Heart rate as a measure of emotional arousal in evolutionary biology

1. General introduction

Since the mid 1970s, starting with a seminal book by Hans Seyle [1], investigating individual responses to challenges in the environment, i.e. stressors, has been an important area of research in evolutionary biology. Physiological activation helps organisms to adapt to environmental changes and maintain homeostasis [2]. Activation of the sympathetic branch of the automatic nervous system can be captured by an increase in heart rate, defined as the number of contractions of the heart per minute (beats per minute, bpm) [3], which indicates the initiation of a physiological stress response, defined as a coping response to external and internal stimuli [1] and emotional arousal, a state of heightened psychological activity [4]. These concepts are closely related, where the physiological stress response is often seen as a general response to all stimuli, whereas emotional arousal takes the animal's perception of the situation into consideration [5]. Measures of heart rate variability, i.e. variation in the time interval between heartbeats, have been suggested to reflect the balance between the regulation of the sympathetic and parasympathetic branches of the automatic nervous system [6,7]. As these responses are plastic and context-dependent [8], heart rate measures can provide important insight into how animals perceive and respond to their environment [9]. Further, heart rate is positively correlated to metabolic rate and thus can be used as a reliable proxy of energy expenditure [10]. Ultimately, stress-related increases in heart rate can increase the overall energy expenditure of individuals [11], which is constrained by physical and/or physiological conditions [10]. Heart rate response to stress can be considerable: for example, in yellow-eyed penguins, Megadyptes antipodes, heart rate was more than twice as high during partner reunion with mutual calling (174 bpm) compared to heart rate during resting (77 bpm) and heart rate returned to baseline levels within seconds after the stimulus ended [12]. In greater white-fronted geese, Anser albifrons, maximum heart rate during agonistic encounters increased four times in comparison to mean heart rate during resting (110 bpm), whereas mean heart rates during walking, feeding and preening was similar to resting heart rate (98–118 bpm). Mean heart rate during flight was similar to maximum heart rates during agonistic encounters (413 bpm) and heart rate returned to baseline values in less than 10 s [13]. In mammals, heart rate during exercise (running) increased linearly with running speed and depended on body size (resting heart rates: 63–581 bpm; maximum heart rate during running: 183–679 bpm) [14]. In domestic pigs, Sus scrofa, mean heart rate increased from 52 bpm during resting to 79 bpm during walking and a maximum heart rate of 157 bpm when losing a physical encounter [15].

Here, I review studies that used heart rate as a measure of emotional arousal in the fields of social behaviour, animal cognition, animal welfare and personality research (table 1). A first study, published in the late 1970s, used heart rate as a measure of anthropogenic disturbance, measured by custom-designed wearable transmitter harnesses in bighorn sheep, Ovis canadensis [30]. After this early study in free-ranging animals, only 10 years later, the first laboratory studies in primates measured heart rate responses to social interactions and different social stimuli via electrodes attached to the skin surface or implanted transmitters [23,36,37]. It took another 10 years before heart rate became a more broadly used measure of emotional arousal in a different context in the late 1990s, with researchers also starting to use different technologies (implanted radio-transmitter, wearable heart rate belts, artificial eggs; table 1). I further provide an outlook into future research avenues that could potentially be explored with new technologies allowing for heart rate measurements to be more broadly applied in different contexts, freely moving animals and a wider variety of species.

