Stress, corticosterone responses and avian personalities
- First Online:
- Cite this article as:
- Cockrem, J.F. J Ornithol (2007) 148: 169. doi:10.1007/s10336-007-0175-8
Birds are constantly responding to stimuli from their environment. When these stimuli are perceived as threatening, stress responses are initiated, with activation of the hypothalamo-pituitary-adrenal axis and the release of corticosterone from the adrenal gland. The basic emotion of fear is also experienced during a stress response. Corticosterone stress responses and behavioural responses to stimuli vary markedly between individual birds, raising questions about the significance of these individual differences and about the relationship between corticosterone responses and fearfulness in birds. Although fearfulness can be challenging to measure, data from several species indicate that corticosterone responses and fear behaviour responses are linked in individual birds. Consistent profiles of behavioural responses of birds to a wide range of stimuli can be identified and are called personalities. Personalities vary along a continuum, but are usually classified as proactive or reactive. Individual corticosterone and behaviour responses depend on each bird’s personality. Birds with proactive personalities have relatively active behavioural responses and relatively low corticosterone stress responses, whilst birds with reactive personalities have relatively passive behavioural responses and large corticosterone responses. Relationships between the physiological and behavioural characteristics of avian personalities can now be explored in detail to determine the significance of individual differences in stress responses and personalities in birds.
KeywordsCorticosteroneStressFearPersonalityBehaviourJapanese quailGreat titChickenAdelie penguin
Birds live in complex environments that can change at any time. When changes in their local environment are potentially harmful, then birds, like other vertebrates, adjust by generating stress responses. These are integrated neuroendocrine responses (Ellis et al. 2006) that involve activation of the hypothalamo-pituitary-adrenal (HPA) axis and secretion of cortisol or corticosterone. When an animal perceives a stimulus to be a threat and the HPA axis is activated, the animal also experiences the basic emotion of fear (LeDoux 1996; Labar and LeDoux 2001). Corticosterone, the major glucocorticoid hormone in birds (Cockrem 2007), has metabolic actions such as increasing blood glucose levels, can affect behaviour, and may have other actions that help birds respond to stressors (Sapolsky et al. 2000). Stress responses can differ markedly between individual birds (Littin and Cockrem 2001; Cockrem and Silverin 2002a; Cockrem 2005) and mammals (Guimont and Wynne-Edwards 2006). Animals also differ in their personalities, with personalities defined as consistent behavioural profiles (Dingemanse and Reale 2005; Steimer and Driscoll 2005). It has been suggested that different stress responses in mammals may be associated with different personalities (Korte et al. 2005). The study of personalities is well established in birds (Carere et al. 2005; Groothuis and Carere 2005); stress responses, measured as corticosterone responses to handling, have been studied in many species of birds (Silverin 1998), and fear behaviour has been examined in chickens, Gallusdomesticus, and Japanese quail, Coturnixcoturnix japonica (Jones 1996). Stress responses, fear behaviour and personalities will be considered together in the current review to determine how endocrine stress responses are related to personality in birds.
Stress and corticosterone responses
It has been said that the normal expression of adaptive behaviour depends on some optimal level of stress (Levine 1971), but what is stress? A surprising number of publications that discuss stress in animals do not actually define the term stress. When definitions are given they vary from general (e.g. the nonspecific response of the body to any demand made upon it; Selye 1974) to specific [e.g. a stressor is any stimulus (behavioural, environmental or demographic) that provokes a physiological stress response, as measured by an increase in glucocorticoid secretion; Creel 2001]. The definition that will be used here is that “stress is a state when the hypothalamo-pituitary-adrenal (HPA) axis is activated with increased secretion of glucocorticoids in response to a stressor”.
An increase in plasma corticosterone can be used to indicate when and to what degree a bird is experiencing stress. This contrasts with measurements of increases in sympathetic nervous system activity or observations of changes in behaviour which do not unequivocally indicate that a bird is experiencing stress. An example of the use of corticosterone responses to determine when birds were experiencing stress was a study of corticosterone and behavioural responses of captive great tits, Parusmajor (Cockrem and Silverin 2002b). Great tits in an aviary were exposed for 60 min to a predator (stuffed Tengmalm’s owl, Aegoliusfunereus) or to a stuffed brambling, Fringillamontifringilla, a passerine bird that is not a threat to great tits. Great tits showed similar changes in behaviour when exposed to the owl and the brambling, whereas there was a corticosterone and hence stress response to the sight of the owl but no corticosterone response to the brambling.
