colleagues (2013), o ex- ample, a gued o an app oach o he in es iga ion o social cogni ion ha is based on in e ac ion a he han shee obse - a ion. Ou indings suppo such a second-pe son app oach by demons a ing ha he pe cep ion o ano he pe son on a compu e sc een does no ha e he same e ec on au onomic ne ous sys em i he pe cep ion is no equipped wi h a bidi- ec ional exchange o gaze wi h ano he conscious mind. The acial EMG measu emen s indica ed ha he pe cep- ion o di ec gaze elici ed an inc ease in zygoma ic and de- c ease in co uga o muscle ac i i y as compa ed o a e ed gaze. This e ec was ound in all condi ions ega dless o whe he he seen ace was ha o a li e pe son o one on a compu e sc een and whe he he o he pe son was able o see he pa icipan o no . Simila acial EMG esponses o a li e pe son's di ec e sus a e ed gaze ha e also been obse ed be o e (Hie anen e al.,2018), bu as a as we know, his is he i s s udy o compa e he a ec i e acial esponses o di e en p esen a ion o ms o di ec and a e ed gaze s im- uli. The p esen esul s indica e ha , con a y o he a ousal esponses, he acial eac ions a e elici ed by he me e pe - cep ion o di ec gaze and hey a e no mode a ed by he ex- pe ience o being seen. The disc epancy be ween au onomic a ousal and acial eac ions impo an ly implies ha he ac i a ion o he zy- goma ic muscle and he elaxa ion o he co uga o muscle do no necessa ily e lec a genuine a ec i e expe ience. These EMG esponses a e associa ed wi h posi i e emo ional eac ions (e.g., Dimbe g, Thunbe g, & G unedal,2002), bu hey may also e lec communica ion o social mo i es (e.g., Pa kinson,2005). Hie anen and colleagues (2018) p oposed ha enhanced zygoma ic ac i i y o ano he pe son's di ec gaze may be caused by a highly au oma ized a ilia i e e- sponse, a lea ned eac ion o smiling o he pe cep ion o a ace looking you way. The p esen esul s lend u he em- pi ical suppo o hei p oposal because he acial eac ions we e dissocia ed om emo ional a ousal and because hey we e also obse ed in a con ex whe e bidi ec ional commu- nica ion was no possible. Subjec i e e alua ions o alence and a ousal we e no ound o di e be ween he gaze di ec ions o condi ions. This was unexpec ed and in con as o he psychophysiological measu emen s, al hough simila inconsis encies ha e been obse ed be o e (e.g., Roseb ock, Hoxha, No is, Cacioppo, & Gollan,2016). The p esen esul s also con adic p e i- ous indings om ou own labo a o y ha ha e ound a small bu signi ican gaze di ec ion e ec on bo h scales (Hie anen e al.,2008; Pönkänen, Alhoniemi, e al.,2011). One possible explana ion o his disc epancy is ha , a e he block du ing which he physiological esponses we e eco ded, pa icipan s we e ins uc ed o assess hei emo ional expe ience du ing he subsequen p esen a ions, and hen, p esen ed wi h he aces again. In he p e ious s udies, pa icipan s we e asked o espond jus based on ecall o how hey el while seeing he aces du ing he p eceding s imulus block. I is plausible ha an online e alua ion o alence and a ousal du ing s imulus p esen a ions leads o di e en esponses han an e alua ion based on ecall o p e ious emo ional s a es. Ano he possi- ble explana ion s ems om he la ge numbe o ials ha p e- ceded he e alua ion. In he p esen s udy, pa icipan s we e p esen ed wi h he same s imulus ace a much la ge numbe o imes (e en as many as 91 imes in he las condi ion) be- o e he e alua ions, which could ha e caused habi ua ion o TABLE 1 Mean sco es and s anda d de ia ions on he Sel -Assessmen Manikin by gaze di ec ion and condi ion  Li e Video call Video Di ec A e ed Di ec A e ed Di ec A e ed Scale M SD M SD M SD M SD M SD M SD Valence 6.28 1.49 6.09 1.53 6.22 1.48 5.72 1.51 6.16 1.74 5.88 1.72 A ousal 2.63 1.60 2.38 1.19 2.38 1.41 2.50 1.11 2.44 1.41 2.44 1.39 No e: 95% con idence in e als o he mean alues a e p esen ed in Appendix B.