2. Social modulation of heart rate

Heart rate has been shown to be strongly affected by social interactions in a wide range of species. Individuals' responses during social interactions can be quantified using mean or maximum heart rate during the encounter [61], heart rate increase (calculated as the difference between the momentary baseline heart rate value a couple of seconds (2–3) before the interaction and either the mean heart rate value during the interaction or the instantaneous maximum heart rate [21,61]), recovery period until the heart rate returns to baseline [62], or excess heart rate (calculated as the mean heart rate increase over baseline heart rate times the duration of that increase [21]). Heart rate during and after agonistic encounters usually increases (blind cave crayfish, Orconectes australis packardi: [16]; bighorn sheep: [30]; domestic pigs: [15]; greylag geese, Anser anser: [19] (see figure 1a in comparison to heart rate during take-off figure 1b); greater white-fronted geese: [13]; king penguins, Aptenodytes patagonicus: [21]; pigtail macaque, Macaca nemestrina: [23]; table 2). In greylag geese, the mean heart rate during aggressive encounters (157 bpm) is significantly higher compared to mean heart rate during other behaviours, for example resting (84 bpm), head up (131 bpm), preening (117 bpm) or feeding (118 bpm). Heart rate in greylag geese increased from 134 bpm during standing and walking, to 159 bpm during swimming and 368 bpm when flying [63]. Heart rate increases after agonistic encounters in greylag geese on average lasted for 8 s (range: 1–261 s; C.A.F. Wascher 2006, unpublished data). Affiliative behaviours generally reduce heart rate (cattle, Bos taurus: [17]; pigtail macaque [23]; rhesus macaques, Macaca mulatta: [26]) and being paired with a long-term social partner was shown to buffer an increase in heart rate during aggressive encounters in male greylag geese [64]. Responses often differ depending on an individual's role as initiator or receiver in the interaction, or whether the individual is dominant or subordinate to its interaction partner. For example, in rats, Rattus norvegicus, and greylag geese, individuals winning agonistic encounters had significantly higher levels of emotional arousal compared to the defeated individuals [24,61]. Already the approach by a dominant group member and the associated risk of aggression were found to significantly increase heart rate in rhesus macaques [26], and even indirect social interactions can affect heart rate, for example observing aggressive interactions as a bystander [20,21], or visual and olfactory presentation of dominant or aggressive individuals [36,46]. The relationship between social status and heart rate can manifest itself in rank-dependent differences in resting heart rates, as described in red deer, Cervus elaphus, where subordinate individuals have lower resting heart rates compared to dominant individuals, which might result in a better capacity to minimize energy requirements [25].

• Download figure
• Open in new tab
• Download PowerPoint

Current research shows that heart rate provides a suitable proxy for estimating individual investment into the social domain, and that measuring heart rate in a variety of social contexts facilitates investigating the adaptive benefits of social relationships across different time frames. However, presently, only a few studies report detailed individual physiological responses to social interactions and often these studies describe short-term effects only instead of investigating any long-term impacts of emotional arousal on energy expenditure, health, reproductive output or longevity.

3. Measuring heart rate in the context of animal cognition research

Generally, the study of animal cognition investigates mechanisms by which animals acquire, process, store and act upon information in the environment [65]. A major constraint in understanding these processes is that cognition can only be indirectly assessed, for example by observing behaviours in response to experimental manipulation, and that behavioural responses to stimuli can be difficult to interpret if individuals differ in the expression of emotional arousal—or do not show any notable behavioural change in response to stimuli [20]. However, given that physiological changes to an external stimulus strongly depend on an individual's perception and evaluation of the event [5], measuring physiological parameters such as heart rate can be a promising quantitative tool to assess cognitive processes and arousal in response to stimuli in the environment. Lambs, Ovis aries, and rats responded with a heart rate increase to novel, sudden and unpredictable stimuli [47,48]. When engaging in learning tasks, heart rate increased in goats, Capra hircus [45], horses, Equus caballus [42] and cattle [35] but decreased in European starlings, Sturnus vulgaris [41]. In the future, improved biologging technology to measure heart rate in different contexts could allow more objective quantification of cognitive function in a wide range of species, as already shown by the use of neurologgers to measure electroencephalograms in order to quantify sleep patterns in free-living animals [66–69].

Heart rate monitoring allows us to gain insights into the perceived level of difficulty of a cognitive challenge or whether individuals can differentiate between specific stimuli. Studies measuring heart rate during the presentation of conspecific and heterospecific photographs showed that chimpanzees, Pan troglodytes, respond differently to familiar humans and strangers [37], and their different heart rate response to aggressive, friendly or unfamiliar conspecifics suggests that chimpanzees can also discriminate between those categories [36]. Similarly, heart rate measured in response to auditory stimuli indicated that dogs, Canis familiaris, can differentiate between conspecific barks given in different contexts [39], and blue penguin chicks, Eudyptula minor, had a higher heart rate during playbacks of calls from familiar individuals (siblings and neighbours) compared to calls of unfamiliar individuals [44]. Interestingly, penguin chicks did not express observable behavioural differences between the treatments, highlighting again that heart rate can reveal individuals' ability to discriminate between stimuli where it could not have been inferred from behavioural observations.