Corticosterone secretion from the cortex of the adrenal gland in birds is stimulated by adrenocorticotropic hormone (ACTH) from the pituitary gland, which in turn is stimulated by corticotropin-releasing factor (CRF) and arginine vasotocin (AVT) from the hypothalamus (Carsia and Harvey 2000). The HPA axis is activated in response to stressors, with an increase in plasma corticosterone concentrations detectable one to two minutes after initial exposure to a stressor (Dawson and Howe 1983; Romero and Reed 2005). Corticosterone responses in birds are responses to stimuli from the environment. It is difficult to measure responses of birds to natural stimuli, so most studies of corticosterone responses in birds describe responses to artificial stimuli. Corticosterone responses measured in birds are usually responses of free-living birds or captive birds in aviaries to capture and handling, or responses of captive birds in cages to removal from their cage and handling. It is generally assumed that birds experience capture and handling as a form of predation (Silverin 1998), although corticosterone responses to capture by a predator have not actually been measured for any bird. Free-living birds are caught, an initial blood sample is collected and then the birds are held in bags, with further samples collected for 30–60 min after capture. Corticosterone concentrations in free-living birds usually rise rapidly for 10–15 min after capture then rise more slowly for the remainder of the sampling period (e.g. Wingfield et al. 1992, 1995). Corticosterone responses to handling of free-living birds held in captivity are generally similar to those of birds in their natural habitat [e.g. responses of captive great tits (Cockrem and Silverin 2002a) and free-living willow tits, Parusmontanus (Silverin 1997]. Domesticated birds such as chickens and Japanese quail have corticosterone responses of much lower amplitude than those of free-living birds, e.g. responses to 15 min of continuous handling (chicken; Fraisse and Cockrem 2006) or to 10 min of physical restraint (Japanese quail; Satterlee and Johnson 1988).
Intra- and interspecific variation in corticosterone responses
Corticosterone secretion increases when a bird perceives a stressor (Cockrem and Silverin 2002b), and corticosterone promotes changes in behaviour and metabolism that are thought to help birds to adjust to stressful situation (Cockrem et al. 2004). An acute rise in corticosterone in response to a stressor is generally considered to be beneficial to birds. However, if the corticosterone response to a stressor is larger than normal, if it persists for longer than usual, or if basal corticosterone levels are high then it is widely held that corticosterone may have negative effects. The existence of variation between birds in the size of their corticosterone responses, illustrated in Fig. 2, raises questions about why this variation exists and how it arises. If a corticosterone response helps a bird respond to a stressor then why do some birds have relatively small responses? Conversely, why do some birds have relatively high corticosterone responses if high responses are potentially deleterious? Cockrem (2005) suggested that birds with smaller responses are likely to be more successful in a constant environment and that birds with larger responses may be more successful in changing environments. This implies that there is no optimal corticosterone response for all situations, a concept that is consistent with explanations of variation in stress responses in mammals (Ellis et al. 2006).
Whilst it has been proposed for mammals that variation in stress responses results from genetic variation and adaptive phenotypic plasticity (Ellis et al. 2006), the relative contributions of genetic variation and phenotypic plasticity to variation in corticosterone responses in birds remain unknown. The existence of genetic variation in corticosterone responses in birds is clearly demonstrated by the heritability of the magnitude of corticosterone responses in lines of Japanese quail selected for low and high responses (Satterlee and Johnson 1988; Odeh et al. 2003). It has been suggested that genetic variation in corticosterone and cortisol responses arises from natural selection, which maintains genes for relatively high and low aggression and produces phenotypes called hawks and doves (Korte et al. 2005). The hawk–dove model describes physiological and behavioural characteristics of the two groups of animals and suggests that both groups can be successful but in different environmental conditions. Breeding success, survival and fitness and hence the relative success of birds with low or high corticosterone responses have not yet been studied in free-living birds.
