My Reflections on Understanding Animal Emotions for Improving the Life of Animals in Zoos

ABSTRACT

Scientists are often reluctant to attribute emotions to nonhuman animals that are similar to human emotions. When the author published her early studies, reviewers prohibited the word fear. Fearful behavior had to be described as agitated. The core emotional systems described by Panksepp may provide a useful framework for people who work hands-on with animals. The core systems are fear, rage, panic (separation distress), seek, lust, nurture, and play. Some scientists who deny that animals have real emotions often fail to review important areas of the literature. The areas that are sometimes left out are the effects of psychiatric medications on animals and genetic influences on differences in animal behavior. In both people and animals, genetics has an influence on both fearfulness and novelty seeking. Visualizing the seven core emotional systems as separate volume controls on a music mixing board may help zoo professionals determine the motivation of both normal and abnormal behavior. It may also help them to design more effective environmental enrichments.

KEYWORDS:

• Animal welfare
• zoos
• emotions
• fear
• exploratory behavior
• separation distress

During a long career working with nonhuman animals, my views and understanding of animal behavior have evolved and changed. When I started working with beef cattle in the 1970s, I thought I could fix all problems with handling animals by designing better handling facilities. Later, I learned that management, stockmanship, and how people interact with animals were equally important. Many studies have clearly shown that when animals on farms fear people, they are less productive (Hemsworth & Coleman, 2010; Herman & Panksepp, 1981; Rushen & dePassille, 2015). During restraint for vaccinations and other procedures, animals often become highly behaviorally agitated. My early research showed individual differences in how an animal reacted to being held in a squeeze chute (Grandin, 1993). When I wrote that article, I was informed that the word fear should not be used because it was not scientific. I originally chose the word fear because I had seen it in Hemsworth’s (1989) study. In the 1980s when I was in graduate school, Hemsworth visited the University of Illinois. We had long discussions about animals being able to feel fear. At this point, Hemsworth and I were ahead of our time. Researchers in the livestock and veterinary field were still reluctant to use the word “fear.”

Successfully training the untrainable animal

Nancy Irlbeck, the nutritionist at the Denver Zoo, approached me about how to obtain low-stress vitamin E samples from nyala antelopes. To obtain these samples, we successfully trained unsedated animals to enter a box and stand still for blood sampling from the rear leg (Grandin, Rooney, Phillips, & Cambre, 1995; Phillips, Grandin, Graffam, Irlbeck, & Cambare, 1998). Many people thought that training the antelopes would be impossible because the animals were so flighty. We learned that to prevent the animals from becoming frightened, we had to slowly and carefully habituate them to the equipment. It took 10 days to train them to tolerate a sliding door suddenly opening (Grandin et al., 1995). Habituation had to be conducted before we could use operant conditioning to get them to stand still for blood sampling. We were still not allowed to use the “fear” word. Instead, the words “habituation” and “desensitization” were used (Grandin et al., 1995). In the next study, bongo antelopes were conditioned to go in the box. After 20 minutes in the box, the mean cortisol levels were only 6.4 ± 3.8 ng/mL (Phillips et al., 1998). The glucose levels in our conditioned animals were only 61.25 ± 19.45 mg/dL, and when they had been previously darted or pole-injected, glucose levels had risen to 166.5 ± 54.59 mg/dL (Phillips et al., 1998).

The original title for the bongo study was “Low-Stress Blood Sampling of Bongo Antelope.” Because this study was submitted in the mid-1990s, one reviewer absolutely would not accept this title. Many people in the zoo field did not want to admit that darting and forcefully restraining animals was highly stressful. One of my coauthors on both antelope articles was the zoo veterinarian. The animals feared him. He was the only person who could never work with our habituated antelope. Unfortunately, this observation was not included in the articles. It is difficult for a veterinarian who loves animals to admit that the animals feared him.

Introducing the word ‘fear’ in the veterinary and animal science research literature

One problem in many fields is silos. Each group of practitioners, such as veterinarians, animal scientists, neuroscientists, and ethologists, have their separate sets of journals. The ease of reviewing literature on the Internet has helped poke holes in these silos. In 1997, I wrote one of the first articles for animal scientists that reviewed neuroscience literature on fear (Grandin, 1997). Another review was published in the Journal of Animal Science by Morriss et al. (2011). I was a coauthor on this paper. Fear research has been around for decades, but it was new to many scientists who work with farm animals or animals in zoos. Many of the old experiments referenced in the 2011 review used highly invasive methods. Electrodes implanted in the animal’s brain were used to either stimulate a specific subcortical area or destroy it. Due to concerns about animal welfare, some of these experiments should not be replicated. However, knowledge gained from these experiments clearly showed that animals have emotions. The information gained from this research indicated that stimulation of certain areas elicits both behavioral and hormonal changes associated with rage and fear. Researchers who are not familiar with the older literature often emphasize that emotions such as fear activate a wide network of brain circuits. This is definitely true when functional magnetic resonance imaging brain scanning is used on people. However, neuroimaging cannot determine which node on a network turns on the entire system. The old electrode experiments were able to determine the subcortical node that was the primary activator of the fear network. Stimulation of nodes that were not the primary activator had no effect.

