Content uploaded by Anouck Haverbeke
Author content
All content in this area was uploaded by Anouck Haverbeke on Oct 19, 2018
Content may be subject to copyright.
Behavioral and cortisol responses of shelter dogs
to a cognitive bias test after olfactory enrichment
with essential oils
Stefania Uccheddu1, Chiara Mariti2, Adinda Sannen3, Hilde Vervaecke3,
Heidi Arnout3,4, Jara Gutierrez Rufo2, Angelo Gazzano2, Anouck Haverbeke1,*
1 Vet Ethology, Leemveldstraat44, 3090 Overijse, Belgium
2 Department of Veterinary Sciences, University of Pisa, Viale delle Piagge 2, 56124 Pisa, Italy
3 Odisee University College, Agro- & Biotechnology, Ethology & Animal Welfare, Hospitaalstraat 23,
9100 Sint-Niklaas, Belgium
4 University of Antwerp, Departmentof Engineering Management, Prinsstraat 13, 2000 Antwerp, Belgium
Abstract: A shelter environment tends to present dierent types of stressors dogs need to cope with. Recent work
has shown that olfactory enrichment with essential oils might be able to modify the aective states of certain species
(dogs, cats, horses, zoo animals…). In these studies, the welfare measurements included physiological indicators, such
as corticosteroid levels, and/or behaviors related to chronic stress. e olfactory eects of 9 essential oils (Cananga od-
orata,Cistus ladaniferus, Citrus aurantium, Cupressus sempervirens, Juniperus communis var. montana, Lavandula an-
gustifolia, Laurus nobilis, Litsea citrata, Pelargonium graveolens) and a blend of these oils were explored on a cognitive
bias test, cortisol levels and the behaviors of 110 shelter dogs (n = 10 dogs within each group). Olfactory enrichment
with the blend resulted in a reduced latency to the ambiguous cue, indicating a more optimistic bias and improved
welfare. e results of this study suggest that olfactory enrichment with essential oils can have specic eects on the
aective states and behaviors of shelter dogs, and could therefore be useful for shelter management. In addition, as not
all of the essential oils tested individually were eective, more research should be conducted to better understand the
eects of each individual essential oils on dogs.
Key Words: behavior; cognitive bias; cortisol; dogs; essential oils; olfactory enrichment; shelter; welfare.
* Corresponding Author: anouck.haverbeke@vetethology.be
Introduction
A shelter environment tends to present dierent types of stressors dogs need to cope with: social
stressors (reduced intraspecic and/or interspecic social contacts), environmental stressors
(restraint for medical procedures, separation from a caretaker or handler) or psychogenic stressors
(separation anxiety, use of aversive training methods by a previous owner/lack of ethological
knowledge in caretakers). Moreover, stressors are known to cause activation of metabolic and
endocrine responses in sheltered animals (Titulaer et al., 2013).
Recent work has shown that essential oils might be able to modify the aective states of
certain species (dogs, cats, horses and zoo animals: Wells, 2004; Graham et al., 2005; Ferguson
et al., 2013; Wells & Egli, 2015; Binks et al., 2018). In these studies, the welfare measurements
included physiological indicators, such as corticosteroid levels (Beerda et al., 1998) or behaviors
related to chronic stress, such as repetitive behaviors, nosing, paw-liing, increased locomotion,
displacement behavior or excessive drinking (Beerda et al., 1998; Haverbeke et al., 2008).
However, interpretation of these indicators can be dicult (Titulaer et al., 2013). erefore
the detection of a cognitive bias might be a complementary solution. A recent and innovative
approach utilizes the inuence of aective states on the interpretation of current experience.
e resulting aect-induced cognitive biases can be measured (Mendl et al., 2009) through
cognitive bias tests as indicators of the animal’s psychological well-being (Mendl et al., 2009; Paul
et al., 2005). A cognitive bias test in this context refers to the propensity of a subject to show
Dog Behavior, 2-2018, pp. 1-14
doi 10.4454/db.v4i2.87
Submitted, 07/29/2018
Accepted, 09/12/2018
behavior indicating the anticipation of either relatively positive or relatively negative outcomes in
response to aectively ambiguous stimuli (Mendl et al., 2009). Changes in cognitive bias reect an
individual’s experience of positive and negative events and thus its aective valence and welfare
(Mendl et al., 2010). e eects of environmental enrichment have been already tested through
cognitive bias test in dierent species such as rats (Brydges et al., 2011), pigs (Douglas et al., 2012)
and European starling (Bateson & Matheson, 2007).
Several studies have found correlations between cognitive biases and aective states in a wide
range of species, including mammals (Mendl et al., 2009; Doyle et al., 2010) and birds (Matheson
et al., 2008; Salmeto et al., 2011). e aim of the current study was to assess whether olfactory
enrichment through essential oils inuences the aective states of sheltered dogs. To do that, the
possible eects of 9 dierent of essential oils (Cananga odorata, Cistus ladaniferus, Citrus auran-
tium, Cupressus sempervirens, Juniperus communis var. montana, Lavandula angustifolia, Laurus
nobilis, Litsea citrata, Pelargonium graveolens) and a blend of these oils on a cognitive bias test,
cortisol levels and behavior of 110 shelter dogs were explored.
Materials and Methods
Participants
One hundred ten dogs ranging in age from 1 to 10 years, of both sexes, and of either pure or
mixed breed, were enrolled in the study and randomly allocated to one of 11 dierent groups
(Table 1). e dogs lived in groups of three in kennels with an indoor section measuring 1.5
meters x1.5 meters and an outdoor run measuring 1.5 meters x 2 meters, joined by a metal door
operated by sta. Water was available ad libitum, and food was provided twice per day, at 8 am
and 6 pm.
