Human attention affects facial expressions in domestic dogs

Abstract

Most mammalian species produce facial expressions. Historically, animal facial expressions have been considered inflexible and involuntary displays of emotional states rather than active attempts to communicate with others. In the current study, we aimed to test whether domestic dog facial expressions are subject to audience effects and/ or changes in response to an arousing stimulus (e.g. food) alone. We presented dogs with an experimental situation in which a human demonstrator was either attending to them or turned away, and varied whether she presented food or not. Dogs produced significantly more facial movements when the human was attentive than when she was not. The food, however, as a non-social but arousing stimulus, did not affect the dogs’ behaviour. The current study is therefore evidence that dogs are sensitive to the human’s attentional state when producing facial expressions, suggesting that facial expressions are not just inflexible and involuntary displays of emotional states, but rather potentially active attempts to communicate with others.

Full text

1
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
www.nature.com/scientificreports
Human attention aects facial
expressions in domestic dogs
Juliane Kaminski , Jennifer Hynds, Paul Morris & Bridget M. Waller
Most mammalian species produce facial expressions. Historically, animal facial expressions have
been considered inexible and involuntary displays of emotional states rather than active attempts
to communicate with others. In the current study, we aimed to test whether domestic dog facial
expressions are subject to audience eects and/ or changes in response to an arousing stimulus (e.g.
food) alone. We presented dogs with an experimental situation in which a human demonstrator was
either attending to them or turned away, and varied whether she presented food or not. Dogs produced
signicantly more facial movements when the human was attentive than when she was not. The
food, however, as a non-social but arousing stimulus, did not aect the dogs’ behaviour. The current
study is therefore evidence that dogs are sensitive to the human’s attentional state when producing
facial expressions, suggesting that facial expressions are not just inexible and involuntary displays of
emotional states, but rather potentially active attempts to communicate with others.
Most mammalian species produce facial expressions, which form meaningful and adaptive components of the
animal’s behavioural repertoire. e facial architecture underlying such facial expressions is highly conserved
among mammals1, suggesting that human facial expression is based on evolutionarily ancient systems. erefore,
it would seem reasonable to debate the extent to which such facial expressions are underpinned by sophisticated
cognitive processes. Historically, animal facial expressions (including human, to an extent) have been considered
inexible and involuntary displays e.g.2,3, reecting an individual’s emotional state rather than active attempts to
communicate with others. ere is some evidence that non-human primate facial expressions can be mediated by
the presence of an audience, suggesting that the sender has some understanding of whether the expressions can
be seen by others4–8. Waller et al. (2015) showed that the production of facial expressions in orangutans is more
intense and more complex during play when a recipient is directed towards them suggesting that the production
of these expressions is not necessarily an automated response and subject to audience eects6. Similarly, Scheider
et al. (2016) showed that Gibbons presented their facial expressions more oen and over a longer duration when
facing other individuals compared to non-facing situation7.
To date there is no systematic experimental evidence, however, that facial expressions in species other than
primates, are produced with similar sensitivity to the attention of the audience.
With the current study, we aimed to test whether domestic dog facial expressions change in response to an
highly arousing but non-social stimulus (food) and/or the changing attentional state of their human audience.
Domestic dogs are a potentially interesting model for this kind of research as they have a unique history. Dogs
have been living with humans for about 30,000 years9, during which time selection pressures seem to have acted
on dogs’ ability to communicate with humans. [see for a review10 and11,12 for a recent discussion].
ere is broad evidence that domestic dogs attend to a human’s attentional state13–15, which is one indicator
of intentionality16. Aer being told not to take a piece of food, dogs steal the food more oen when the human’s
eyes are closed compared to situations during which the human’s eyes are open, the human has her back turned
to the dog or she is distracted13,15. Dogs are also sensitive to the human’s attentional state during communicative
interactions with humans. Dogs follow communicative gestures more once the humans eyes are visible and the
gesture is clearly directed at them16. Dogs also follow the gaze of a human to a target only if eye contact had been
established prior to the gaze shi17.
