Validation of modified open field behaviour as a measure of trait anxiety in the dog
Introduction
Anxiety is a core emotion which has defined neurophysiological processes underpinning its existence in a range of mammalian species (Millan, 2003). Many behavioural disorders in the dog are believed to be caused by or directly related to the presence of excessive anxiety. These disorders include aggression, separation anxiety, canine compulsive disorder and phobias (Overall, 2013). Evidence for this includes neurophysiological research into affected dogs and clinical improvement of cases with the use of anxiolytic pharmacotherapy (Simpson et al., 2007, Vermeire et al., 2011). Anxiety in the dog is believed to cause sympathetic signs such as panting and elevated heart rate, along with displacement behaviours such as lip-licking and yawning (Overall, 1997). These behaviours can be used to help identify a dog that is in an anxious state; however this is a qualitative assessment due to the current lack of studies that use quantitative measurements on a relevant scale. As such, no study to date has demonstrated that highly anxious dogs are more likely to be affected by the serious behaviour problems mentioned above. Anxiety can be measured either as a state or a trait. State anxiety is defined as the level of anxiety being experienced at one particular time, and is contextual and able to be conditioned. This differs from trait anxiety, which is the individual tendency for an animal to experience anxiety across time and context (Goes et al., 2009). For trait anxiety to be measured in a behavioural paradigm, the measurement of anxiety must occur in an unconditioned environment and have test-retest repeatability within an animal on novel and subsequent exposures (Teixeira-Silva et al., 2009). Many previous studies measuring the fear or anxiety of a dog occur largely in the presence of a potentially conditioned stimulus such as an unfamiliar dog or a thunderstorm or a human (Svartberg, 2005, Planta and De Meester, 2007, Araujo et al., 2013). For this reason the behaviour may be strongly affected by the individual past experiences (both positive and negative) and socialisation of the individual dog. Therefore, having the ability to measure unconditioned trait anxiety in dogs would allow future testing of the hypothesis that trait anxiety predisposes to behaviour problems in dogs. Breeding animals based on their anxiety has been demonstrated extensively in the laboratory (Murphree et al., 1969, Stead et al., 2006, Bickell et al., 2009). If dogs with high levels of trait anxiety could be excluded from domestic dog breeding programs, then this could potentially reduce the incidence of behaviour problems.
Behavioural paradigms designed to measure both trait and state anxiety have been developed in rodents (Ohl, 2003). These paradigms include the open field test, elevated plus maze and the free exploratory paradigm (Teixeira-Silva et al., 2009). In humans, trait anxiety (measured using a questionnaire) is elevated in people with anxiety disorders (Van Dam et al., 2013). Some studies have measured aspects of canine anxiety using behavioural paradigms. King and colleagues tested greyhounds and beagles with behavioural paradigms originally designed to test anxiety in rodents (King et al., 2003). More recently, canine-specific tests have been developed to measure state anxiety in the dog (Åkerberg et al., 2012, Araujo et al., 2013). These studies focus on state anxiety in the presence of potentially preconditioned stimuli (such as the presence of a human or the sound of a thunderstorm).
Careful design of behavioural paradigms with consideration to the ethology of the species can give face validity to the experiment. However, construct validity requires some degree of pharmacological or neurophysiological validation (Rodgers et al., 1997). Diazepam is an anxiolytic drug commonly used for pharmacological validation of paradigms used to measure anxiety in a range of mammalian species (Ohl, 2003)In previous studies, diazepam has reduced some of the behavioural signs of anxiety-related behaviour problems (such as decreased activity in response to a simulated thunderstorm). (Herron et al., 2008, Araujo et al., 2013). The effect of diazepam on activity has been reported previously in dogs and other species depending on the dosage used (Crawley, 1985, Herron et al., 2008). Other studies have assessed the repeatability of behavioural tests in dogs, finding high test-retest correlations within dog for many of the behaviours measured (King et al., 2003, Svartberg et al., 2005). Typically, behavioural assessments involve the test for some form of fear or aggression in response to commonly encountered contexts which may have been preconditioned. For example, some dogs within a population may have already developed fear of unfamiliar humans or thunderstorm phobias, while others may have habituated to these stimuli. In this case the individual reaction will be influenced by the past experience of those dogs, such as the amount of human socialisation and previous exposure to thunderstorms. This study aimed to measure anxiety-related behaviours with high test-retest repeatability that are relatively unconditioned. This was attempted by using exposure to novel stimuli and a detailed examination of canine open field behaviour in the presence or absence of diazepam. In addition, a novel substrate preference test was included to further investigate the interaction between exploratory drive and neophobia. As canine exploratory behaviour in a novel environment is believed to be related to an underlying behavioural trait (or traits), it is expected to have significant test-retest repeatability. Furthermore, anxiety is believed to cause an inhibition of exploratory behaviour in a novel environment, and therefore an anxiolytic drug (such as diazepam) is expected to remove this inhibition.
