10.07.2015 Views

Dogs do look at images: eye tracking in canine cognition ... - Springer

Dogs do look at images: eye tracking in canine cognition ... - Springer

Dogs do look at images: eye tracking in canine cognition ... - Springer

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Anim Cogn (2012) 15:163–174DOI 10.1007/s10071-011-0442-1ORIGINAL PAPER<strong>Dogs</strong> <strong>do</strong> <strong>look</strong> <strong>at</strong> <strong>images</strong>: <strong>eye</strong> <strong>track<strong>in</strong>g</strong> <strong>in</strong> can<strong>in</strong>e <strong>cognition</strong> researchSanni Somppi • He<strong>in</strong>i Törnqvist • Laura Hänn<strong>in</strong>en •Christ<strong>in</strong>a Krause • Outi Va<strong>in</strong>ioReceived: 11 February 2011 / Revised: 6 July 2011 / Accepted: 1 August 2011 / Published onl<strong>in</strong>e: 23 August 2011Ó Spr<strong>in</strong>ger-Verlag 2011Abstract Despite <strong>in</strong>tense research on the visual communic<strong>at</strong>ionof <strong>do</strong>mestic <strong>do</strong>gs, their cognitive capacitieshave not yet been explored by <strong>eye</strong> <strong>track<strong>in</strong>g</strong>. The aim of thecurrent study was to expand knowledge on the visual<strong>cognition</strong> of <strong>do</strong>gs us<strong>in</strong>g contact-free <strong>eye</strong> movement <strong>track<strong>in</strong>g</strong>under conditions where social cue<strong>in</strong>g and associ<strong>at</strong>ivelearn<strong>in</strong>g were ruled out. We exam<strong>in</strong>ed whether <strong>do</strong>gsspontaneously <strong>look</strong> <strong>at</strong> actual objects with<strong>in</strong> pictures andcan differenti<strong>at</strong>e between pictures accord<strong>in</strong>g to their noveltyor c<strong>at</strong>egorical <strong>in</strong>form<strong>at</strong>ion content. Eye movements ofsix <strong>do</strong>mestic <strong>do</strong>gs were tracked dur<strong>in</strong>g present<strong>at</strong>ion ofdigital color <strong>images</strong> of human faces, <strong>do</strong>g faces, toys, andalphabetic characters. We found th<strong>at</strong> <strong>do</strong>gs focused their<strong>at</strong>tention on the <strong>in</strong>form<strong>at</strong>ive regions of the <strong>images</strong> withoutany task-specific pre-tra<strong>in</strong><strong>in</strong>g and their gaz<strong>in</strong>g behaviordepended on the image c<strong>at</strong>egory. <strong>Dogs</strong> preferred the facial<strong>images</strong> of conspecifics over other c<strong>at</strong>egories and fix<strong>at</strong>ed ona familiar image longer than on novel stimuli regardless ofthe c<strong>at</strong>egory. <strong>Dogs</strong>’ <strong>at</strong>traction to conspecifics over humanfaces and <strong>in</strong>anim<strong>at</strong>e objects might reflect their n<strong>at</strong>ural<strong>in</strong>terest, but further studies are needed to establish whether<strong>do</strong>gs possess picture object re<strong>cognition</strong>. Contact-free <strong>eye</strong>movement <strong>track<strong>in</strong>g</strong> is a promis<strong>in</strong>g method for the broaderexplor<strong>at</strong>ion of processes underly<strong>in</strong>g special socio-cognitiveskills <strong>in</strong> <strong>do</strong>gs previously found <strong>in</strong> behavioral studies.S. Somppi (&) L. Hänn<strong>in</strong>en O. Va<strong>in</strong>ioFaculty of Veter<strong>in</strong>ary Medic<strong>in</strong>e, University of Hels<strong>in</strong>ki,P. O. Box 57, 00014 Hels<strong>in</strong>ki, F<strong>in</strong>lande-mail: sanni.somppi@hels<strong>in</strong>ki.fiH. Törnqvist C. KrauseFaculty of Behavioral Sciences, University of Hels<strong>in</strong>ki,P. O. Box 9, 00014 Hels<strong>in</strong>ki, F<strong>in</strong>landKeywords Can<strong>in</strong>e Domestic <strong>do</strong>gs Eye movement<strong>track<strong>in</strong>g</strong> Visual <strong>cognition</strong>IntroductionClose co-evolution, shar<strong>in</strong>g the same liv<strong>in</strong>g habit<strong>at</strong>, andsimilarities <strong>in</strong> can<strong>in</strong>e and human socio-cognitive skillsmake the <strong>do</strong>g a unique model for compar<strong>at</strong>ive <strong>cognition</strong>studies (Hare and Tomasello 2005; Miklósi et al. 2007;Topál et al. 2009). Domestic<strong>at</strong>ion has equipped <strong>do</strong>gs withsophistic<strong>at</strong>ed sensitivity to respond to human visual communic<strong>at</strong>ivecues, <strong>in</strong>clud<strong>in</strong>g gaze direction and po<strong>in</strong>t<strong>in</strong>ggestures (Hare and Tomasello 2005; Miklósi et al. 2003;Soproni et al. 2002; Virányi et al. 2004). Despite the<strong>in</strong>tense research around the visual communic<strong>at</strong>ion of<strong>do</strong>mestic <strong>do</strong>gs, their cognitive capacities have not yet beenexplored by <strong>eye</strong> <strong>track<strong>in</strong>g</strong>.Animals with developed visual systems control theirgaze direction us<strong>in</strong>g <strong>eye</strong> movements (Land 1999). Eyemovements are l<strong>in</strong>ked to <strong>at</strong>tention (Buswell 1935); thevisual-cognitive system directs a gaze toward importantand <strong>in</strong>form<strong>at</strong>ive objects, and vice versa, gaze directionaffects several cognitive processes (Henderson 2003).Based on behavioral observ<strong>at</strong>ions, <strong>do</strong>gs can perceive twodimensionalvisual <strong>in</strong>form<strong>at</strong>ion and can be tra<strong>in</strong>ed to performvisual tasks, such as classify<strong>in</strong>g photographs of n<strong>at</strong>uralstimuli by means of a perceptual response rule (Rangeet al. 2008). They can correl<strong>at</strong>e the photographs and voiceof their owner (Adachi et al. 2007) and are able to dist<strong>in</strong>guishthe facial <strong>images</strong> of two <strong>in</strong>dividual humans or <strong>do</strong>gs(Racca et al. 2010).However, <strong>in</strong> previous experiments, <strong>do</strong>gs have beeneither manually restra<strong>in</strong>ed (Adachi et al. 2007; Guo et al.2009; Racca et al. 2010) or their choices were re<strong>in</strong>forced123