Dogs were selected based on the following criteria: (a) success at the training phase, (b) no
previous diagnosis of anxiety or aggressive behavior, (c) some socialization prerequisites, such as
the ability to deal with people without fear, (d) the veterinarian’s agreement and (e) ability to walk
on leash (f) good medical health.
Table 1. Description of the study protocol.
Group Number
of dogs
Pre-test
training Cognitive test 1 Exposure to collar
for 3 hours Cognitive test 2
1 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
No addition
(control group)
Aer exposure to collar and aer
collection of saliva at T1
2 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
e blend Aer exposure to collar and aer
collection of saliva at T1
3 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Litsea citrata Aer exposure to collar and aer
collection of saliva at T1
4 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Cupressus
sempervirens
Aer exposure to collar and aer
collection of saliva at T1
5 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Citrus aurantium Aer exposure to collar and aer
collection of saliva at T1
6 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Pelargonium
graveolens
Aer exposure to collar and aer
collection of saliva at T1
2
3
7 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Lavandula
angustifolia
Aer exposure to collar and aer
collection of saliva at T1
8 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Cananga odorata Aer exposure to collar and aer
collection of saliva at T1
9 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Juniperus
communis var.
Montana
Aer exposure to collar and aer
collection of saliva at T1
10 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Cistus ladaniferus Aer exposure to collar and aer
collection of saliva at T1
11 10 Ye s
Before exposure to collar
and aer collection of
saliva at T0
Laurus nobilis Aer exposure to collar and aer
collection of saliva at T1
Olfactory enrichment
A blend of 9 essential oils (Arhomani, Belgium) and each separate oil of the blend (Flora s.r.l.,
Pisa, Italy) were tested, for a total of 10 treatments and a control group (Table 1). Essential oils were
diused through a cotton collar worn by the dogs for 3 hours before starting the second cognitive
test procedure (see below). e collar, just before being applied to the dog, had 1 drop of an indi-
vidual oil or of the blend added to it. e control group, as the experimental groups, wore a cotton
collar for 3 hours but without any oils or other addition.
During this part of the experiment, the dogs stayed in the pen where they routinely spent time.
Dogs were allowed to mix in the same pen only if they were allocated to the same essential oil
group. In order to avoid odour contamination, there was a distance of 500m between the dierent
pens.
Test protocol
Cognitive test
All dogs of the 11 groups were subjected to two cognitive bias (CB) tests (modied from
(Mendl et al., 2010), one prior to (CB 1) and one aer essential oil exposure (CB 2). CB 2 was
performed 3 hours aer T0. To avoid more stress caused by a dierent routine in the shelter,
we could not control for order eects. All the dogs were tested on the second cognitive bias test
following olfactory enrichment.e protocol of (Mendl et al., 2010) and (Owczarczak-Garstecka
& Burman, 2016) was modied in this study based on a pilot study we carried out, in which we
observed that dogs in the shelter were unable to maintain attention during the original cognitive
test as proposed in (Mendl et al., 2010) and (Owczarczak-Garstecka & Burman, 2016). For this
reason, we used a shortened version. During the training session, all dogs received a minimum
of 8 training trials instead of 15. During the test, we used just one ambiguous location instead of
three ambiguous locations. e test phase involved 6 trials (instead of the 32 proposed by (Mendl
et al., 2010).
In addition, during the pilot study, we realized that shelter dogs were much more interested
in humans than in food, so we changed the original protocol by having the researcher behind the
camera rather than behind the bowl put on the ground, in order to avoid the dog choosing that
bowl for its closeness with a person (the researcher).
Training and cognitive tests were performed with each dog enrolled in the experiment
individually led to a test area (6 meters x 6 meters) within the shelter, the same for all sessions and
all dogs. e setting is described in gure 1. e bowl was placed at one of three predetermined
locations (two during the training) 4 meters in front of the dog’s xed starting position. e
4
latency to reach the bowl, was dened as the time elapsed between release from the lead and the
dog putting its head into the bowl, or touching the rim of the bowl with its nose (Mendl et al.,
2010). CB tests were video recorded and then analysed as described below.
Figure 1. Experimental setting.
Training
Dogs were rst trained to associate a certain location with a reward. During the training, the
distance between the two bowls (Positive and Negative) was 1meter.When the bowl was placed at
the ‘positive’ location (P) on one side of the test area, it contained food, and when it was placed
at the ‘negative’ location (N) on the opposite side of the test area, it was empty. Two visually
identical bowls were used for rewarded (P) and non-rewarded (N) locations, and both bowls had
a piece of food taped to their bottom sides that were inaccessible to the dogs to control for odour
cues. Training was complete when the dogs reached a pre-set criterion, that is, when the dog ran
to the positive location faster than to the negative one twice consecutively.
Each training session started aer a 10-minute period of habituation with the researchers in the
experimental area (Figure 1). e dog was put on a lead and held by one of the researchers behind
a barrier, while the other researcher stood at the far end of the room and baited (or did not bait,
depending on trial type) a food bowl with 50 gr of commercial dog food. e dog was released
to approach the bowl. Each dog received at least 8 training trials conducted so that no location
was repeated more than twice. Each training session started with two positive (rewarded) trials
to encourage participation, followed by two negative (non-rewarded) trials. e remaining trials
were randomly assigned to be rewarded or non-rewarded. e latency to reach the bowl, dened
as the time elapsed between release from the lead and the dog putting its head into the bowl or
touching the rim of the bowl with its nose, was recorded for each trial using a stopwatch. e
maximum time allowed per trial was 30 seconds. If the dog did not reach the food bowl within
that time, the maximum time was scored.