Phenotypic plasticity has been defined as “the morphological and physiological responses of an organism’s phenotype to a change in environmental conditions” (Schlichting 1989), and as “the extent to which an organism can change its physiology, behaviour, morphology and/or development in response to environmental cues” (Dufty et al. 2002). Phenotypic plasticity arises from influences of the environment throughout the life of an animal (Dufty et al. 2002), with maternal influences both before and after birth contributing to differences between adult mammals in their cortisol stress responses (Maccari et al. 2003; Cameron et al. 2005). Exposure of mammals to stressors during development is generally thought to lead to increased reactivity of the HPA axis in adults (Meaney 2001). However, an alternative view is that both relatively low or high levels of exposure to adversity during development lead to increased responsiveness to stressors in adulthood.
The extent to which birds show phenotypic plasticity of corticosterone responses and the origins of such plasticity have not been characterised in birds. Maternal effects on corticosterone responses of offspring could be mediated via maternal corticosterone deposited in eggs (Hayward and Wingfield 2004), and food availability for chicks may affect corticosterone responses of the birds when they are adults (Pravosudov and Kitaysky 2006). There are various approaches to the quantification of phenotypic plasticity, with the most widely used being the slope of a reaction norm (Valladares et al. 2006). A reaction norm is a function that describes for a given genotype the relationship between strength of an environmental variable and phenotype. Reaction norms can be used to indicate the relative importance of genetic contributions and environmental effects in the generation of phenotypes (Van Noordwijk 1989; Postma and Van Noordwijk 2005). Animals have individual reaction norms, so phenotypic variation for a trait can be described for a population from the reaction norms of the individuals within the population. A reaction norm for corticosterone responses in a population of birds could be determined from individual responses to a stressor that was applied at different intensities. This approach could be used to measure reaction norms in species of birds living in different environments to determine whether the slopes of reaction norms, and hence the amount of phenotypic plasticity in the corticosterone responses of birds, differ between species. It is predicted that species living in relatively constant environments will show lower plasticity then species living in more variable environments.
Corticosterone responses have been measured in many species of birds in different situations and at different times of year (Silverin 1998; Romero 2002). Mean responses of groups of birds of the same species can differ in relation to weather conditions at the time of sampling, between seasons, between populations and between years, so it is difficult to interpret the significance of interspecific differences in corticosterone responses. However, differences between species in the relative variation between individual birds may be less than is first apparent. The magnitude of the corticosterone responses shown in Fig. 2 differed between the four species of birds. The maximum mean corticosterone concentrations in the Adelie penguins (at 30 min; Fig. 2c) were 11 times greater than in the chickens (at 15 min; Fig. 2b). However, when the variation between individual birds at these times is expressed as coefficients of variation, the coefficients are the same for the penguins and chickens (51.4 and 50.5%, respectively). It is apparent from this example that the relative amount of variation between individual birds in their corticosterone responses can be similar across species that have quite different absolute corticosterone responses.
Fear and corticosterone responses
Activation of the HPA axis when an animal perceives a stimulus to be a threat is considered to occur simultaneously with the basic emotion of fear (LeDoux 1996; Labar and LeDoux 2001). Fear can be defined as the state or situation in which an animal perceives a stimulus to be a threat and in which the animal is generating a behavioural and/or a physiological response to the stimulus. Fear behaviour helps an animal to avoid or reduce the possible consequence of exposure to danger. The quantification of fear is problematical (Murphy 1978; Miller et al. 2006), but fear is nonetheless widely measured in chickens and quail (Jones 1996). The degree of fear experienced by a bird in a particular situation can potentially be assessed from behavioural observations, such as measuring the distance to which a bird can be approached before it flees (flight initiation distance; Stankowich and Blumstein 2005). However, it is difficult to determine whether observed behaviours that may be called fear behaviours actually reflect an underlying state of fear. Instead, studies of fear often use behavioural tests in controlled situations to measure fearfulness, where underlying fearfulness is “the propensity to be more or less easily frightened” (Jones 1996), or “a temperament trait defining the general susceptibility of an animal to react to a variety of potentially threatening situations” (Boissy 1995).