Heart rate can also reveal whether and to what degree animals respond to emotions of other individuals, i.e. emotional contagion. When watching conspecifics experiencing stimuli of negative emotional valence, dogs and horses responded with a higher heart rate compared to seeing others experiencing a positive treatment [43,70], and hens, Gallus gallus domesticus, significantly increased heart rate when their young were exposed to a mild stressor [38]. Similarly, greylag geese responded with a higher increase in heart rate when observing interactions involving closely affiliated individuals (pair-partner and family members) instead of non-affiliated individuals, even in the apparent absence of a behavioural response to these interactions (figure 1c)) [20].

Although heart rate presents a good means for quantifying the level of emotional arousal and does so more accurately compared to behavioural measures, it does not provide information about the valence of emotional arousal, because heart rate increases in response to both positive and negative experiences [18,71]. Potentially, different parameters of heart rate variability may allow us to better disentangle effects of positive versus negative valence [72,73]. Future research investigating heart rate during cognitive experiments could provide not only insights into the cognitive abilities of individuals and advance the field of animal cognition, but also improve our understanding of heart rate modulation in response to a variety of carefully chosen stimuli.

4. Individual differences in heart rate

Consistent inter-individual variation in phenotypic traits, including behavioural and physiological traits, has been described in a wide variety of species. With regard to physiology, individuals differ in both reactivity to challenges in the environment as well as in their baseline levels [74]. In some species, individuals with more aggressive behavioural traits have been shown to increase their heart rate more during aggressive interactions compared to less aggressive individuals [50,75]. The heart rate of chickens from different pedigrees (low and high feather pecking lines) did not significantly differ during baseline conditions or during restraint stress, but heart rates in high feather pecking lines tended to be elevated for longer [49]. In yellow-eyed penguins, individuals differed in their responses and habituation to anthropogenic disturbances [76]. Domestication is also assumed to affect physiological traits. Wild animals had higher average heart rates compared to farmed individuals in Atlantic salmon, Salmo salar [77], and dogs have higher resting heart rate compared to wolves, Canis lupus: [78].

While most studies reported in this review focus on heart rate or heart rate variability, research on individual responses to challenges could also include more detailed investigations into individual recovery periods [22]. Also, repeatability of physiological responses over time and across contexts has hardly been investigated but would provide insights into the development and maintenance of consistent phenotypes, i.e. ‘animal personality’. In summary, a better understanding of individual differences in physiological responses across different contexts could reveal links between behaviour and physiology that affect behavioural decision making and the evolution of different life-history strategies [74,79]. For example, individual stress responses can be linked to the energy demands of various reproductive stages [80,81], season [11,82,83], moulting [84,85], length of light–dark cycle [86], predation risk [87] and parasite load [88].

5. Implications for animal welfare

Good animal welfare is generally assumed when animals are not experiencing unnecessary stress. Since heart rate can be used to quantify the activation of the physiological stress response, monitoring heart rate can ultimately be considered a useful tool when attempting to improve animal welfare in a variety of contexts. The benefit of using heart rate as a measure of the physiological stress response, in contrast to, for example, glucocorticoid levels, is that heart rate can reflect individual levels of arousal accurately and at a high temporal resolution, and even small physiological responses to specific stimuli, in the absence of visible behavioural reactions, can be isolated [89].

Heart rate can be used to monitor arousal levels of individuals when evaluating the effectiveness of measures intended to reduce stress, for example, when enrichment or other changes are introduced in animal shelters or laboratories. In kennelled dogs, the presentation of auditory and olfactory enrichment decreased heart rate and changed heart rate variability, indicating stress reduction [52,53], and mice, Mus musculus, that were housed in enriched conditions had a significantly lower heart rate compared to individuals housed under minimal husbandry conditions without enrichment and periods of individual housing [55]. Generally, social isolation can cause significant increases in heart rate (alpaca, Vicugna pacos: [90]; dogs: [78]; cattle: [54]; cynomolgus monkeys, Macaca fascicularis: [91]; goats: [18]; wolves [78]). However, affiliative behaviours, even if performed by heterospecifics, can have a stress-reducing effect; grooming carried out by a familiar human decreased heart rate in laboratory-kept rhesus monkeys [92], dairy cows [93] and lambs [94]. Measuring heart rate can also reveal ineffective enrichment designs or unsuccessful attempts at stress reduction in animal housing. For example, individually housed rhesus monkeys in a laboratory environment did not show any changes in heart rate in response to increasing cage sizes [56] and auditory enrichment did not result in the expected reduction in heart rate in African green monkeys, Chlorocebus aethiops [51].