164 Anim Cogn (2012) 15:163–174(Range et al. 2008), which may have affected the outcome.<strong>Dogs</strong> are highly sensitive and respond to humangestures and <strong>at</strong>tentional st<strong>at</strong>es (reviewed <strong>in</strong> Miklósi et al.2007; Topáletal.2009), so the owner or experimentermay even subconsciously <strong>in</strong>fluence the <strong>do</strong>g’s expect<strong>at</strong>ionsand behavior; the phenomenon called Clever Hanseffect (Pfungst 1907). In the previous studies, this effecthas been m<strong>in</strong>imized by avoid<strong>in</strong>g the human <strong>in</strong>terferencewith the <strong>do</strong>gs or leav<strong>in</strong>g the restra<strong>in</strong>er ignorant of thepurpose of the experiment (Adachi et al. 2007; Guo et al.2009; Racca et al. 2010). The Clever Hans effect mightbe particularly prevalent <strong>in</strong> <strong>do</strong>gs (Lit et al. 2011); thus,tasks th<strong>at</strong> are performed without human presence couldbe useful.In traditional visual discrim<strong>in</strong><strong>at</strong>ion tests the <strong>look</strong><strong>in</strong>gbehavior is measured manually from the orient<strong>at</strong>ionbehavior (Adachi et al. 2007; Guo et al. 2009; Racca et al.2010; Faragó et al. 2010), which may <strong>in</strong>clude both bouts ofactive <strong>in</strong>form<strong>at</strong>ion process<strong>in</strong>g and blank stares (Asl<strong>in</strong>2007). When only behavioral measures are used, it isimpossible to def<strong>in</strong>e which fe<strong>at</strong>ures the <strong>do</strong>gs’ <strong>at</strong>tention isdrawn to. Is the <strong>do</strong>g actually <strong>look</strong><strong>in</strong>g <strong>at</strong> the picture? Eyemovement <strong>track<strong>in</strong>g</strong>, a technique for directly assess<strong>in</strong>ggaz<strong>in</strong>g behavior, enables measurement of the sub-elementsof the <strong>look</strong><strong>in</strong>g behavior, which likely <strong>in</strong>dic<strong>at</strong>e the underly<strong>in</strong>gmechanisms of the performance better than traditionalparadigms (Qu<strong>in</strong>n et al. 2009).Eye <strong>track<strong>in</strong>g</strong> provides an objective method for understand<strong>in</strong>gthe ongo<strong>in</strong>g cognitive and emotional processes,such as visual-sp<strong>at</strong>ial <strong>at</strong>tention, semantic <strong>in</strong>form<strong>at</strong>ion process<strong>in</strong>g,and motiv<strong>at</strong>ional st<strong>at</strong>es of animals (Henderson2003; Kano and Tomonaga 2009). Eye movement <strong>track<strong>in</strong>g</strong><strong>in</strong> <strong>do</strong>gs has been used to diagnose ocular motor abnormalitiessuch as nystagmus (Jacobs et al. 2009; Dell’Ossoet al. 1998), and a head-mounted <strong>eye</strong>-<strong>track<strong>in</strong>g</strong> camera hasbeen recently tested on a s<strong>in</strong>gle <strong>do</strong>g (Williams et al. 2011),but to d<strong>at</strong>e, the method has not been utilized <strong>in</strong> can<strong>in</strong>ecognitive studies. Although <strong>do</strong>gs’ visual acuity is lowerthan th<strong>at</strong> of humans and they have compar<strong>at</strong>ively reducedcolor perception (dichrom<strong>at</strong>ic color vision), the visualsystem of the <strong>do</strong>g <strong>do</strong>es not impose crucial limit<strong>at</strong>ions onthe use of the <strong>eye</strong> movement <strong>track<strong>in</strong>g</strong> method. The can<strong>in</strong>eoptic chiasm has a crossover of about 75%, which allowsfor good b<strong>in</strong>ocular vision (Miller and Murphy 1995). <strong>Dogs</strong>can detect details, even <strong>in</strong> very small photographs presented<strong>at</strong> a short distance (Range et al. 2008).The spontaneous visual discrim<strong>in</strong><strong>at</strong>ion abilities of human<strong>in</strong>fants and non-human prim<strong>at</strong>es have traditionally beentested with novelty preference paradigms (Fantz 1964;Joseph et al. 2006; Dahl et al. 2007; Colombo and Micthell2009). In a commonly used protocol, a s<strong>in</strong>gle stimulus isfirst presented repe<strong>at</strong>edly (familiariz<strong>at</strong>ion) and then changedto a novel stimulus. If the subject discrim<strong>in</strong><strong>at</strong>es thesimilarity of the repe<strong>at</strong><strong>in</strong>g present<strong>at</strong>ion, there is a decl<strong>in</strong>e <strong>in</strong><strong>look</strong><strong>in</strong>g, and a rebound <strong>in</strong> <strong>look</strong><strong>in</strong>g occurs when the stimulusis shifted if the subject discrim<strong>in</strong><strong>at</strong>es a novel stimulus froma familiar one (Houston-Price and Nakai 2004). Recently, ithas been demonstr<strong>at</strong>ed th<strong>at</strong> <strong>do</strong>gs