Waller et al. (2013) analysed the facial expressions of dogs waiting to be rehomed in shelters, and found
a negative correlation between the frequency of facial movements the dogs produced when interacting with a
stranger, and the rate at which they were re-homed. e more oen dogs produced a specic facial movement,
Action Unit 101 (which raises the inner eyebrow) the quicker they were re-homed18. Raising the inner eyebrow
changes the visual appearance of the eyes and makes them look bigger, a key feature of paedomorphism (juvenile
University of Portsmouth King Henry 1st Street, Portsmouth, PO12DY, UK. Correspondence and requests for
materials should be addressed to J.K. (email: juliane.kaminski@port.ac.uk)
Received: 18 May 2017
Accepted: 18 September 2017
Published: xx xx xxxx
OPEN
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
2
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
features present in the adult). One hypothesis is therefore that by picking dogs that raise their inner eye brow
more, humans simply follow their preference for paedomorphic facial characteristics, a preference which might
have acted as a selection pressure during dog domestication18,19.
It is possible, therefore, that dogs have also evolved the ability to use these facial expressions dierentially
depending on their audience. In which case, during domestication dogs may have gained additional cognitive
control of their facial expressions.
e current study investigated whether dog facial expressions can be subject to so called audience eects, and
can therefore be tailored to the human’s attentional state, which might suggest some social communicative func-
tion and possible voluntary control. e alternative is that dog facial expressions are a simple emotional display
based on the dog’s state of arousal. In order to try and discriminate between these two possible explanations, the
human, depending on the condition, presented a piece of food, as a non-social and arousing stimulus (a recent
study, using thermal imaging, shows that food seems to be more arousing for dogs than social contact with a
human as long as the human remains silent20).
erefore, if dogs produce facial expressions merely as an emotional display, we would expect them to not
necessarily dierentiate between the social (human attention) and the non social (food) conditions. However, if
the dogs behave in dierent ways in responds to the social and the non-social stimuli, this would provide some
evidence that dogs discriminate between the conditions based on social context and one possible explanation for
such discrimination that dogs exercise some voluntary control.
Results
Analysis of FACS Coding. A 2 × 2 × 2 multivariate MANOVA was conducted as the design used repeated
measures and multiple dependent variables. ere were three repeated measures: attention (attentive vs. not
attentive), food (food present vs. food absent) and trial (trial 1 vs. trial 2). e multiple dependent variables were
the nine AUs (see Table1 for a list of AUs used in the analysis). To reduce the data and focus on the most com-
monly occurring movements, Action Units that one third of the dogs or more never produced in any of the four
conditions were excluded from the analysis. e analysis showed no 3 - way interaction of attention, food and trial
(Wilks’ λ = 0.058, F(8,16), p = 0.24, ηp2 = 0.43) and so we excluded the factor trial from further analysis.
A 2 × 2 doubly multivariate MANOVA showed that there was no main effect of food, Wilks’ λ = 0.072,
F(9,15) = 0.65, p = 0.74, ηp2 = 0.28 and no attention x food interaction, Wilks’ λ = 0.63, F(9,15) = 95, p = 0.51,
ηp2 = 0.36. ere was a signicant main eect of attention with a large eect size, Wilks’ λ = 0.071, F(9,15) = 21.78,
p < 0.0001, ηp2 = 0.93. e origin of the signicant main eect of attention was that for all except one of the
AUs (AU 145: blink), there was more activity in the attention condition than in the no attention condition (see
Table1).
Two of the AUs, AU 101 (“eye brow raiser”, see Fig.1) and AD 19 (“tongue show”) reached signicance indi-
vidually with the main factor attention having an eect on the amount of AU101 and AD19 movements produced
(see Table1, Fig.1).