5
Test
When the training was completed, the test started. Each dog was presented a food bowl in
three locations, positive (P), negative (N) and intermediate (M).e Middle bowl was located
between Positive and Negative bowl. e distance between Positive (or Negative) and Middle
Bowl was 50 cm.e bowl was presented in each location twice (P1, M1, N1 and P2, M2, N2) but
in dierent order. e accessible food was only present in the positive location (P). Negative (N)
and Intermediate (M) locations remained empty but with olfactory control cues. All the tests were
videorecorded. Since in each CB the locations were tested twice, we used mean values for each
location in CB 1 and in CB 2 in further analyses.
Behavioral observations
e observations of dog behaviors were carried out on the videos recorded during the whole
test. Each dog was observed using a continuous sampling method.
e behavioral analysis was conducted using the ethogram reported in Table 2 and 3 (Haverbeke
et al., 2008). Depending on the type of behavior, either the duration (in seconds) or the number
of occurrences was recorded.
Table 2. Behaviors scored in terms of number of occurrences.
Behavior Description
Oral behaviors:
Yawning Mouth open to apparent fullest extent while eyes are closed
Non-directed licking
Snout licking
Tongue out, the tip of the tongue is briey extended
Part of the tongue is shown and moved along the upper lip
No oral behaviors
Paw liing Fore paw lied into a position of approximately 45°
Urinatingsquat Urinating by squatting while keeping both hind limbs on cage oor
Urinating, limbraised Urinating while raising one hind limb
Defecating Excreting the contents of the bowels
Table 3. Behaviors analysed in terms of duration (seconds).
Behaviors
Repetitive or stereotypicbehavior
Pacing Immediately repeating a path just taken and continuing in the repetition in circles, in a
gure eight pattern or fence/wall-line running
Circling Continuous walking in short circles, apparently chasing its tail or hind limbs
Other behaviors Manipulating environment (Stereotypic interactions with elements from the
environment, such as digging (scratching the oor with the forepaws in a way that is
similar to how dogs dig holes), oor licking (licking the oor with the tongue)), Auto
grooming
Notseen Unable to determine behavior of the dog owing to darkness or the position of the dog
Miscellaneousoralbehaviors
Barking loud, rough noise
Roaring loud, deep sound
Growling low, rough sound
Whining long, high sound
Yelping sudden, short, high sound
Panting Increased frequency of inhalation and exhalation oen in combination with the opening
of the mouth
Teethclapping Making short loud noise by hitting teeth together
6
Notseen Unable to determine the behavior of the dog owing to darkness or the position of the
dog
Locomotive states
Prone, head down Trunk of body on oor, chin or side of head in contact with the oor, paws or limbs
Prone, head up Trunk of body on the oor, no part of the head in contact with the paws
Sitting Only hindquarters and front paws in contact with the oor
Standing Upright with at least three paws in contact with the oor without any walk
Walking Takes at least one step, shiing body position
Highly active Any motion across oor faster than a walk, including trotting and jumping
Changing from one state of locomotion to another
Notseen Unable to determine behavior of dog owing to darkness or the position of the dog
Postures
High e breed specic posture as shown by dogs under neutral conditions, but in addition
the tail is positioned higher or the position of the head is elevated, and the ears are
pointed forwards, or the animal is standing extremely erect
Neutral e breed posture shown by dogs under neutral conditions
Halow Two or more of the following three features are displayed: a lowered position of the tail
(compared to the neutral posture), a backward position of the ears and bent legs
Low e position of the tail is lowered, the ears are positioned backwards, and the legs are
bent
Verylow Low posture, but now the tail is curled forward between the hind legs
Notseen Unable to determine the behavior of the dog owing to darkness or the position of the
dog
Cortisol
Saliva samples for the assessment of plasma cortisol concentrations were collected, at the same
time in the day, before the addition of the oils at T0 (to identify the basal cortisol levels) and at T1,
i.e. aer 3 hours exposure to the collar for all groups, including the control one. Collection was
always carried out before the cognitive bias tests at T0 and T1. Saliva samples were collected using
Salivette Cortisol code blue(Sarstedt, Nümbrecht, Germany) and stored at-20 °C until they were
further processed using a commercial ELISA kit (Diametra, Milano, Italy).
Statistical analysis
e statistical analysis was performed using IBM SPSS Statistics for Windows, version 22.0
(Armonk, NY: IBM Corp). For each of the oils under study, the dierence in the variables
measured before and aer exposure was tested using a Wilcoxon signed-rank test. is paired
dierence test was used because each subject is measured twice, resulting inpairsof observations.
is reduces the eect of confounders like individual dierences (e.g. in pace length or in interest
in food) between dogs.e test statistics (sum of positive ranks) as well as the two-sided p-values
are reported in the results below. P values ≤ 0.05 were deemed statistically signicant.
We additionally tested for T0 as well as T1whether the dogs’ responses during the cognitive
bias tests were appropriate (i.e. dogs were slower to approach the ‘negative’ location N when
compared to the ‘positive’ location P) by using a one-sided paired t-test comparing latency to
approach N versus latency to approach P. Statistical p values ≤ 0.05 were deemed statistically
signicant.
7
Results
Cognitive test
We explored the dogs’ latency to approach P and N, just to make sure dogs’ response to the CB
test 1 (before exposure) and CB test 2 (aer exposure) was appropriate (i.e. animals were slower
to approach N than P). e results are reported in Table 4.
Table 4. Statistical results of the comparison between latency to reach positive and negative locations be-
fore exposure and after exposure to essential oils (CB 1: Cognitive test before exposure; CB 2: Cognitive test
after exposure).
Mean (seconds) N Standard. Deviation Standard Error Mean
CB 1 Latency before
exposure P location 19.32 110 17.01 1.62
Latency before
exposure N location 24.18 110 27.28 2.60
CB 2 Latency aer
exposure P location 2.64 110 0.48 0.05
Latency aer
exposure N location 13.81 110 13.53 1.29
e analysis revealed a significant eect of the blend “e blend” in reducing the latency
to reach the intermediate position (test statistic=3; n=10; p=0.039). We also observed a trend
towards reducing the latency to reach the intermediate position (test statistic=5; n=10; p=0.078)
for Litsea citrata oil (Table 5).