exhibit an orient<strong>in</strong>gbehavior toward novel pictures, but their preferences aredependent on image c<strong>at</strong>egory (Racca et al. 2010). Thisf<strong>in</strong>d<strong>in</strong>g <strong>in</strong>dic<strong>at</strong>es th<strong>at</strong> <strong>do</strong>gs may have the ability to c<strong>at</strong>egorizevisual objects. Form<strong>in</strong>g c<strong>at</strong>egories by sort<strong>in</strong>g objects<strong>in</strong>to associ<strong>at</strong>ive classes is one of the fundamental cognitiveabilities of humans, but little is known about spontaneousc<strong>at</strong>egoriz<strong>at</strong>ion <strong>in</strong> animals (Murai et al. 2005).The purpose of our study was to expand knowledge onthe cognitive abilities of <strong>do</strong>gs us<strong>in</strong>g a new method, contactfree<strong>eye</strong> movement <strong>track<strong>in</strong>g</strong>. Our aim was to assess how<strong>do</strong>gs view two-dimensional photographs under free view<strong>in</strong>gconditions where social cue<strong>in</strong>g is ruled out. To studythis, we tested whether <strong>do</strong>gs differenti<strong>at</strong>e between picturesaccord<strong>in</strong>g to their novel or c<strong>at</strong>egorical <strong>in</strong>form<strong>at</strong>ion contentwithout any task-specific pre-tra<strong>in</strong><strong>in</strong>g.M<strong>at</strong>erials and methodsAll experimental procedures were approved by the EthicalCommittee for the use of animals <strong>in</strong> experiments <strong>at</strong> theUniversity of Hels<strong>in</strong>ki.Animals and experimental setupSix priv<strong>at</strong>ely owned <strong>do</strong>gs, aged 1–5 years (Table 1), particip<strong>at</strong>ed<strong>in</strong> the study. All experiments were conducted <strong>at</strong>the Veter<strong>in</strong>ary Faculty of the University of Hels<strong>in</strong>ki.Prior to conduct<strong>in</strong>g the experiments (1–2 months), the<strong>do</strong>gs were tra<strong>in</strong>ed by their owners as <strong>in</strong>structed by the firstauthor. <strong>Dogs</strong> were tra<strong>in</strong>ed us<strong>in</strong>g an operant-positive condition<strong>in</strong>gmethod (clicker) to lie on a 10-cm-thick Styrofoamm<strong>at</strong>tress and lean their jaw on a purpose-designedu-shaped ch<strong>in</strong> rest for up to 60 s (see Fig. 1 for experimentaldetails). Dur<strong>in</strong>g the last 2–3 weeks of the tra<strong>in</strong><strong>in</strong>gperiod, the <strong>do</strong>gs and their owners visited the experimentalTable 1 Breed, age, and gender of the <strong>do</strong>gs particip<strong>at</strong><strong>in</strong>g <strong>in</strong> theexperimentID Breed Age (years) GenderD1 Beauce shepherd 1 FemaleD2 Beauce shepherd 2 FemaleD3 Beauce shepherd 2.5 FemaleD4 Rough collie 1 FemaleD5 Hovawart 4 Castr<strong>at</strong>edD6 Gre<strong>at</strong> pyreness 5 Female123


170 Anim Cogn (2012) 15:163–174Fig. 4 An example of <strong>eye</strong> gaze p<strong>at</strong>terns dur<strong>in</strong>g present<strong>at</strong>ion of theDOG image (a) and the BLANK screen (b). The focus map representsaveraged fix<strong>at</strong>ions of the right <strong>eye</strong> of five <strong>do</strong>gs and three repetitionspresented consecutively. The color cod<strong>in</strong>g represents the average offix<strong>at</strong>ion dur<strong>at</strong>ions; m<strong>in</strong>imum 5 ms <strong>in</strong>dic<strong>at</strong>ed by light blue and themaximum of 100 ms or over by bright red. The area correspond<strong>in</strong>g tothe image size and placement is overlaid on the BLANK screen as adashed l<strong>in</strong>eFig. 5 The effect of novelty st<strong>at</strong>us of the image on the number offix<strong>at</strong>ions per frame (mean ± SEM) and the total dur<strong>at</strong>ion of fix<strong>at</strong>ionsper frame (mean <strong>in</strong> ms ± SEM) <strong>in</strong> six <strong>do</strong>gs. In the novelty paradigm,the same stimulus image repe<strong>at</strong>s 3–5 times (last frame is consideredas FAMILIAR) and then changes (SHIFT) to another stimulus imagefrom the same stimulus c<strong>at</strong>egory (NOVEL). St<strong>at</strong>istically significantdifferences between the 1st, FAMILIAR, and NOVEL frames arepresented as different letters (MIXED, P \ 0.05)LETTER n = 113) were found for the number of fix<strong>at</strong>ions,the total dur<strong>at</strong>ion of fix<strong>at</strong>ions, and the rel<strong>at</strong>ive fix<strong>at</strong>iondur<strong>at</strong>ion (Table 3). The <strong>do</strong>gs fix<strong>at</strong>ed on DOG <strong>images</strong> moreoften (2.0 ± 0.3) than on HUMAN (1.6 ± 0.3, P = 0.014),ITEM (1.2 ± 0.3, P = 0.006), or LETTER <strong>images</strong>(0.5 ± 0.5, P = 0.002). LETTER <strong>images</strong> g<strong>at</strong>hered fewerfix<strong>at</strong>ions than other c<strong>at</strong>egories (LETTER vs. DOG P =0.000; LETTER vs. HUMAN P = 0.025; LETTER vs.ITEM P = 0.043; Fig. 7).The mean dur<strong>at</strong>ion of a s<strong>in</strong>gle fix<strong>at</strong>ion was 214 ms (95%CI 154–289 ms) on average and did not differ between thec<strong>at</strong>egories. However, the total fix<strong>at</strong>ion dur<strong>at</strong>ion among thefour image c<strong>at</strong>egories differed st<strong>at</strong>istically significant(P = 0.000; Table 3). <strong>Dogs</strong> fix<strong>at</strong>ed longest on the <strong>images</strong>Fig. 6 The effect of the repe<strong>at</strong>ed image