Section snippets
Subjects
The subjects used in this study were healthy desexed male and female Greyhound dogs (n = 63) aged between 2 and 9 years and weighing between 26 and 43 kg. The dogs were originally sourced from the greyhound racing sector in the state of Victoria (Australia) and were housed at the Melbourne University Canine Blood Bank, where they were bled regularly but no more than once per month. The bleeding procedure was performed by a veterinarian who removed approximately 450 ml of blood by jugular
Repeated exposure group, trial 1 and 2
The average age of this group of dogs was 6.64 years (SE = 0.26). Test-retest repeatability for a range of behavioural measures during the first two trials was measured using Spearman’s rank correlation coefficient (displayed in Table 1). Large positive correlations for the factors of total distance travelled (ρ = 0.813, 95% CI: 0.664–0.900, P < 0.001), displacement from the speaker during the acoustic stimulus (ρ = 0.699, 95% CI: 0.484–0.834, P < 0.001) and distance midline during the acoustic stimulus
Discussion
This study demonstrated the repeatability of a range of behavioural measures within dog across time, and the effect of diazepam on those behaviours in both test-naïve and test-experienced dogs. The behavioural measures with the highest repeatability in the REx group (distance travelled and distance midline during acoustic stimulus) were also the measures that were significantly affected by diazepam treatment in the AEx group.
In rodent models of anxiety, test-retest repeatability of behavioural
Conclusion
This study demonstrated that in an open field there are a range of repeatable behavioural measures including locomotor activity and aversion to a noise stimulus. The use of diazepam before behavioural testing has the effect of increasing locomotor activity and reducing noise stimulus aversion in this paradigm. This effect of diazepam on these test-retest repeatable behavioural measures supports the hypothesis that they are indices of trait anxiety in the dog. Further validation of this test in
Acknowledgment
The technical support of Melanie Conley, and the use of dogs from the Melbourne University Canine Blood Bank are gratefully acknowledged. AJL is a NHMRC Principal Research Fellow (1020737). We also acknowledge the Victorian Government’s Operational Infrastructure Support Program.
References (30)
- et al.
Test for personality characteristics in dogs used in research
J. Vet. Behav. Clin. Appl. Res.
(2012) - et al.
Development and validation of a novel laboratory model of sound-induced fear and anxiety in Beagle dogs
J. Vet. Behav. Clin. Appl. Res.
(2013) - et al.
Fear responses to noises in domestic dogs: prevalence, risk factors and co-occurrence with other fear related behaviour
Appl. Anim. Behav. Sci.
(2013) Exploratory behavior models of anxiety in mice
Neurosci. Biobehav. Rev.
(1985)- et al.
Trait and state anxiety in animal models: is there correlation?
Neurosci. Lett.
(2009) - et al.
Open field activity and human interaction as a function of age and breed in dogs
Physiol. Behav.
(1997) - et al.
Fear of novel and startling stimuli in domestic dogs
Appl. Anim. Behav. Sci.
(2003) The neurobiology and control of anxious states
Prog. Neurobiol.
(2003)Testing for anxiety
Clin. Neurosci. Res.
(2003)- et al.
The open field as a paradigm to measure the effects of drugs on anxiety-like behaviors: a review
Eur. J. Pharmacol.
(2003)