Analysis of Behavioural Coding. We also examined the other behavioural measures to see if condition
had an eect. We compared the frequency of vocalizing bouts across the four conditions using a Friedman test
(data were non normally distributed). Condition had an eect on the frequency of vocalizations produced with a
medium eect size, χ2(3, N = 23) = 17.69, p = 0.001, W = 0.26. e pattern of medians revealed that the human’s
attention increased the frequency of the vocalizations produced by the dogs (attention/food Mdn = 2.76; atten-
tion/no food Mdn = 2.65; no attention/food Mdn = 2.37; no attention/no food Mdn = 2.11). Wilcoxon pair-
wise comparisons revealed signicant dierences between attention/food vs. no attention/food, Z(23) = 2.31,
p = 0.021; attention/no food vs. no attention/food, Z(23) = 2.65p = 0.008; and attention/no food vs. no attention/
no food, Z(23) = 2.49, p = 0.013. No other combinations revealed signicant eects.
We also compared the frequency of tail wagging across the four conditions. ere was no eect of condi-
tion, χ2(3, N = 23) = 6.49, p = 0.09, although the pattern of medians was similar to that found for vocalizing
(attention/food Mdn = 2.76; attention/ no food Mdn = 2.65; no attention/food Mdn = 2.37; no attention/no food
Mdn = 2.2) with attention and food increasing the frequency of the behaviour. None of the pairwise comparisons
was signicant.
AU/AD F(df) p d ηp2M(SD) Attention M(SD) No Attention
101 102.58 (1,23) <0.0001 1.62 0.82 0.121 (0.04) 0.056 (0.04)
145 0.04 (1,23) =0.84 0.02 0.002 0.197 (0.06) 0.199 (0.09)
12 0.89 (1,23) =0.35 0.09 0.04 0.044 (0.05) 0.039 (0.06)
25 1.33 (1,23) =0.26 0.17 0.06 0.128 (0.12) 0.108 (0.11)
26 1.98 (1,23) =0.17 0.21 0.08 0.139 (0.13) 0.113 (0.12)
118 0.88 (1,23) =036 0.18 0.04 0.049 (0.07) 0.037 (0.05)
19 5.87 (1,23) =0.024 0.19 0.20 0.071 (0.07) 0.057 (0.07)
102 2.21 (1,23) =0.15 0.19 0.09 0.014 (0.02) 0.011 (0.01)
105 2.76 (1,23) =0.11 0.26 0.11 0.047 (0.05) 0.036 (0.03)
Table 1. AU unit activity as a function of attention or not for individual univariate ANOVAs Signicant results
in bold.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
3
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
We conducted similar analyses on the other behaviours (laying, sitting, standing and moving, see Table1) but
no signicant eects were revealed.
Discussion
e study has two main ndings. First, human attentional state aected the production of dogs’ facial expressions.
Dogs produced signicantly more facial expressions when the human was oriented towards them, than when the
human had her back turned to the dog. is eect was strongest for AUs, AU101 (“inner brow raiser”) and AD19
(“tongue show”). Human attentional state also aected one of the dogs other behaviours, the frequency of vocal-
izations produced. e visibility of the food, however, did not aect dogs’ facial movements and there is also no
conclusive evidence that it aected any of the dogs other behaviours. So, while dogs produce more facial expres-
sions when the human is oriented towards them and in a position to communicate, the visibility of non-social but
arousing stimulus (the food) did not alter their facial movements in the same way.
One interpretation of the dogs’ behaviour could be that dogs produce their facial expressions communica-
tively and that the dogs’ facial expressions are not just mediated by the individual’s emotional state. erefore,
dogs increase the frequency of their production dependent on the other individual’s attentional state but not
in response to being presented with a non social but arousing stimulus (the food). However, our data show an
increase of all facial movements when the human is attentive, but no evidence that dogs specically modulate
their facial movements depending on the attentional state of the human. Human attention had an eect on one
of the dogs other behaviours, the frequency of the vocalizations produced, but not their non-communicative
behaviours, such as sitting and standing. erefore, it seems that there might be a specic communicative func-
tion of this sensitivity to human attention. ere is substantial evidence supporting the importance of visibility
of the human’s eyes for dogs during communicative interactions with humans. Teglas et al. (2012) showed that
dogs do not follow the gaze of a human to a certain location unless eye contact had been established beforehand17.