present<strong>at</strong>ion on a the numberof fix<strong>at</strong>ions (mean ± SEM) and the dur<strong>at</strong>ion of a s<strong>in</strong>gle fix<strong>at</strong>ion(mean <strong>in</strong> ms ± SEM) per frame b the total dur<strong>at</strong>ion of fix<strong>at</strong>ions perframe (mean <strong>in</strong> ms ± SEM) and the rel<strong>at</strong>ive fix<strong>at</strong>ion dur<strong>at</strong>ion of thefix<strong>at</strong>ions targeted to the object (% ± SEM) <strong>in</strong> six <strong>do</strong>gs dur<strong>in</strong>g anexperimental trial where a total of six frames of two different <strong>images</strong>from the same stimulus c<strong>at</strong>egory are shown as a novelty paradigm.St<strong>at</strong>istically significant differences between the frames are presentedas different letters (MIXED, P \ 0.05)of <strong>do</strong>gs (534 ± 80 ms) and shortest (94 ± 120 ms) on the<strong>images</strong> of alphabetic characters (DOG vs. HUMANP = 0.024; DOG vs. ITEM P = 0.001; DOG vs. LETTER123


Anim Cogn (2012) 15:163–174 171Fig. 7 The effect of the image c<strong>at</strong>egory (DOG, HUMAN, ITEM, andLETTER) on the number of fix<strong>at</strong>ions (mean ± SEM) and the totaldur<strong>at</strong>ion of fix<strong>at</strong>ions (mean <strong>in</strong> ms ± SEM) per frame <strong>in</strong> six <strong>do</strong>gs.St<strong>at</strong>istically significant differences between the image c<strong>at</strong>egories arepresented as different letters (MIXED, P \ 0.05)P = 0.000; LETTER vs. HUMAN P = 0.005; LETTERvs. ITEM P = 0.029; Fig. 7). The ma<strong>in</strong> effect of the imagec<strong>at</strong>egory on the rel<strong>at</strong>ive fix<strong>at</strong>ion dur<strong>at</strong>ion of the object wasst<strong>at</strong>istically significant (P = 0.042), but the pairwisecomparisons did not specify which c<strong>at</strong>egories differed fromeach other (DOG 65.4 ± 6.4%; HUMAN 56.2 ± 6.7%;ITEM 60.4 ± 8.4%; and LETTER 39.8 ± 13.3%).DiscussionThe current study produced evidence on the can<strong>in</strong>e visual<strong>cognition</strong> by a new method: a contact-free <strong>eye</strong> movement<strong>track<strong>in</strong>g</strong>. <strong>Dogs</strong> focused their <strong>at</strong>tention on the <strong>in</strong>form<strong>at</strong>iveregions of the <strong>images</strong> without any task-specific pre-tra<strong>in</strong><strong>in</strong>g,and their gaz<strong>in</strong>g behavior depended on the imagec<strong>at</strong>egory.<strong>Dogs</strong> focused most of their fix<strong>at</strong>ion dur<strong>at</strong>ion on theactual image compared with the surround<strong>in</strong>g monitor an<strong>do</strong>n the actual object compared with the background image,as it has previously been reported for humans, chimpanzees,and monkeys (Yarbus 1967; Nahm et al. 1997; Kanoand Tomonaga 2009). <strong>Dogs</strong> spontaneously prefer <strong>images</strong>of conspecifics over human faces and <strong>in</strong>anim<strong>at</strong>e objects,suggest<strong>in</strong>g they might be able to discrim<strong>in</strong><strong>at</strong>e <strong>images</strong> ofdifferent c<strong>at</strong>egories.Various animal species can form visual c<strong>at</strong>egories whentra<strong>in</strong>ed us<strong>in</strong>g a m<strong>at</strong>ch-to-sample procedure (Bovet andVauclair 2000). However, animals quickly learn to repe<strong>at</strong>the re<strong>in</strong>forced behaviors (Sk<strong>in</strong>ner 1938), even uncharacteristicones, and hence, the n<strong>at</strong>ural <strong>look</strong><strong>in</strong>g behavior couldrema<strong>in</strong> hidden (Dahl et al. 2007). Explicit reward<strong>in</strong>g forcerta<strong>in</strong> pre-def<strong>in</strong>ed criterion could lead to <strong>at</strong>ypical responsestr<strong>at</strong>egies and limit the comparability between studies <strong>do</strong>newith naïve human subjects (Murai et al. 2005; Dahl et al.2009). In the current study, the differences between stimulusc<strong>at</strong>egories arise from unprompted <strong>at</strong>tention measureddirectly as fix<strong>at</strong>ional <strong>eye</strong> movements.The c<strong>at</strong>egories g<strong>at</strong>hered different numbers of fix<strong>at</strong>ions,while the average dur<strong>at</strong>ion of a s<strong>in</strong>gle fix<strong>at</strong>ion was the same<strong>in</strong> all c<strong>at</strong>egories. The role of the fix<strong>at</strong>ion is to keep the gazestable enough for stimulus encod<strong>in</strong>g <strong>in</strong> the photoreceptors(Land 1999; Yarbus 1967); thus, a sufficiently long dur<strong>at</strong>ionfor s<strong>in</strong>gle fix<strong>at</strong>ion is needed to identify the object. Inhumans, targets <strong>in</strong>terpreted as <strong>in</strong>form<strong>at</strong>ive or <strong>in</strong>terest<strong>in</strong>g<strong>at</strong>tract more re-fix<strong>at</strong>ions (Buswell 1935; Henderson andHoll<strong>in</strong>gworth 1999). In the current study, the <strong>images</strong> ofletters received the least number of fix<strong>at</strong>ions and thus theshortest overall fix<strong>at</strong>ion dur<strong>at</strong>ion. The f<strong>in</strong>d<strong>in</strong>g correspondswith behavioral observ<strong>at</strong>ions <strong>in</strong> a recent study of Faragóet al. (2010) who suggested th<strong>at</strong> <strong>do</strong>gs might consider n<strong>at</strong>uralobjects more <strong>in</strong>terest<strong>in</strong>g than abstract ones. On