Fearfulness in birds is most commonly measured in behaviour tests in which birds are placed in new situations (open field tests) or are presented with novel or startling stimuli (novel object or startle tests). Another test that is widely used in chickens and quail is the tonic immobility test. Tonic immobility (TI) is an unlearned response easily induced by brief manual restraint in which a bird remains still and exhibits reduced responsiveness to external stimulation (Jones 1986). TI is thought to be a form of antipredator behaviour (Gallup 1977), and it has been suggested that TI is equivalent to the immobility response within a predation episode (Ratner 1967). Characteristics of TI in chickens include temporary suppression of the righting response, reduced vocalisation, and intermittent eye closure (Gallup 1977; Jones 1986). TI is induced in chickens by placing a bird on its side or back and lightly restraining it, with one hand held over its head and the other hand placed on its sternum for 15 s (Fraisse and Cockrem 2006). The handler then slowly removes their hands, moves back and observes the bird. If the bird rights itself within 15 s, the induction is repeated up to five-times. The time to first head movement and the duration of TI are considered to be positively related to fearfulness, and the numbers of inductions and head movements to be negatively related to fearfulness (Jones 1988).
Some strains of chicken are generally considered to be more flighty and fearful than others. Fraisse and Cockrem (2006) compared corticosterone responses and fearfulness in white Leghorn and brown Hyline laying hens and found that corticosterone responses to handling were greater in white hens. The duration of tonic immobility, latency to first head movement and number of head movements in tonic immobility tests were greater in white than brown hens, whereas the number of inductions was less for tonic immobility tests. The tonic immobility and corticosterone response results were consistent and indicated greater fearfulness and larger corticosterone responses to potentially threatening stimuli in white than brown hens. Data from lines of Japanese quail selected for low and high corticosterone responses to restraint also support an association between corticosterone responses and fearfulness in birds. Quail selected for low corticosterone responses show lower fearfulness than quail selected for high corticosterone responses (Jones et al. 1994a, 1994b).
Personality and corticosterone responses
Animals differ in their responses to other animals and to the environment. For example, some animals are more aggressive and more likely to explore new situations than other animals. Characteristic patterns of behaviour in individual animals represent strategies that the animals use to cope with demands from their environment. These patterns have been called coping strategies or coping styles, where coping styles are defined as coherent sets of behavioral and physiological stress responses to common challenges faced by animals (Koolhaas et al. 1999). Coping styles are consistent over time, are equivalent to behavioural syndromes (Sih et al. 2004) and animal temperaments (Reale et al. 2007), and can also be called personalities (Carere and Eens 2005). Personalities have been defined as suites of correlated behaviours that are expressed across different situations (Carere and Eens 2005). The existence of consistent individual differences in personalities can be explained from a consideration of life-history trade-offs (Wolf et al. 2007). It has been suggested that animals with proactive personalities are likely to be more successful in an environment that remains constant than those with reactive personalities (Cockrem 2005). Conversely, the more cautious style of reactive animals may be more successful in a changing environment and hence there is no optimal personality for all situations.
Personalities in free-living animals have been better studied in birds, especially the great tit in the Netherlands, than in any other animal group (Groothuis and Carere 2005). Although personalities vary along a continuum, individuals are commonly divided into two personality groups on the basis of their behavioural responses to various situations. Great tit personalities are assigned from their performance in novel environment and novel object tests (Groothuis and Carere 2005). Birds are released individually into a novel room or presented with novel objects in their cage. The rates of exploration in the room and approach to the novel object are scored and used to classify birds as fast or slow explorers. These tests are used both in captive birds and in free-living birds that are brought to the laboratory, tested and then released at their capture site. Great tits have been selected for fast or slow exploratory behaviour and the personalities of birds in each selection line examined in detail (Groothuis and Carere 2005). Fast and slow personalities in great tits are equivalent to the proactive (active) and reactive (passive) coping styles or personalities identified in mammals (Koolhaas et al. 1999). For example, behavioural characteristics of lines of mice selected for high and low aggression (short and long attack latencies; SAL and LAL lines; Benus et al. 1991; Veenema et al. 2003) and classified as having proactive and reactive personalities are generally similar to characteristics of fast and slow great tits.