Kaminski et al. (2012) showed that dogs follow a human’s communicative gestures (e.g. pointing or gazing) but
ignore a human’s actions which resembled the communicative gestures but are not intended to be communicative
and during which the humans eyes were not directed at the dog16. Here we now add to this evidence by showing
that the visibility of the human’s eyes might be important for dogs for the production of facial expressions. is
might be evidence that dogs produce facial expressions as a exible signal and that its production depends on the
attentional state of the receiver of the signal. Without taking any physiological measures it is obviously impossible
to say to what extent seeing the human’s and the human’s eyes is also arousing for the dog. But our study high-
lights that a non-social stimulus which has been proven to be arousing for dogs, does not have any eect on the
production of their facial expressions. It is however possible impossible to say whether dogs behaviour in this and
other studies is evidence for a exible understanding of another individuals perspective, hence constituting a true
understanding of another individual’s mental state, or is a rather hardwired or learnt response to seeing the face
or the eyes of another individual (see for a discussion21,22).
Interestingly there are two Action Units that stand out from the overall analysis, the AD 19 (tongue show) and
the AU101 (inner brow raiser). e tongue shows movement can potentially be associated with stress (e.g. nose
lick behavior) but could also indicate panting behavior, which dogs use for heat regulation23. However, a facial
expression ethogram based on systematic analysis using tools like DogFACS does not exists, which is why we can
only speculate. Interestingly in the general dog literature a relaxed open mouth with tongue show is sometimes
described as generally attentive, which would be an interpretation in line with the results we see here.
However, AU101 may be of greatest signicance as the response of humans to this unit may have had the
greatest inuence on selection. Waller et al. (2013) showed that dogs from a shelter that produced the AU101
more frequently were rehomed quicker18. is could be for two possible reasons. Firstly, AU101 resembles a facial
movement which in humans indicates sadness, hence potentially making humans feels more empathic towards
dogs that produce this movement more. Another possibility is that the AU101 lets the eyes of the dogs appear
Figure 1. Mean rate (±1 SD) of facial movement AU101 (Inner eye brow raise) as a function of condition.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
4
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
bigger and more infant like potentially tapping into the preference of humans for paedomorphic characteristics
and/ or humans innate tendency to respond to ostensive cues, one of which is ‘eyebrow raising’24,25. Regardless of
the exact mechanism, it seems that humans are particularly responsive to this facial movement in dogs. Increased
production of this movement in response to human attention could benet dogs in their interaction with humans,
therefore.
In conclusion, we have demonstrated that dogs’ production of facial expressions is subject to audience eects,
and can be tailored to the human attentional state suggesting some communicative function and are not simple
emotional displays based on the dogs arousal state. Facial expressions are oen considered to be an automatic,
reexive and emotionally based system3, but these data point to a more exible system (at least in domestic dogs)
combining both emotional and potentially cognitive processes.
Methods
Subjects. 24 family dogs (13 male and 11 female) of various breeds and ages (age range = 1–12 years,
Mage = 4.75, SD = 3.33) participated in the study (see Table2). e dogs were normal family dogs with a train-
ing background typical for a pet dog. Dogs were randomly selected from a database of dogs at the Max Planck
Institute for Evolutionary Anthropology in Leipzig/ Germany. e only criterion for selection was that dogs had
to be comfortable to be without their owner and comfortable with a stranger in a strange environment. Research
was non-invasive and strictly adhered to the legal requirements of Germany. e study was ethically approved by
an internal committee at the Max Planck Institute for Evolutionary Anthropology (members of the committee are
Prof. M. Tomasello, Dr. J. Call and Susanne Mauritz). e animal research complies with the “Guidelines for the
Treatment of Animals in Behavioral Research and Teaching” of the Association for the Study of Animal Behavior
(ASAB). IRB approval was not necessary because no special permission for the use of animals in purely behav-
ioural or observational studies is required in Germany (TierSchGes §7 and §8). Dogs were fed by their owners
according to their normal daily routine and not food deprived in any way. Water was available to the dogs ad
libitum. In the conditions during which food was presented to the dogs, ©Frolic was used, which is a food mainly
used by owners as a treat in between meals. Dogs received regular breaks and observations were stopped in case
the dogs showed any sign of severe stress.