theother hand, these complex pictures might conta<strong>in</strong> more<strong>in</strong>form<strong>at</strong>ion to process. The lowest level for the pictureperception can occur on the basis of physical fe<strong>at</strong>ures (i.e.,color, <strong>in</strong>tensity, contrasts, or edge orient<strong>at</strong>ion), which <strong>do</strong>esnot require the understand<strong>in</strong>g the represent<strong>at</strong>ional contentof the image (Fagot et al. 1999, Bovet and Vauclair 2000).In <strong>eye</strong> movements, the targets of the fix<strong>at</strong>ions could bedriven not only by higher level conceptual <strong>in</strong>form<strong>at</strong>ion ofthe objects but also by low-level visual saliency (Henderson2003; Henderson and Holl<strong>in</strong>gworth 1999). In fact, themore complex the image, the more it conta<strong>in</strong>s details th<strong>at</strong>might <strong>at</strong>tract fix<strong>at</strong>ions and hence <strong>in</strong>crease the time it takesto view and process the image. Therefore, the c<strong>at</strong>egorydependentgaz<strong>in</strong>g behavior could also be the result of differences<strong>in</strong> physical complexity. In humans, the high- andlow-level mechanisms of the guidance of <strong>eye</strong> movementsare dependent on the task and could altern<strong>at</strong>e by turns(E<strong>in</strong>häuser et al. 2008a, b). Thus, based only on <strong>eye</strong>movement d<strong>at</strong>a, we cannot yet draw any conclusions as towhether the <strong>at</strong>tention of <strong>do</strong>gs was ma<strong>in</strong>ly directed bystimulus fe<strong>at</strong>ures or semantic <strong>in</strong>form<strong>at</strong>ion, or both.The letter <strong>images</strong> were much simpler <strong>in</strong> their fe<strong>at</strong>uresthan were the other c<strong>at</strong>egories, but the <strong>images</strong> of <strong>do</strong>g faceswere not apparently more complex from their physicalproperties than presented human faces. However, <strong>do</strong>gsfix<strong>at</strong>ed for more times and longer total dur<strong>at</strong>ion on can<strong>in</strong>efaces than on human faces. A preference for conspesificfaces has been suggested to <strong>in</strong>dic<strong>at</strong>e expertise <strong>in</strong> the perceptionof the faces of own species (Dahl et al. 2009;H<strong>at</strong>tori et al. 2010). Recent behavioral studies <strong>in</strong> <strong>do</strong>gs alsofound species-dependent <strong>look</strong><strong>in</strong>g behavior when view<strong>in</strong>ghuman and can<strong>in</strong>e faces (Guo et al. 2009; Racca et al.2010). Face detection plays an important role <strong>in</strong> non-verbalcommunic<strong>at</strong>ion <strong>in</strong> prim<strong>at</strong>es and probably also <strong>in</strong> other123


172 Anim Cogn (2012) 15:163–174social mammals (Leopold and Rhodes 2010). The <strong>do</strong>gmight perceive conspecific faces to be more <strong>in</strong>terest<strong>in</strong>g or<strong>in</strong>form<strong>at</strong>ive than human faces due to different social relevanciesand communic<strong>at</strong>ion str<strong>at</strong>egies between <strong>do</strong>gs and<strong>do</strong>gs, and <strong>do</strong>gs and humans. One factor th<strong>at</strong> might haveaffected c<strong>at</strong>egory preferences is a vari<strong>at</strong>ion with<strong>in</strong> a c<strong>at</strong>egory;the <strong>do</strong>g faces represent many breeds, while humanfaces were all Caucasian.As has been found <strong>in</strong> human <strong>in</strong>fants (Fantz 1964;reviewed <strong>in</strong> Colombo and Micthell 2009) and monkeys(Joseph et al. 2006), the first frame <strong>at</strong>tracted the highesttotal <strong>look</strong><strong>in</strong>g time, which decreased when the image wasrepe<strong>at</strong>ed, probably <strong>in</strong>dic<strong>at</strong><strong>in</strong>g habitu<strong>at</strong>ion to the stimulus.However, we could not detect expected rebound <strong>in</strong> <strong>look</strong><strong>in</strong>gwhen the novel picture was presented. Instead, the totaldur<strong>at</strong>ion of fix<strong>at</strong>ions decreased after the stimulus changed.It is likely th<strong>at</strong> <strong>do</strong>gs did not notice the change or they mayhave generalized familiar pictures to novel ones. In generaliz<strong>at</strong>ion,two stimuli, even if readily dist<strong>in</strong>guishable, aresimilar enough with respect to their physical properties toevoke the same response (Shepard 1994, Chirlanda andEnqvist 2003; generaliz<strong>at</strong>ion of n<strong>at</strong>ural 2D pictures hasbeen demonstr<strong>at</strong>ed <strong>in</strong> <strong>do</strong>gs by Range et al. 2008). Most ofthe familiar and the familiar–novel pairs were r<strong>at</strong>her similar;for example, humans were paired accord<strong>in</strong>g to gender,age, hair color, and expression. When the stimuli arecomplex, <strong>do</strong>gs’ performance could be limited as comparedwith other species with better visual acuity and color perception(Range et al. 2008). It is possible th<strong>at</strong> the <strong>do</strong>gs <strong>do</strong>not perceive the m<strong>in</strong>or details of the picture and thuscannot discrim<strong>in</strong><strong>at</strong>e the <strong>images</strong> with<strong>in</strong> the same c<strong>at</strong>egory.An <strong>in</strong>teraction between c<strong>at</strong>egory and novelty st<strong>at</strong>us wasnot established, which