Table 5. Latency (mean ± Standard Deviation in seconds before and after 3 hours of exposure) and cortisol
values (mean ± Standard Deviation in ng/ml before and after 3 hours of exposure) to each essential oil or
after 3 hours without any exposure in the control group (P < 0.05, *).
Before
exposure
Aer 3 hours
of exposure
(T1)
Statistical
results
Before
exposure
Aer 3 hours
of exposure
(T1)
Statistical
results
Latency value
(seconds)
Cortisol value
(ng/ml)
Control group
(no exposure) 20.60±11.00 15.95±10.98 P=0.38 2.406 ± 0.30* 1.762 ± 0.435 P=0.03*
Cananga odorata 18.65±7.84 16.92±9.35 P=0.84 1.923 ± 0.70 1.512 ± 0.111 P=0.08
Cistus
ladaniferus 18.77±11.78 14.98±11.93 P=0.54 1.538 ± 0.22 1.424 ± 0.132 P=0.18
Citrus aurantium 17.22±12.48 11.21±11.98 P=0.35 1.642 ± 0.21 1.507 ± 0.196 P=0.43
Cupres
sussempervirens 24.47±7.14 18.46±11.32 P=0.19 1.766 ± 0.58 2.175 ± 0.424 P=0.12
Juniperus
communis var.
Montana
21.93±9.30 14.06±13.71 P=0.20 1.397 ± 0.30 1.497 ± 0.364 P=0.74
Laurus nobilis 20.80 ±11.35 15.45±9.59 P=0.10 1.082 ± 0.45 1.435 ± 0.198 P=0.14
Lavandula
angustifolia 22.19±9.60 16.70±14.08 P=0.29 1.821 ± 0.39*1.549 ± 0.245 P=0.03*
Litsea citrate 21.97±9.34 14.70±10.39 P=0.078 1.467 ± 0.30 1.919 ± 0.313 P=0.078
Pelargonium
graveolens 20.74±9.58 15.48±10.70 P=0.10 1.287 ± 0.33 1.596 ± 0.504 P=0.10
e blend 23.83±9.80 13.46±11.28 P=0.039* 1.557 ± 0.49 1.316 ± 0.119 P=0.25
8
Behavioral observations
Only the olfactory enrichment with Laurus nobilis induced a signicantly longer duration of
high posture among these dogs (test statistic=26.5; n=10; p=0.047).
e analysis revealed non-signicant trends for dierent oils: Cananga odorata reduced the
“nosing” time (test statistic=9; n=10; p=0.064), Citrus aurantium (test statistic=46; n=10; p=0.064)
and Cupressus sempervirens (test statistic=39; n=10; p=0.055) increased the time spent in “tail
wagging”, and Pelargonium graveolens (test statistic=3; n=10; p=0.078) reduced the time spent
in “non-oral stress behaviors” (circling, pacing, manipulation of environment, autogrooming).
Cortisol
Olfactory enrichment with Lavandula angustifolia induced a signicant reduction in saliva
cortisol levels (test statistic=3; n=8; p=0.039). A similar signicant reduction was also found in the
control group (test statistic= 0; n=6; p=0.031) (Table 5).
Discussion
Cognitive test
In the present study, authors applied a cognitive test to evaluate the eectiveness of olfactory
enrichment with essential oils in reducing the level of stress in sheltered dogs. Olfactory enrichment
with the blend of oils resulted in a reduced latency to the ambiguous cue in the cognitive test,
indicating a more optimistic bias and, consequently, an improved welfare (Mendl et al., 2010).
ese results provide support for the idea that the interactions between compounds oen result in
biological activity that is greater than the activity of the isolated compounds(Galindo et al., 2010).
Many domestic dogs are kept in rescue and rehoming shelters which are frequently stressful
and impoverished environments. Dog’s welfare is oen compromised within these environments
and there is a need to determine new practical and eective methods to improve the welfare
of these kenneled dogs (Binks et al., 2018).e development of objective methods to assess the
aective states of non-human animals is a crucial step in improving animal welfare (e.g. Dawkins,
2008.). Mendl (Mendl et al., 2009) enumerated several potential advantages of the cognitive bias
test, including the ability to make a priori predictions for dierent species: mammals (Mendl
et al., 2009; Doyle et al., 2010), birds (Matheson et al., 2008; Salmeto et al., 2011) and insects
(Salmeto et al., 2011). Douglas et al. (2012) support the hypothesis that an enriched environment
induces a more optimistic cognitive bias indicative of a more positive aective state and better
welfare in pigs.
Negative eects from inadequate environmental manipulations have been investigated by
several researchers. Environmental manipulations chosen to induce negative eect produce
pessimistic cognitive biases in animals’ responses to ambiguous stimuli (reviewed in (Mendl et
al., 2009)). Rats show pessimistic responses when housed in impoverished cages but switch to
optimistic responses when moved to enriched cages (Brydges et al., 2011).
To our knowledge, this is the rst time that a cognitive bias test has been applied to assess the
eect of olfactory enrichment with essential oils. Although this is a rst study on this topic and
the number of dogs tested in each experimental group was relatively low, it is remarkable that
the statistical analysis revealed some signicant dierences. In particular, the results regarding
the blend of oils are in line with previous studies that reported improved optimism through
environmental enrichment (Douglas et al., 2012) in pigs.
However, we should take into account that the medium latency for reaching the positive location
in the cognitive bias test 2 (P2- aer exposure) is signicantly lower than the medium latency
9
for reaching the positive location in test 1 (P1- before exposure). So although the dogs respond
appropriate to each CB-test (i.e. approaching N signicantly slower than P), they approach both
P and N signicantly faster during CB2 than during CB1, which might suggest some eagerness to
participate in the test.