Biologging technologies to measure heart rate could be widely used to assess the effects of anthropogenic disturbance on wildlife. This is especially relevant as previous studies have shown that individuals can show pronounced emotional arousal, as measured by heart rate, in the absence of obvious behavioural changes. For example, American black bears, Ursus americanus, and bighorn sheep showed no behavioural responses but significantly increased heart rate in response to anthropogenic disturbances, such as drone overflights and vehicle traffic [28,30]. Anthropogenic sounds alone were identified as a stressor that led to elevated heart rate in farm animals [95] and fish [33]. Also, direct human contact can increase heart rate in wild animals; for example, brown bears, Ursus arctos, significantly increased heart rate in response to dog hunts and human encounters [31] and in captive koalas, Phascolarctos cinereus, heart rate increased in the presence of tourists compared to situations without humans [32]. Human approach towards nesting birds may not always lead to females leaving the nest, but elevated heart rate for extended periods can increase energy expenditure (Adélie penguins, Pygoscelis adeliae: [27]; Humboldt penguin, Spheniscus humboldti: [12]; yellow-eyed penguin [12]; wandering albatrosses, Diomedea exulans: [34]). By contrast, other populations might be found to be more resilient; for example, nesting American oystercatchers, Haematopus palliates, did not significantly increase heart rate in response to a variety of human disturbances, including human approach, off-road vehicles and aircraft overflights [29], and greylag geese did not significantly increase heart rate in response to familiar humans approaching [96] (figure 1d). Responses to human approach can also be affected by previous experience with humans and wild animals can habituate to the presence of humans [22,97]. Knowledge about physiological responses of wild animals to human disturbances can be used to develop guidelines and recommendations to minimize the impact of human activities [98].

A core application for using heart rate measurements to assess animal welfare is in the context of livestock management, given the large numbers of animals kept for food production and increasing demands for improved ethical production of animal products. Decisions on transport and husbandry routines should aim to minimize arousal levels, and heart rate monitoring can provide important evidence for approaches that aim to improve animal welfare. For example, when transporting sheep, heart rate revealed differences in the perceived stress depending on the methods of driving and loading sheep [99,100], and cognitively enriched pigs have a lower heart rate during feeding announcement compared to a control group [101]. Recently, heart rate measures have also been implemented in precision livestock farming, defined as the use of advanced technologies for continuous automated real-time animal monitoring to optimize the contribution of each animal [58–60,102,103]. In future, heart rate could be used on a large scale to identify early signs of decreased welfare or poor health in groups of farmed animals, pets or critically endangered animals in a conservation context (table 1).

6. Summary and future research directions

Heart rate provides an accurate and comprehensive measure of emotional arousal in non-human animals [9]. Here, I propose that research on heart rate modulation in a wide variety of species and different contexts could be of interdisciplinary interest and would provide important insights for basic research, animal welfare and conservation.

One of the biggest advantages of heart rate measures is that it allows relatively accessible measuring of short-term levels of emotional arousal in response to single events (e.g. vocalizations, social interactions and human disturbance), but heart rate can also assess the impact of long-term stressors on individuals' physiology (e.g. seasonal effects, noise pollution) and reveal chronic stress. Assessment of heart rate can be cheaper compared to other measures of activation of the physiological stress response, for example, non-invasive measures of glucocorticoid metabolites in faeces [104].

A major methodological challenge in evolutionary biology research is that, when quantifying activation of the physiological stress response, studies often compare absolute levels of arousal or stress, with little consideration about the adaptive benefit or cost associated with the response [105]. Activation of the physiological stress response has been proposed to have both adaptive and maladaptive aspects [105]. Short-term increases in heart rate are considered adaptative to help individuals to cope with a given stimulus in the environment. However, increases in heart rate result in an increase in energy expenditure and long-term maintenance of an elevated physiological stress response might be costly and have negative effects on an individual's health and fitness (e.g. [53]). For example, stress through group instability caused increased coronary artery atherosclerosis in dominant male cynomolgus monkeys compared to subordinate individuals [106]. A better understanding of potential fitness consequences of activation of the physiological stress response in behavioural biology is important for interpreting findings about heart rate responses in different contexts. Recording multiple physiological parameters simultaneously, e.g. body temperature, heat flux, blood and tissue oxygenation, blood metabolites and gases, gastric pH and motility, or diving air volume and respiratory rate, could help to achieve this by providing a more comprehensive representation of an individual's physiological status [57,107,108].