is partly contradictory to the recentstudy of Racca et al. (2010) <strong>in</strong> which <strong>do</strong>gs also directedshorter <strong>look</strong><strong>in</strong>g times to novel can<strong>in</strong>e faces, but longerdur<strong>at</strong>ions <strong>at</strong> novel human faces and objects. C<strong>at</strong>egorydependentnovelty responses have also been reported forhuman observers, who prefer familiarity <strong>in</strong> human facesbut novelty <strong>in</strong> n<strong>at</strong>ural scenes, and had no clear preferencefor geometric figures (Park et al. 2010). The differentresults may also be due to many metho<strong>do</strong>logical reasons;we exam<strong>in</strong>ed fix<strong>at</strong>ional <strong>eye</strong> movements, not overall orient<strong>in</strong>gtime as Racca et al. (2010). The overall orient<strong>in</strong>gtime <strong>in</strong>cludes also the <strong>in</strong>active view<strong>in</strong>g (‘‘blank stares’’;Asl<strong>in</strong> 2007). Moreover, the present<strong>at</strong>ion setup differed. Inthe current study, the design was a modified version of th<strong>at</strong>used for monkeys (Joseph et al. 2006). The stimulus wasrepe<strong>at</strong>ed several times, and <strong>do</strong>gs were rewarded after thesixth frame. All <strong>do</strong>gs were highly motiv<strong>at</strong>ed, but thedecrease <strong>in</strong> the number and total dur<strong>at</strong>ion of fix<strong>at</strong>ionsdur<strong>in</strong>g the last frames of the trial suggest th<strong>at</strong> they mighthave got tired of the monotony of the task. It is also likelyth<strong>at</strong> they anticip<strong>at</strong>ed the reward and therefore were notfocus<strong>in</strong>g their <strong>at</strong>tention <strong>at</strong> the monitor <strong>at</strong> the end of the trial.Vary<strong>in</strong>g the length of trials could have represented a betterdesign for can<strong>in</strong>e research. <strong>Dogs</strong> might detect the differencebetween familiar and novel <strong>images</strong> better if the<strong>images</strong> are presented side by side as <strong>in</strong> visual pairedcomparison study of Racca et al. (2010). Different metho<strong>do</strong>logieshave led to contradict<strong>in</strong>g novelty responses also<strong>in</strong> <strong>in</strong>fant studies (Houston-Price and Nakai 2004).The absence of preferential <strong>look</strong><strong>in</strong>g <strong>do</strong>es not necessarilymean the absence of discrim<strong>in</strong><strong>at</strong>ion (Asl<strong>in</strong> 2007). Noveltypreferences can vary <strong>in</strong>dividually; the subjects might haveperceived certa<strong>in</strong> c<strong>at</strong>egories more <strong>at</strong>tractive or might haveused <strong>in</strong>dividual str<strong>at</strong>egies for detect<strong>in</strong>g a new stimulus, assuggested previously for monkeys and humans (Josephet al. 2006). Preference may even focus on certa<strong>in</strong> details ofthe object, for example, the head area <strong>in</strong> whole-body picturesQu<strong>in</strong>n et al. (2009). Also the physical similarity anddifferences <strong>in</strong> <strong>at</strong>tractiveness between consecutive <strong>images</strong>could affect novelty responses. A rebound <strong>in</strong> <strong>look</strong><strong>in</strong>g islikely to occur when consecutive <strong>images</strong> differ more fromeach other and when the novel image is more <strong>in</strong>terest<strong>in</strong>g <strong>in</strong>content than the previous one (Dahl et al. 2007).The <strong>do</strong>gs <strong>in</strong> the current study performed the tasks<strong>in</strong>dependently while the owners and experimentersrema<strong>in</strong>ed hidden beh<strong>in</strong>d an opaque barrier. The <strong>do</strong>gs weretra<strong>in</strong>ed neither to fix<strong>at</strong>e on the monitor nor to discrim<strong>in</strong><strong>at</strong>e<strong>images</strong>. However, all measured <strong>eye</strong> movement variables<strong>in</strong>dic<strong>at</strong>ed th<strong>at</strong> the <strong>do</strong>gs were more <strong>in</strong>terested <strong>in</strong> <strong>look</strong><strong>in</strong>g <strong>at</strong>the monitor when <strong>images</strong> were displayed than when thescreen was blank. This f<strong>in</strong>d<strong>in</strong>g confirms th<strong>at</strong> the <strong>do</strong>gs hadnot learned to fix their gaze on the monitor <strong>in</strong> anticip<strong>at</strong>ionof a reward or a response to social cue<strong>in</strong>g. To the authors’knowledge, blank screen view<strong>in</strong>g has not been previouslymeasured <strong>in</strong> animal studies. <strong>Dogs</strong> targeted some fix<strong>at</strong>ionstoward the blank screen, which is typical also <strong>in</strong> human<strong>look</strong><strong>in</strong>g behavior because the re-activ<strong>at</strong>ion of memoryrepresent<strong>at</strong>ions drives the <strong>eye</strong>s to previously viewed loc<strong>at</strong>ions(Ferreira et al. 2008).Clearly, we cannot yet say unequivocally wh<strong>at</strong> the <strong>do</strong>gssee <strong>in</strong> the pictures. In humans, a picture is someth<strong>in</strong>g <strong>in</strong>which objects can be recognized, even though the objectsthemselves are not actually present. It is under deb<strong>at</strong>ewhether animals recognize pictures as represent<strong>at</strong>ions ofreal-world objects (Bovet and Vauclair 