Materials. Dogs were observed in a quiet room (2.85 m × 3.60 m) and to ensure that dogs did not move
around extensively so their facial expressions could be observed easily, dogs were tied with a lead (1 m) to a prede-
termined spot in the room. Dogs were situated 1 m away from the human who was standing on a predetermined
and marked spot. A video camera was placed on a stationary tripod such that the dogs’ faces were visible (see
Fig.2).
Name Breed Gender Age (Years)
Anouk Eurasier Male 1
Arik Hovawart Male 4
Bacardi Mongrel (German Shepherd mix) Female 11
Balou Schapendoes Male 11
Basma Basenji Female 2
Caja Mongrel (Doberman Mix) Female 8
Cody Mongrel Male 4
Dusky Basenji Male 2
Fefo Parson Jack Russell Male 3
Gerda Mongrel (Poodle & Labrador Mix) Female 3
Gordo Mongrel (Canario & Doberman Mix) Male 4
Guenni Whippet Male 2
Guiness Dalmation Male 5
Kendra Border Collie Female 2
Kenny Labrador Male 3
Lea Mongrel (Leonberger & German Shepherd Mix) Female 11
Luna German Shepherd Female 4
Mira Mongrel (Podenco & Magyar Vizsla Mix) Female 5
Paul Golden Retriever Male 3
Romy Mongrel (Rottweiler Mix) Male 2
Scully Border Collie Female 5
Sparky Boxer Male 2
Tina Mongrel Female 12
Wilma Mongrel (Rottweiler and Rhodesian Ridgeback Mix) Female 5
Table 2. List of the dogs included in the study with information about breed, gender and age (years).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
5
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
Procedure. Before testing started, each dog was allowed to familiarise itself with the experimenter and the
room in which testing was conducted, and was allowed to move around freely for several minutes. Aer that the
dog was attached to a lead and the experimenter positioned herself in front of the dog and behaved according to
the following four conditions:
Attentive Food. e experimenter stood facing the dog with arms pointing towards the dog and palms placed
together displaying the food (see Fig.2A).
Attentive No Food. e experimenter stood facing the dog with palms in the same position but not displaying
food (see Fig.2B).
Not Attentive Food. e experimenter had her back turned towards the dog but had her arms behind her back,
palms placed together displaying the food (see Fig.2C).
Not Attentive No Food. e experimenter had her back turned towards the dog and her arms behind her back,
palms placed together but not displaying food (see Fig.2D).
e design was a within subjects design in which each dog received all four conditions. e order of condi-
tions was counterbalanced across dogs.
During each trial the experimenter stood still and did not respond to any of the dog’s behaviours. e exper-
imenter looked at a predetermined spot at the opposite wall and did not actively seek eye contact with the dog
when she was oriented towards the dog. Aer 2 minutes the trial ended and the human briey interacted with the
dog before she then changed her position according to the condition presented in the next trial.
Each dog received two trials per condition, summing up to 8 trials altogether. Trials were split in two sessions
to prevent fatigue and sessions were presented on two dierent days, with a break of up to four days between
sessions. e order of condition was counterbalanced across dogs.
FACS Coding. Coding of the facial movements was based on the DogFACS manual (Waller et al. 2013: www.
dogfacs.com). DogFACS (Waller et al., 2013) is based on e Facial Action Coding System (FACS), an anatom-
ically based facial expression coding system rst developed for humans (Ekman & Friesen, 1978). It identies
observable facial changes associated with underlying muscle movement (Action units, AUs) allowing an objec-
tive, reliable and standardized measurement of facial movements. DogFACS was used to identify the facial move-
ments produced during each condition. A certied DogFACS coder (JH) coded the frequency and duration of
Figure 2. Experimenter’s position in the (A) Attentive Food (B) Attentive No food (C) Not attentive Food (D)
Not attentive no Food condition.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
6
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
the dierent action units. Action Units that one third of the dogs or more never produced in any of the four con-
ditions were excluded from the analysis (for a list of all possible AUs, see Waller et al. 2013 or www.dogfacs.com).
is le 9 AUs (Action Units), ADs (Action Descriptors) and EADs (Ear Action Descriptors) that were included
in the analysis (see Table3 for a description).