Behavioral observations
e results of the present study indicate that olfactory enrichment with Laurus nobilis induced
high posture among dogs. In volatiles, Laurus nobilis has been reported to improve vigilance
performance in a discrimination task (Matsubara et al., 2011), which could be interpreted
positively as increased self-condence (Fatjó et al., 2007) or alternatively as a heightened alertness
due to a negative state. In humans, a high individual variability in response to olfactory exposure
to 1,8-cineol (major component of Laurus nobilis), jasmine absolute ether, linalyl acetate and
peppermint essential oil has been observed on vigilance (Heuberger & Ilmberger, 2010). In the
present study, dog’s high posture aer olfactory exposure with Laurus nobilisis not accompanied
by other signs of alertness and can therefore be interpreted as a sign of improved self-condence
in line with Haverbeke et al. (2008).
Some essential oils had a slight eect on behavior. Olfactory enrichment with Cananga odorata
reduced the “nosing” time. is could indicate a decrease of stress among the dogs (Beerda et al.,
1998). In line with these interpretations, Hongratanaworakit and Buchbauer (2004) showed that
in humans Cananga odorata decreases blood pressure and pulse rate and increases subjective
attentiveness and alertness. Olfactory enrichment with Citrus aurantium and Cupressus
sempervirens increased the time spent in “tail wagging”. Tail wagging can be seen in the interactive
social context or to facilitate interaction and could have ambivalent interpretations going from an
increase to a decrease in condence in dogs (Gasci et al., 2005). In the present study, the exposure
of dogs to Citrus aurantium and Cupressus sempervirens are not accompanied by other changes.
erefore it is likely that in this study tail wagging is a sign of relaxation. is is in line with
previous results that have demonstrated anti-anxiety eects of both Citrus aurantium (in humans
(Akhlaghi et al., 2011; Carvalho-Freitas et al., 2002; De Moraes Pultrini et al., 2006) and rats (Leite
et al., 2008) and Cupressus sempervirens(in humans Bouguenoun et al., 2006).
Olfactory enrichment with Pelargonium graveolens reduced the time spent in “non-oral
stress behaviors” (circling, pacing, manipulation of environment, autogrooming). is observed
reduction of stress behaviors (Haverbeke et al., 2008) is in line with the ndings of Rashidi Fakari
et al., 2015, who observed an anxiolytic and sedative eect of Pelargonium graveolens in humans.
Cortisol
e observed reduction in saliva cortisol with Lavandula angustifoliais in line with Atsumi
& Tonosaki who have observed a decrease of salivary cortisol level on humans aer smelling
lavender essential oil (Atsumi & Tonosaki, 2007). In addition, a previous study using olfactory
enrichment with Lavandula angustifolia on sheltered dogs showed a change in dogs’ activities
(resting time) suggestive of relaxation (Graham et al., 2005).
We also observed a reduction of cortisol levels in the control group. is nding is in line with
previous research in dogs (Shiverdecker et al., 2013; Cobb et al., 2016). One possible explanation
is that, the mere application of a cognitive test can result in a stress relieving factor, being a sort
of cognitive enrichment for sheltered dogs. However, this does not explain why the cognitive test
with essential oil exposure had no eect on cortisol levels except in the Lavandula angustifolia
group. Another explanation is that essential oils (except Lavandula angustifolia) has increased
neophobia (i.e. the fear of novelty, which can be sometimes observed in captive animals that
have received little or no previous novel sensory stimulation (Mason et al., 1991) as observed in
Goeldi’s monkeys exposed to peppermint oil (Boon, 2003) and in a young tiger exposed to catnip
10
(Todd, 2015). However, as dogs belong to a species who tends to be very neophilic (Kaulfuss
& Mills, 2008) this explanation might probably not be considered for the canine species.As the
interpretation remains open, further studies are required in order to demonstrate through a more
detailed and rigorous analysis the eects of Lavandula angustifolia essential oil on cortisol levels
versus the eects of the other essential oils.
Although saliva collection was carried out at dierent times of the day, it is unlikely that
the dierences we observed were inuenced by this. In fact, previous research has not found
a circadian rhythm in the HPA (Hypothalamic Pituitary Adrenal) activity of dogs: neither in
laboratory dogs at 30 minutes intervals over a period of 28 hours (Takahashi et al., 1981) nor
at 20 minutes intervals over a period of 25 hours (Kemppainen & Sartin, 1994), nor in working
dogs exposed to defense training and trailing tasks at 90-180 minutes intervals over a period of 24
hours (Kolevska et al., 2003).
e saliva cortisol collected aer T1 could not have reected an earlier emotional state (pre-
olfactory enrichment), because cortisol concentrations rise approximately 20 minutes aer a dog
encounters a stressor (Vincent & Michell, 1992). Moreover, previous authors (Kirschbaum &
Hellhammer, 2000) have shown that changes in plasma and salivary cortisol levels are closely
synchronized: aer injections of cortisol, salivary levels increase within 1 minute and peak
concentrations in blood are seen 2-3 minutes later in saliva.
Some methodological limitations have been encountered during this study. Firstly, we used
a short version of the cognitive bias test because the sheltered dogs were not able to perform
the longer test (author’s observations in an unpublished, pilot study). eir limited performance
might be due to the fact that these dogs were not accustomed to be involved in cognitive activities
in their actual environment (presence of physical and social stimuli).
Secondly, being a study carried out in the eld and not in a laboratory setting, many factors
could not be controlled. For instance, there is a potential risk of olfactory confounding eects.