2000; Jitsumori2010). <strong>Dogs</strong> can associ<strong>at</strong>e visual image with acoustical<strong>in</strong>form<strong>at</strong>ion, suggest<strong>in</strong>g they are capable of form<strong>in</strong>g mentalrepresent<strong>at</strong>ions through pictures (Adachi et al. 2007, Faragóet al. 2010). It has also been demonstr<strong>at</strong>ed through afetch<strong>in</strong>g task th<strong>at</strong> <strong>at</strong> least some <strong>do</strong>gs are able to m<strong>at</strong>chphotographs of items to actual objects (Kam<strong>in</strong>ski et al.2009). Nevertheless, <strong>eye</strong> movement <strong>track<strong>in</strong>g</strong> is a promis<strong>in</strong>gmethod for compar<strong>in</strong>g visual perception str<strong>at</strong>egies andabilities between humans and <strong>do</strong>gs.123


Anim Cogn (2012) 15:163–174 173In conclusion, contact-free <strong>eye</strong> movement <strong>track<strong>in</strong>g</strong> canbe used to assess can<strong>in</strong>e visual <strong>cognition</strong>. This promis<strong>in</strong>gmethod represents a tool for the broader explor<strong>at</strong>ion ofprocesses underly<strong>in</strong>g special socio-cognitive skills <strong>in</strong> <strong>do</strong>gspreviously established through behavioral studies. <strong>Dogs</strong>’<strong>at</strong>traction to conspecific and human faces over <strong>in</strong>anim<strong>at</strong>eobjects might reflect the n<strong>at</strong>ural <strong>in</strong>terests of <strong>do</strong>gs, but furtherstudies are needed to establish if <strong>do</strong>gs possess pictureobject re<strong>cognition</strong>.Acknowledgments This work was f<strong>in</strong>ancially supported by theAcademy of F<strong>in</strong>land and University of Hels<strong>in</strong>ki. The authors aregr<strong>at</strong>eful to Antti Flyckt, M<strong>at</strong>ti Pastell, Aleksander Alafuzoff, TeemuPeltonen, Jaana Simola, Timo Murtonen, and Kristian Törnqvist fortheir support <strong>in</strong> conduct<strong>in</strong>g the experiment. Authors also thank IKEAgroup for the permission to use the photos of children’s toys.Conflict of <strong>in</strong>terestof <strong>in</strong>terest.ReferencesThe authors declare th<strong>at</strong> they have no conflictAdachi I, Kuwah<strong>at</strong>a H, Fujita K (2007) <strong>Dogs</strong> recall their owner’s faceupon hear<strong>in</strong>g the owner’s voice. Anim Cogn 10:17–21Asl<strong>in</strong> RN (2007) Wh<strong>at</strong>’s <strong>in</strong> a <strong>look</strong>? Dev Sci 10:48–53Bovet D, Vauclair J (2000) Picture re<strong>cognition</strong> <strong>in</strong> animals andhumans. Behav Bra<strong>in</strong> Res 109:143–165Buswell GT (1935) How people <strong>look</strong> <strong>at</strong> pictures; a study of thepsychology of perception <strong>in</strong> art. University of Chicago Press,ChicagoChirlanda S, Enqvist M (2003) A century of generaliz<strong>at</strong>ion. AnimBehav 66:15–36Colombo J, Micthell DW (2009) Infant visual habitu<strong>at</strong>ion. NeurobiolLearn Mem 92:225–234Dahl CD, Logothetis NK, Hoffman KL (2007) Individu<strong>at</strong>ion andholistic process<strong>in</strong>g of faces <strong>in</strong> rhesus monkeys. Proc Rev Soc BBiol Sci 274:2069–2076Dahl CD, Wallraven C, Bülthoff HH, Logothetis NK (2009) Humansand macaques employ similar face-process<strong>in</strong>g str<strong>at</strong>egies. CurrBiol 19:509–513Dell’Osso LF, Williams RW, Jacobs JB, Erchul DM (1998) Thecongenital and see-saw nystagmus <strong>in</strong> the prototypical achiasmaof can<strong>in</strong>es: comparison to the human achiasm<strong>at</strong>ic prototype. VisRes 38:1629–1641E<strong>in</strong>häuser W, Rutishauser U, Koch C (2008a) Task-demands canimmedi<strong>at</strong>ely reverse the effects of sensory-driven saliency <strong>in</strong>complex visual stimuli. J Vis 8:1–19E<strong>in</strong>häuser W, Spa<strong>in</strong> M, Perona P (2008b) Objects predict fix<strong>at</strong>ionsbetter than early saliency. J Vis 8:1–26Fagot J, Mart<strong>in</strong>-Malivel J, Dépy D (1999) Wh<strong>at</strong> is the evidence for anequivalence between objects and pictures <strong>in</strong> birds and nonhumanprim<strong>at</strong>es? Curr Psychol Cogn 18:923–949Fantz RL (1964) Visual experience <strong>in</strong> <strong>in</strong>fants: decreased <strong>at</strong>tention tofamiliar p<strong>at</strong>terns rel<strong>at</strong>ive to novel ones. Sci 146:668–670Faragó T, Pongràcz P, Miklósi Á, Huber L, Virányi Z, Range F(2010) <strong>Dogs</strong>’ expect<strong>at</strong>ion about signalers’ body size virtue oftheir growls. PLoS One 12:1–8Ferreira F, Apel J, Henderson JM (2008) Tak<strong>in</strong>g a new <strong>look</strong> <strong>at</strong><strong>look</strong><strong>in</strong>g <strong>at</strong> noth<strong>in</strong>g. Trends Cogn Sci 12:405–410Guo K, Me<strong>in</strong>ts K, Hall C, Hall S, Mills D (2009) Left gaze bias <strong>in</strong> humans,rhesus monkeys and <strong>do</strong>mestic <strong>do</strong>gs. Anim Cogn 12:409–418Hare B, Tomasello M (2005) Human-like social skills <strong>in</strong> <strong>do</strong>gs?Trends Cogn Sci 9:439–444H<strong>at</strong>tori Y, Kano F, Tomonaga M (2010) Differential sensitivity toconspecific and allospecific cues <strong>in</strong> chimpanzees and humans: acompar<strong>at</strong>ive <strong>eye</strong>-<strong>track<strong>in</strong>g</strong> study. Biol Lett 6:610–613Henderson JM (2003) Human gaze control dur<strong>in</strong>g real-world