Behavioural Coding. In addition to the FACS coding we also coded the general behaviour of the dogs to
see if we could identify any changes across conditions. We coded lying, sitting, standing, moving, tail-wagging,
yawning, nose licking and vocalizing (see Table3 for a list of the coded behaviours and their denitions).
Reliability coding. For the FACS coding JH (a trained FACS coder) did intra-rater reliability coding of 20%
of the original material (frequency of AUs produced) 1 year aer her original coding. Reliability was excellent
with all ric > 0.88, N = 40, p < 0.001. For the behavioural data (frequency of behaviours), a second coder unaware
of the research question coded 20% of the original material for the dierent behaviours. Reliability was excellent
for each of the behaviours (Stand: ric = 0.099, N = 40, p < 0.0001, Sit: ric = 0.99, N = 40, p < 0.0001 Move: r = 0.98,
N = 40, p < 0.0001, Lying: ric = 1, N = 40, p < 0.0001, Tail wagging: ric = 0.99, N = 40, p < 0.0001Vocalize: r = 0.99,
N = 40, p < 0.0001).
Data Availability Statement. All data will be made available.
References
1. Diogo, ., Wood, B. A., Aziz, M. A. & Burrows, A. On the origin, homologies and evolution of primate facial muscles, with a
particular focus on hominoids and a suggested unifying nomenclature for the facial muscles of the Mammalia. J. Anat. 215, 300–319
(2009).
2. Darwin, C. e expression of emotion in animals and man. Lond. Methuen1877 Biogr. Sketch Infant. Mind 2, 285–294 (1872).
3. Tomasello, M. Origins of Human Communication (MIT Press, 2010).
4. Poss, S. ., uhar, C., Stoinsi, T. S. & Hopins, W. D. Differential use of attentional and visual communicative signaling by
orangutans (Pongo pygmaeus) and gorillas (Gorilla gorilla) in response to the attentional status of a human. Am. J. Primatol. 68,
978–992 (2006).
5. Leavens, D. A., ussell, J. L. & Hopins, W. D. Multimodal communication by captive chimpanzees (Pan troglodytes). Anim. Cogn.
13, 33–40 (2010).
6. Waller, B. M., Caeiro, C. C. & Davila-oss, M. Orangutans modify facial displays depending on recipient attention. PeerJ 3, e827
(2015).
7. Scheider, L., Waller, B. M., Oña, L., Burrows, A. M. & Liebal, . Social use of facial expressions in hylobatids. PloS One 11, e0151733
(2016).
8. Demuru, E., Ferrari, P. F. & Palagi, E. Emotionality and intentionality in bonobo playful communication. Anim. Cogn. 18, 333–344
(2015).
9. almann, O. et al. Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs. Science 342,
871–874 (2013).
10. Marshall-Pescini, S., aminsi, J., aminsi, J. & Marshall-Pescini, S. e social dog: History and evolution. Soc. Dog Behav. Cogn.
3–33 (2014).
11. Hare, B. et al. e domestication hypothesis for dogs’ sills with human communication: a response to and. Anim. Behav. 79, e1–e6
(2010).
12. Udell, M. A., Dorey, N. . & Wynne, C. D. Wolves outperform dogs in following human social cues. Anim. Behav. 76, 1767–1773
(2008).
13. Xitco Jr, M. J., Gory, J. D. & uczaj II, S. A. Dolphin pointing is lined to the attentional behavior of a receiver. Anim. Cogn. 7,
231–238 (2004).
Behaviour Denition
AU 101(1) Inner brow raiser. Liing of the inner brow region performed by the frontalis muscle.
AU 145(1) Blink: the relaxation of the levator palpebrae muscle and contraction of the orbicularis occuli circular muscle act to move the upper and lower
eyelid, closing the eye. e retractor anguli occuli lateralis pulls the outer corner of the eye caudally, aiding in closing the eye.
AU 12(1) Lip corner puller. e zygomaticus pulling the lip corners towards the ears.