Heart rate measurements allow us to validate behavioural indicators of emotional arousal and activation of the physiological stress response. Whereas commonly used physiological measures that require laboratory analysis of hormonal changes are expensive, time consuming or practically not feasible, behavioural observations and analyses of certain indicative behaviours, such as vigilance, can be used to analyse responses to different stimuli. Correlating behavioural observations with simultaneous measures of heart rate provides an important validation for using behavioural indicators of individual arousal and stress. In horses, heart rate and heart rate variability are correlated with snorting in stressful conditions [109], ear posture is indicative of emotional arousal in dairy cows [110], and in greylag geese, the maximum heart rate during aggressive encounters is positively related to the frequency of shaking and duration of vigilance after the conflict [62]. However, other studies found no relationship between considered behavioural measures and heart rate in some contexts [111,112]. If a robust relationship between frequency or intensity of behavioural measures and emotional arousal is validated, behavioural observations can be used to assess levels of arousal at low material cost. However, more studies reporting heart rate are needed to carefully validate behavioural indicators of stress under consideration of factors like species, context and individual variation.

More research is needed to investigate in which contexts emotional arousal and activation of the physiological stress response are adaptive and help an animal to cope with stressors, and in which contexts the activation of the physiological stress response is maladaptive and has negative impacts on an individual's health [113]. Here, long-term monitoring over months or years is technologically feasible and could, especially in managed animal species (e.g. farm animals, animals in shelters), be linked to outcomes of daily routines or veterinary inspections. Long-term monitoring of heart rate would also allow us to disentangle short-term reactivity to environmental stimuli from long-term activation of the physiological stress response, which is not well understood yet [11,103]. Heart rate recordings could be incorporated in automated health surveillance processes in a wide variety of different settings [114,115].

A major constraint to assessing heart rate is the availability of suitable devices. In this review, I have presented research applying different biologging technology, ranging from fully implantable transmitters to wearable heart rate belts that can be applied non-invasively (table 1). Different technology makes measuring heart rate possible in a wide range of species [107,116], including fish [51], marine mammals and turtles [28,95], domestic pets and farmed animals [33], but different approaches have specific social, ethical and regulatory challenges [117]. For example, implanting transmitters is an invasive process and might not be ethically acceptable in all contexts and species. Commercially available wearable technology, such as heart rate belts, are not easily applicable to wild animals because they often need readjustment and manual removal. Especially for small species, e.g. birds, there are additional ethical considerations to be made regarding to the weight of the device. To minimize ethical risks of prolonged invasive measuring techniques, robust statistical methods to analyse the collected data are needed, which can help to refine the study design in accordance with the 3Rs (reduce, refine and replace) [118,119]. Next to ethical considerations, different devices also differ in their price range and method of actual heart rate recording. Most heart rate equipment records an electrocardiogram [120] and some require microphones to acoustically record the frequency of heartbeats [121]. All technical solutions measuring heart rate in non-human animals need to be carefully corrected for outlier values, which can arise when the algorithm used to measure the electrocardiogram is not carefully validated for the species in question or when the signal is temporarily lost in un-restrained animals [122]. Therefore, similar to other physiological parameters, heart rate measurements require careful validation and removal of outliers that are not biologically meaningful [120,122,123].

To summarize, using heart rate to gain an improved understanding of how individuals cope with challenges in their environment is of interest for several areas of evolutionary biology, including research on social behaviour, animal cognition and individual differences. For example, understanding the adaptive costs and benefits of social behaviour is a central topic in evolutionary ecology, or quantifying individual levels of stress in response to anthropogenic disturbances or environmental challenges and evaluating the effectiveness of measures intended to reduce stress is of particular importance in applied animal welfare research and conservation. Technological advances in biologging can help to further advance the application of heart rate monitors in diverse areas of animal behaviour research and contribute to studying emotional arousal in different contexts.

Data accessibility

This article has no additional data.

Competing interests

I declare we have no competing interests.

Funding

This work has received no funding.