sceneperception. Trends Cogn Sci 7:498–504Henderson JM, Holl<strong>in</strong>gworth A (1999) High-level scene perception.Annu Rev Psychol 50:243–271Houston-Price C, Nakai S (2004) Dist<strong>in</strong>guish<strong>in</strong>g novelty andfamiliarity effects <strong>in</strong> <strong>in</strong>fant preference procedures. Infant ChildDev 13:341–348Jacobs JB, Dell’Osso LF, Wang ZI, Acland GM, Bennett J (2009)Us<strong>in</strong>g the NAFX to measure the effectiveness over time ofgene therapy <strong>in</strong> can<strong>in</strong>e LCA. Invest Ophthalmol Vis Sci50:4685–4692Jitsumori M (2010) Do animals recognize pictures as represent<strong>at</strong>ionsof 3D objects? Comp Cogn Behav Rev 5:136–138Joseph JE, Powell DK, Andersen AH, Bh<strong>at</strong>t RS, Dunlap MK, FoldesST, Forman E, Hardy PA, Ste<strong>in</strong>metz NA, Zhang Z (2006) fMRI<strong>in</strong> alert, behav<strong>in</strong>g monkeys: an adapt<strong>at</strong>ion of the human <strong>in</strong>fantfamiliariz<strong>at</strong>ion novelty preference procedure. J Neurosci Methods157:10–24Kam<strong>in</strong>ski J, Tempelmann S, Call J, Tomasello M (2009) Domestic<strong>do</strong>gs comprehend human communic<strong>at</strong>ion with iconic signs. DevSci 12:831–837Kano F, Tomonaga M (2009) How chimpanzees <strong>look</strong> <strong>at</strong> pictures: acompar<strong>at</strong>ive <strong>eye</strong>-<strong>track<strong>in</strong>g</strong> study. Proc Biol Sci 276:1949–1955Land MF (1999) Motion and vision: why animals move their <strong>eye</strong>s.J Comp Physiol A Neuroethol Sens Neural Behav Physiol185:341–352Leopold DA, Rhodes G (2010) A compar<strong>at</strong>ive view of faceperception. J Comp Psychol 124:233–251Lit L, Schweitzer JB, Oberbauer AM (2011) Handler beliefs affectscent detection <strong>do</strong>g outcomes. Anim Cogn 14:387–394Miklósi Á, Kub<strong>in</strong>yi E, Topál J, Gacsi M, Viranyi Z, Csanyi V (2003)A simple reason for a big difference: wolves <strong>do</strong> not <strong>look</strong> back <strong>at</strong>humans, but <strong>do</strong>gs <strong>do</strong>. Curr Biol 13:763–766Miklósi Á, Topál J, Csányi V (2007) Big thoughts <strong>in</strong> small bra<strong>in</strong>s?<strong>do</strong>gs as a model for understand<strong>in</strong>g human social <strong>cognition</strong>.Neuroreport 18:467–471Miller PE, Murphy CJ (1995) Vision <strong>in</strong> <strong>do</strong>gs. J Am Vet Med Assoc207:1623–1634Murai C, Kosugi D, Tomonaga M, Tanaka M, M<strong>at</strong>suzawa T, Itakura S(2005) Can chimpanzee <strong>in</strong>fants (Pan troglodytes) form c<strong>at</strong>egoricalrepresent<strong>at</strong>ions <strong>in</strong> the same manner as human <strong>in</strong>fants (Homosapiens)? Dev Sci 8:240–254Nahm FKD, Perret A, Amaral DG, Albright TD (1997) How <strong>do</strong>monkeys <strong>look</strong> <strong>at</strong> faces? J Cogn Neurosci 9:611–623Park J, Shimojo E, Shimojo S (2010) Roles of familiarity and novelty<strong>in</strong> visual preference judgments are segreg<strong>at</strong>ed across objectc<strong>at</strong>egories. Proc N<strong>at</strong>l Acad Sci USA 107:14552–14555Pfungst O (1907) Das Pferd des Herrn von Osten (der Kluge Hans):E<strong>in</strong> Beitrag zur experimentellen Tier-und Menchenpsychologie.Johann Ambrosius Barth, LeipzigQu<strong>in</strong>n PC, Doran MM, Reiss JE, Hoffman JE (2009) Time course ofvisual <strong>at</strong>tention <strong>in</strong> <strong>in</strong>fant c<strong>at</strong>egoriz<strong>at</strong>ion of c<strong>at</strong>s versus <strong>do</strong>gs:evidence for a head bias as revealed through <strong>eye</strong> <strong>track<strong>in</strong>g</strong>. ChildDev 80:151–161Racca A, Amadei E, Ligout S, Guo K, Me<strong>in</strong>ts K, Mills D (2010)Discrim<strong>in</strong><strong>at</strong>ion of human and <strong>do</strong>g faces and <strong>in</strong>version responses<strong>in</strong> <strong>do</strong>mestic <strong>do</strong>gs (Canis familiaris). Anim Cogn 13:525–533Range F, Aust U, Steurer M, Huber L (2008) Visual c<strong>at</strong>egoriz<strong>at</strong>ion ofn<strong>at</strong>ural stimuli by <strong>do</strong>mestic <strong>do</strong>gs. Anim Cogn 11:339–347Shepard R (1994) Perceptual-cognitive universals as reflections of theworld. Psychon Bull Rev 1:2–28123


174 Anim Cogn (2012) 15:163–174Sk<strong>in</strong>ner (1938) The behavior of organisms: an experimental analysis.D. Appleton-century company. New York, p 457Soproni K, Miklósi Á, Topál J, Csányi V (2002) <strong>Dogs</strong>’ (Canisfamiliaris) responsiveness to human po<strong>in</strong>t<strong>in</strong>g gestures. J CompPsychol 116:27–34Topál J, Miklósi Á, Gácsi M, Dóka A, Pongrácz P, Kub<strong>in</strong>yi E,Virányi Z, Csányi V (2009) The <strong>do</strong>g as a model forunderstand<strong>in</strong>g human social behavior. In: Brockmann HJ, RoperTJ, Naguib M, Wynne-Edwards KE, Mitani JC, Simmons LW(eds) Advances <strong>in</strong> the study of behavior, vol 39. Academic Press,Burl<strong>in</strong>gton, pp 71–116Virányi Z, Topál J,Gácsi M, Miklósi Á, Csányi V (2004) <strong>Dogs</strong>respond appropri<strong>at</strong>ely to cues of humans’ <strong>at</strong>tentional focus.Behav Process 66:161–172Williams FJ, Mills DS, Guo K (2011) Development of a headmounted,<strong>eye</strong>-<strong>track<strong>in</strong>g</strong> system for <strong>do</strong>gs. J Neurosci Methods94:259–265Yarbus AL (1967) Eye movements and vision. Plenum Press, New York123

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!