However, in order to reduce the risk of crossed stimulation among dierent essential oils groups,
a distance of 500 meters from one pen to another was set. Further, in a shelter environment the
quantity of olfactory stimuli is high and similar for all dogs. us a possible eect of odours other
than essential oils should be equally distributed for all dogs, which is not the case in the present
study. Lastly, even if we might consider any olfactory confounding eect, the main olfactory
eect should still remain the one obtained by the tested essential oil as it is the most proximate
odour from the dog’s nose. In order to conrm our results, further research should investigate the
maximum or minimum distance necessary to create an olfactory eect with essential oils.
irdly, the ndings should be interpreted with caution because it is possible that the dogs’
behavior was inuenced by a learning eect and a decreased interest because the cognitive bias
test was repeated twice. However, each time that a cognitive bias test is being used, information
processing, including attention, learning, memory and decision-making is being addressed
(Mendl et al., 2009).
Fourthly, in our study all the dogs were tested on the second cognitive bias test following
olfactory enrichment. Unfortunately, we could not control for any order eects, because a dierent
protocol would have caused changes in the shelter routine and therefore additional stress for the
shelter dogs and the sta. Nevertheless, as the order was the same for the tested dogs, the results
of a potential order eect should be the same for all groups. e dierent ndings observed in sub
samples suggest that essential oils have dierent eects: this could be a combination of essential
oils’ stimulation and repetition of the test. Further research should investigate the eects of
single essential oils in dierent conditions.e tendencies or signicant decreases that are found
in dierent behaviors in various groups could be caused by an increase in condence the dogs
experienced in the second CB (they were familiar with the CB and might have been eager to
11
participate and enjoy human contact or the enrichment).If a design would be applied in which
50% of the dogs start with essential oils (group 1) and 50% of the dogs without essential oils
(group 2), it is not possible to conduct the control-CB within the same day as the essential oil-
CB for the rst group. It is quite likely that aer 3 hours of exposure to essential oils, an eect of
essential oils would still be present during the control-CB. Conducting the control-CB at another
day would generate a confounding eect of day.
Lastly, we should take into account that Galindo (Galindo et al., 2010) armed that eects
of essential oils can vary considerably depending on the dosage. In our study, we used the same
dosage for each oil. Further studies will need to focus on the eects which obtained by diusing
dierent concentrations of essential oils.
Conclusions
ese preliminary results suggest that olfactory enrichment with essential oils can inuence the
aective states and behaviors of shelter dogs. More research is needed to understand the impact of
each individual essential oil and its eect on dog’s welfare, considering possible factors aecting
their inuence, including individual factors or dierent concentrations of the essential oils.
Acknowledgments
We would rstly like to thank AHVMA for having supported this research. anks to Giorgia
Ascheri and the sta of the shelter Shardana (Sardinia, Italy) for their help during data collection.
We appreciate the contribution of Stijn Schoelynck during the data analysis.
Author Contributions
e idea for the paper was conceived by Haverbeke A. and Uccheddu S. e experimental
protocol was designed by Uccheddu S., Haverbeke A. and Mariti C. e data were statistically
analysed by Arnouts H. and Sannen A. and discussed by all authors. e videos were analysed by
Gutierrez Rufo J. e cortisol concentration in the saliva was analysed by Mariti C. and Gazzano
A.. e paper was written by Uccheddu S. and Haverbeke A. and discussed by all authors.
Conicts of Interest
ere could be a potential conict of interest because Haverbeke A. has selected the composition
of the oils of the blend. e funding sponsors had no role in the design of the study; in the
collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision
to publish the results.
References
Akhlaghi M., Shabanian G., Raeian-Kopaei M., Parvin N., Saadat M., Akhlaghi M. Flor de Citrus au-
rantium e ansiedade pré-operatória. Rev. Bras. Anestesiol. 2011; 61: 707-712. https://doi.org/10.1590/
S0034-70942011000600002.
Atsumi T., Tonosaki K. Smelling lavender and rosemary increases free radical scavenging activity and
decreases cortisol level in saliva. Psychiatry Res. 2007; 150: 89-96. https://doi.org/10.1016/j.psy-
chres.2005.12.012.
Bateson M., Matheson S. Performance on a categorisation task suggests that removal of environmental
enrichment induces “pessimism” in captive European starlings (Sturnus vulgaris). Anim. Welf. 2007;
16: 33-36.
12
Beerda B., Schilder M.B., van Hoo J.A.R.A., de Vries H.W., Mol J.A., Behavioural, saliva cortisol and
heart rate responses to dierent types of stimuli in dogs. Appl. Anim. Behav. Sci. 1998; 58: 365-381.
https://doi.org/10.1016/S0168-1591(97)00145-7.
Binks J., Taylor S., Wills A., Montrose V.T. The behavioural effects of olfactory stimulation on
dogs at a rescue shelter. Appl. Anim. Behav. Sci. 2018; 202: 69-76. https://doi.org/10.1016/J.AP-
PLANIM.2018.01.009.
Boon M. 2003. Goeldi’s monkeys (Callimico goeldii): olfactory enrichment to stimulate natural behaviour
and greater activity. In: Proc. 5th Annu. Symp. Zoo Res. Marwell Zool. Park. Winchester, UK, 7-8th
July 2003; 212-224.
Bouguenoun I., De Pauw-Gillet M.-C., Bensakhri Z., Baiwir D., Mazzucchelli G., De Pauw E., Bendjed-
dou D. Eet inammatoire des extraits de Cupressus sempervirens planté dans l’Est de l’Algérie: du
modèle murin à la stimulation des cellules BEAS-2B par l’allergène majeur Cup s1. Rev. Fr. Allergol.
2016; 56: 10-19. https://doi.org/10.1016/J.REVAL.2015.09.005.
Brydges N.M., Leach M., Nicol K., Wright R., Bateson M. Environmental enrichment induces optimistic
cognitive bias in rats. Anim. Behav. 2011; 81: 169-175. https://doi.org/10.1016/j.anbehav.2010.09.030.