AU 25(1) Lips part.
AU 26(1) Jaw drop.
AU 118(1) Lip pucker. e buccinators and orbicularis oris muscles act to push the lip corners rostrally, towards a medial point.
AD 19(1) Tongue show. e tongue is shown and it reaches at least the inner lower lip.
EAD 102(1) Ears adductor. e ears are adducted and the base of both pinnas becomes closer together by being pulled towards the head midline.
EAD 105(1) Ears downward. e ears are pulled ventrally, laterally.
Laying(2) e dog’s legs laid at on the ground while the head could rest on the ground, legs, or remain o the ground.
Sitting(2) e dog’s forelegs were extended and perpendicular to the ground while the hind legs were exed with the tarsus resting at on the ground.
Standing(2) All legs were extended and perpendicular to the ground.
Moving(2) e dog displaced its body from one location to another by alternately moving its legs more than two steps.
Tail-wagging e dog’s tail is extended and moves side to side in quick motion.
Vocalising Any sound emitted from the dog.
Table 3. Facial movements (DogFACS: Action Units, AUs, Action Descriptors, ADs and Ear Action Descriptors,
EADs) and general behaviours coded. Listed denitions were partly obtained from (1)Waller et al. 2013 and (2)Call
et al. (2003).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
7
Scientific RepoRts | 7: 12914 | DOI:10.1038/s41598-017-12781-x
14. D’Entremont, B., Hains, S. & Muir, D. A demonstration of gaze following in 3-to 6-month-olds. Infant Behav. Dev. 20, 569–572
(1997).
15. S chwab, C. & Huber, L. Obey or not obey? Dogs (Canis familiaris) behave dierently in response to attentional states of their owners.
J. Comp. Psychol. 120, 169 (2006).
16. Moll, H., Carpenter, M. & Tomasello, M. Fourteen-month-olds now what others experience only in joint engagement. Dev. Sci. 10,
826–835 (2007).
17. Téglás, E., Gergely, A., upán, ., Milósi, Á. & Topál, J. Dogs’ gaze following is tuned to human communicative signals. Curr. Biol.
22, 209–212 (2012).
18. Waller, B. M. et al. Paedomorphic facial expressions give dogs a selective advantage. PLoS One 8, e82686 (2013).
19. Archer, J. & Monton, S. Preferences for infant facial features in pet dogs and cats. Ethology 117, 217–226 (2011).
20. Travain, T. et al. How good is this food? A study on dogs’ emotional responses to a potentially pleasant event using infrared
thermography. Physiol. Behav. 159, 80–87 (2016).
21. Premac, D. & Woodru, G. Does the chimpanzee have a theory of mind? Behav. Brain Sci. 1, 515–526 (1978).
22. Penn, D. C. & Povinelli, D. J. On the lac of evidence that non-human animals possess anything remotely resembling a ‘theory of
mind’. Philos. Trans. R. Soc. Lond. B Biol. Sci. 362, 731–744 (2007).
23. Goldberg, M. B., Langman, V. A. & Taylor, C. . Panting in dogs: paths of air ow in response to heat and exercise. Respir. Physiol.
43, 327–338 (1981).
24. Csibra, G. & Gergely, G. Natural pedagogy. Trends Cogn. Sci. 13, 148–153 (2009).
25. Gergely, G. & Csibra, G. Teleological reasoning in infancy: e naıve theory of rational action. Trends Cogn. Sci. 7, 287–292 (2003).
Acknowledgements
We thank the owners for volunteering their dogs for our research. We also thank Katrin Schumann from the
Max Planck Institute for Evolutionary Anthropology in Leipzig for their help with data collection. We thank
Becky Spooner and Hoi-Lam Jim for help with coding some of the behavioural data of the dogs and for reliability
coding. We thank Jerome Micheletta for help with Figure 2.
Author Contributions
J.K., P.M. and B.W. wrote the main manuscript text and J.H. coded the video materials and wrote parts of the
method section. All authors reviewed the manuscript.
Additional Information
Competing Interests: e authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2017
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com