Survival, dispersal and breeding success of free-living great tits whose personality has been determined in laboratory tests have also been described (Dingemanse et al. 2003, 2004; Dingemanse and De Goede 2004). Survival of adults and the number of offspring that survived and contributed to the breeding population (a measure of fitness) were both related to personality, but the relative success of proactive or reactive great tits differed between years and between sexes (Dingemanse et al. 2004). These findings indicated that selection pressures on personality traits can differ between years, leading to the maintenance in avian populations of genetic variation in personalities. It can be suggested that in general birds with proactive personalities are likely to be more successful in constant or predictable conditions and birds with reactive personalities will be more successful in changing or unpredictable conditions. Clearly there is a need for further studies of relationships between survival, fitness and personality in free-living birds.
Birds vary in their personalities and in their corticosterone responses, so are proactive and reactive personalities associated with different corticosterone responses to stressors? This question can be addressed indirectly from a consideration of behaviour in Japanese quail selected for low and high corticosterone responses. A wide range of behaviour tests have been conducted in lines of quail selected on their corticosterone responses to 10 min of restraint (Satterlee and Johnson 1988). LS (low corticosterone stress response) quail are less fearful than HS (high corticosterone stress response) quail in tests of tonic immobility (Satterlee et al. 1993), open field behaviour (Jones et al. 1992) and avoidance of humans (Jones et al. 1994a, 1994b). LS quail show higher sociality than HS quail (Jones et al. 2002), and also have a higher growth rate (Satterlee and Johnson 1985) and reach puberty earlier (Marin et al. 2002; Satterlee et al. 2002). The behaviour data from the LS and HS quail are consistent with the classification of the two selection lines into proactive and reactive personalities respectively, so a proactive personality in these quail is associated with relatively low corticosterone responses and a reactive personality with relatively high corticosterone responses. Lines of quail have also been selected for short (STI) and long (LTI) durations of tonic immobility (Hazard et al. 2005). STI quail are less fearful than LTI quail in open field as well as tonic immobility tests, but differences between the lines in corticosterone responses have not been shown (Jones et al. 1991, 1994a, 1994b), and there were no significant correlations between corticosterone and behaviour variables in these quail (Mignon-Grasteau et al. 2003). Individual differences in corticosterone responses are more apparent when birds are challenged with mild rather than strong stressors (Beuving and Vonder 1986), and line differences between the STI and LTI quail may become apparent if corticosterone responses to a milder stressor are measured in future.
Relationships between corticosterone responses and behaviour can also been examined in data from great tits and chickens. It has been suggested that corticosterone responses may be smaller in fast (proactive) compared with slow (reactive) great tits (Carere et al. 2003; Groothuis and Carere 2005). White Leghorn hens selected for low (LFP) compared with high (HFP) levels of feather pecking showed less fear in an open field but not in a tonic immobility test (Jones et al. 1995). LFP hens had higher corticosterone responses and lower noradrenaline sympathetic nervous system responses to restraint than HFP hens (Korte et al. 1997). Selection for feather pecking in these hens has not led to lines of hens with different personalities. In another study, white Leghorn hens that showed less fear on the basis of tonic immobility responses also had smaller corticosterone responses, although not significantly so, to a stressor than hens showing more fear (Beuving et al. 1989). Personality differences can also be identified between strains of laying hens. Brown Hyline hens with relatively smaller corticosterone responses to handling and lower fearfulness than white Leghorn hens (Fraisse and Cockrem 2006) have proactive personalities, and white Leghorn hens have reactive personalities.
Proposed characteristics of birds with proactive and reactive personalities
Fast great titsa
Slow great tits
Bold and aggressive
Shy or cautious and not aggressive
Fast and superficial explorers
Slow and thorough explorers
Less flexible and less sensitive to environmental stimuli
More flexible and more sensitive to environmental stimuli
Lower corticosterone responses to stressors
Higher corticosterone responses to stressors
Relatively high sympathetic nervous system reactivity
Relatively low sympathetic nervous system reactivity
Relatively low parasympathetic nervous system reactivity
Relatively high parasympathetic nervous system reactivity
More successful in constant or predictable conditions
More successful in changing or unpredictable conditions
Research in stress endocrinology described here has been supported by the Institute of Veterinary, Animal and Biomedical Sciences, Massey University, and by Antarctica New Zealand.