Carvalho-Freitas M.R.R., Costa M. Anxiolytic and sedative eects of extracts and essential oil from Citrus
aurantium L. Biol Pharm Bull. 2002; 25; 1629-1633. https://doi.org/10.1248/bpb.25.
Cobb M.L., Iskandarani K., Chinchilli V.M., Dreschel N.A. A systematic review and meta-analysis of
saliveray cortisol measurement in domestic canines. Domest. Anim. Endocrinol. 2016; 57: 31-42.
https://doi:10.1016/J.DOMANIEND.2016.04.003.
Dawkins M.S., e Science of Animal Suering. Ethology. 2008; 114: 937-945. https://doi.org/10.1111
/j.1439-0310.2008.01557.
de Moraes Pultrini A., Galindo L.A., Costa M. Eects of the essential oils from Citrus aurantium L.in ex-
perimental anxiety models in mice. Life Sciences 2006; 78: 1720-1725.
Douglas C., Bateson M., Walsh C., Bédué A., Edwards S.A., Environmental enrichment induces optimis-
tic cognitive biases in pigs. Appl. Anim. Behav. Sci. 2012; 139: 65-73. https://doi.org/10.1016/J.AP-
PLANIM.2012.02.018.
Doyle R.E., Fisher A.D., Hinch G.N., Boissy A., Lee C. Release from restraint generates a positive
judgement bias in sheep. Appl. Anim. Behav. Sci. 2010; 122: 28-34. https://doi.org/10.1016/j.ap-
planim.2009.11.003.
Fatjó J., Feddersen-Petersen D., Ruiz de la Torre J.L., Amat M., Mets M., Braus B., Manteca X. Ambivalent
signals during agonistic interactions in a captive wolf pack. Appl. Anim. Behav. Sci. 2007; 105: 274-283.
https://doi.org/10.1016/j.applanim.2006.11.009.
Gácsi M., Zsef J., Csányi T.V., Gyori B., Si M., Fia Z., Kubinyi V.E. A Species-Specic Dierences and
Similarities in the Behavior of Hand-Raised Dog and Wolf Pups in Social Situations with Humans.
2005; 47: 111-122 https://doi.org/10.1002/dev.04(20082).
Ferguson C.D., Kleinman H.F., Browining J. Eect of Lavender aromatherapy on acute-stressed horses. J
Equine Vet Sci. 2013; 1:67-69.
Galindo L.A., de Moraes Pultrini A., Costa M. Biological eects of Ocimum gratissimum L. are due to
synergic action among multiple compounds present in essential oil. J. Nat. Med. 2010; 64: 436-441.
https://doi.org/10.1007/s11418-010-0429-2.
Graham L., Wells D.L., Hepper P.G. e inuence of visual stimulation on the behaviour of dogs housed
in a rescue shelter. Anim. Welf. 2005; 14: 143-148. https://doi.org/10.1016/j.applanim.2004.08.024.
Haverbeke A., Diederich C., Depiereux E., Giroy J.M. Cortisol and behavioral responses of working
dogs to environmental challenges. Physiol. Behav. 2008; 93: 59-67. https://doi.org/10.1016/j.phys-
beh.2007.07.014.
Heuberger E., Ilmberger J. e inuence of essential oils on human vigilance. Nat. Prod. Commun. 2010;
5: 1441-1446.
Hongratanaworakit T., Buchbauer C. Evaluation of the harmonizing eect of ylang-ylang oil on humans
aer inhalation. Planta Med. 2004; 70: 632-636. https://doi.org/10.1055/s-2004-827186.
Kaulfuß P., Mills D.M. Neophilia in domestic dogs (Canis familiaris) and its implication for studies of dog
cognition. Animal Cognition 2008; 11: 553-556. https:// doi: 10.1007/s10071-007-0128-x.
Kirschbaum C., Hellhammer D.H. 2000. Salivary cortisol. In: Fink, G. (Ed.), 2000. Encyclopedia of Stress,
Vol. 3. Academic Press, New York, NY, pp. 379-383.
13
Kolevska J., Bruncklik V., Svodoba M. Circadian rhythm of cortisol secretion in dogs of dierent daily ac-
tivities. Acta Vet Brno. 2003; 72: 599-605, https://doi.org/10.2754/avb200372040599.
Kemppainen R.J.. Sartin J.L. Evidence for episodic but not circadian activity in plasma concentrations of
adrenocorticotrophin, cortisol and thyroxin in dogs. J. Endocrinol.1984; 103: 219-226.
Leite M.P., Fassin J., Baziloni E.M.F., Almeida R.N., Mattei R., Leite J.R. Behavioral eects of essential
oil of Citrus aurantium L. inhalation in rats. Brazilian J. Pharmacogn. 2008; 18: 661-666. https://doi.
org/10.1590/S0102-695X2008000500003.
Mason G. J. Stereotypies: a critical review. Anim. Behav. 1991; 41: 1015-1037.
Matheson S.M., Asher L., Bateson M. Larger, enriched cages are associated with “optimistic” response
biases in captive European starlings (Sturnus vulgaris). Appl. Anim. Behav. Sci. 2008; 109: 374-383.
https://doi.org/10.1016/j.applanim.2007.03.007.
Matsubara E., Fukagawa M., Okamoto T., Fukuda A., Hayashi C., Ohnuki K., Shimizu K., Kondo R. Vola-
tiles emitted from the leaves of Laurus nobilis L. improve vigilance performance in visual discrimina-
tion task. Biomed. Res. 2011; 32: 19-28. https://doi.org/10.2220/biomedres.32.19.
Mendl M., Brooks J., Basse C., Burman O., Paul E., Blackwell E., Casey R. Dogs showing separation-
related behaviour exhibit a ‘pessimistic’ cognitive bias. Curr. Biol. 2010; 20: 839-840. https://doi.
org/10.1016/J.CUB.2010.08.030.
Mendl M., Burman O.H.P., Parker R.M.A., Paul E.S. Cognitive bias as an indicator of animal emotion and
welfare: Emerging evidence and underlying mechanisms. Appl. Anim. Behav. Sci. 2009; 118: 161-181.
https://doi.org/10.1016/j.applanim.2009.02.023.
Owczarczak-Garstecka S.C., Burman O.H.P. Can Sleep and Resting Behaviours Be Used as Indica-
tors of Welfare in Shelter Dogs (Canis lupus familiaris)? PLoS One. 2016; 11: e0163620. https://doi.
org/10.1371/journal.pone.0163620.
Paul E.S., Harding E.J., Mendl M. Measuring emotional processes in animals: e utility of a cognitive ap-
proach. Neurosci. Biobehav. Rev. 2005; 29: 469-491. https://doi.org/10.1016/j.neubiorev.2005.01.002.
Rashidi Fakari F., Tabatabaeichehr M., Kamali H., Naseri M. Eect of Inhalation of Aroma of Geranium
Essence on Anxiety and Physiological Parameters during First Stage of Labor in Nulliparous Women: a
Randomized Clinical Trial. J. caring Sci. 2015; 4: 135-41. https://doi.org/10.15171/jcs.2015.014.
Salmeto A.L., Hymel K.A., Carpenter E.C., Brilot B.O., Bateson M., Sufka K.J. Cognitive bias in
the chick anxiety-depression model. Brain Res. 2011; 1373: 124-130. https://doi.org/10.1016/j.
brainres.2010.12.007.
Shiverdecker M.D., Schiml P.A., Hennessy M.B. Human interaction moderates plasma cortisol and
behavioral responses of dogs to shelter housing. Physiol. Behav. 2013; 109: 75-79. https://doi.
org/10.1016/j.physbeh.2012.12.002.
Takahashi T., Ebihara S., Nakamura Y., Takahashi K. A Model of Human Sleep-Related Growth Hor-
mone Secretion in Dogs: Eects of 3, 6, and 12 Hours of Forced Wakefulness on Plasma Growth
Hormone, Cortisol, and Sleep Stages. Endocrinology. 1981; 109: 262-272, https://doi.org/10.1210/
endo-109-1-262.
Titulaer M., Blackwell E.J., Mendl M., Casey R.A. Cross sectional study comparing behavioural, cognitive
and physiological indicators of welfare between short and long term kennelled domestic dogs. Appl.
Anim. Behav. Sci. 2013; 147: 149-158. https://doi.org/10.1016/j.applanim.2013.05.001.
Todd N.B. Inheritance of the catnip response in domestic cats. J. Hered. 2015; 54-56.
Vincent I.C., Michell A.R. Comparison of cortisol concentrations in saliva and plasma of dogs. Res. Vet.
Sci. 1992; 53: 342-345.
Wells D.L., Egli, J.M. e inuence of olfactory enrichment on the behaviour of captive black-foot-
ed cats, Felis nigripes. Appl. Anim. Behav. Sci. 2015; 85: 107-119. https://doi.org/10.1016/J.AP-
PLANIM.2003.08.013.
Wells D. L. A review of environmental enrichment for kennelled dogs, Canis familiaris. Appl. Anim. Be-
hav. Sci. 2004; 85: 307-317, doi:10.1016/J.APPLANIM.2003.11.005.
14
Risposte comportamentali e del cortisolo di cani di canile sottoposti ad un “Cognitive bias test”
dopo arricchimento olfattivo con olii essenziali
Stefania Uccheddu1, Chiara Mariti2, Adinda Sannen3, Hilde Vervaecke3, Heidi Arnout3,4,
Jara Gutierrez Rufo2, Angelo Gazzano2, Anouck Haverbeke1
1 Vet Ethology, Leemveldstraat44, 3090 Overijse, Belgium
2 Department of Veterinary Sciences, University of Pisa, Viale delle Piagge 2, 56124 Pisa, Italy
3 Odisee University College, Agro- & Biotechnology, Ethology & Animal Welfare, Hospitaalstraat 23,
9100 Sint-Niklaas, Belgium
4 University of Antwerp, Departmentof Engineering Management, Prinsstraat 13, 2000 Antwerp, Belgium
Sintesi
L’ambiente di canile comporta per gli animali diverse forme di stress a cui i cani devono adattarsi. Recenti ricerche
hanno dimostrato che l’arricchimento con olii essenziali potrebbe essere in grado di modicare lo stato emozionale di
certe specie animali (cani, gatti, animali di zoo..). In questi studi la valutazione del welfare includeva indicatori siologici,
come ad esempio le concentrazioni di corticosteroidi e/o comportamenti correlati allo stress cronico.
L’eetto olfattorio di 9 olii essenziali (Cananga odorata, Cistus ladaniferus, Citrus aurantium, Cupressus sempervirens,
Juniperus communis var. montana, Lavandula angustifolia, Laurus nobilis, Litsea citrata, Pelargonium graveolens) e di una
miscela di questi olii è stato valutato sui risultati di un “Cognitive bias test”, sui livelli di cortisolo e sul comportamento
di 110 cani di canile (n= 10 cani per ogni gruppo).
L’arricchimento olfattivo con la miscela di olii ha ridotto la latenza della scelta dello stimolo ambiguo, indicando un
pregiudizio ottimistico ed un miglioramento del welfare.
I risultati di questo studio suggeriscono che l’arricchimento olfattivo con olii essenziali può avere un eetto specico
sullo stato emozionale e sul comportamento dei cani di canile e potrebbe perciò essere utile nel management di queste
strutture.
Inoltre, poiché non tutti gli olii testati singolarmente si sono dimostrati ecaci, ulteriori ricerche dovrebbero essere
eettuate per comprendere meglio gli eetti dei singoli olii sul cane.