Panksepp core emotions

When Catherine Johnson and I wrote Animals in Translation and Animals Make Us Human (2009), we discovered Panksepp’s (1998) influential book Affective Neuroscience. The seven core emotional systems that he referenced made a lot of sense to me. The following is a list with major references for each emotion. This list provides the zoo professional with a concise, easily referenced list. (The next section will contain a more in-depth discussion.)

Fear (negative emotion; Davis, 1992, 1997; Gloor, Olivier, & Quesney, 1981; Kemble, Blanchard, Blanchard, & Takushi, 1984; Rogan & LeDoux, 1996): motivates animals to avoid danger. An animal will either flee, freeze, or fight when it is fearful.

Rage (negative emotion; Bard, 1928; Levison & Flynn, 1985; Olds, 1977; Panksepp, 1971; Siegal & Shaikh, 1997): anger to fight off a predator.

Panic (negative emotion, separation distress, oxytocin system; Herman & Panksepp, 1981; Panksepp, 2003; Panksepp & Bishop, 1981): vocalizations and other agitated behavior that occur when a mother and young are separated. Separation distress may also occur when a single animal is separated from the herd.

Seek (positive emotion, dopamine system; (Burgdorf & Panksepp, 2006; Ikemoto & Panksepp, 1999; Kluver & Bucy, 1939): the urge to explore. The seek system may motivate learning that is rewarded; tends to counteract fear.

Lust (sex; positive emotion; Luttge & Hall, 1973): motivation for mating; affected by androgens.

Nurture (positive motivation, oxytocin system; Lee, Macbeth, & Young, 2009): motivates a mother animal to care for her young. Animals grooming each other, pair bonding, and parental care may be motivated by this system.

Play (positive emotion; Panksepp, 2011; Panksepp & Burgdorf, 2000, 2003): Tickling rats is rewarding and mimics rough and tumble play.

Each one of these core emotions has subcortical brain areas that activate behavior (Panksepp, 1998, 2005, 2011). Earlier work by Joseph LeDoux clearly supported subcortical fear circuits (LeDoux, 2000; Rogan & LeDoux, 1996).

Scientific support for core emotions

This section summarizes research that supports the core emotional systems.

Fear

The amygdala is a central part of the fear system (Davis, 1992, 1997; LeDoux, 2000; Rogan & LeDoux, 1996). Electrical stimulation of the amygdala in a cat will trigger fearful behavior such as hissing (Fernandez, de Molina & Hunsperger, 1959). Fear motivates animals to avoid danger. Electrical stimulation of the human amygdala elicits the fear feeling (Chapman, Schroeder, & Buyer, 1954; Gloor et al., 1981). In animals, stimulation of the amygdala triggers the secretion of stress hormones (Matheson, Branch, & Taylor, 1971; Redgate & Fahringer, 1973; Setckleiv, Skaug, & Kaada, 1961). If the amygdala is destroyed, the animal will no longer fear an aversive stimulus (Davis, 1992). Complete destruction of the amygdala will tame rats in the wild (Kemble et al., 1984).

Rage

When the hypothalamus is electrically stimulated, cats and rats will become aggressive (Bard, 1928; Hess, 1957; Levison & Flynn, 1985; Panksepp, 1971; Siegal & Shaikh, 1997). This area of the brain is separate from fear. There are some overlapping brain areas that trigger both fear and rage (Ursin & Kaada, 1960). When Hess (1957) originally did his aggression experiments, previous researchers called the cat’s reaction “sham rage.” When I took general psychology as a college freshman, I was taught that the rage was not real. Later experiments showed that the rage was real. When the hypothalamus was stimulated, the cat failed to attack Styrofoam blocks but would attack a stuffed or anesthetized rat (Levison & Flynn, 1985). This finding showed that electrical stimulation triggered behaviors that a cat would perform in a natural setting.

Panic

Separation distress is distinct from fear (Panksepp & Bishop, 1981). Separation distress is triggered when a mother animal is separated from her young or an animal is separated from his or her conspecifics. Horses have individual differences in their reaction to being isolated (Lansade, Bouissou, & Erhard, 2008). Experiments with animal breeding clearly have shown that the fear and panic systems are separate. The two emotional systems were separated with clever breeding experiments. In the wild, where there are a lot of predators, an animal would likely be both high in fear and high in panic. Faure and Mills (1998) and Mills and Faure (1986, 1990, 1991) bred quail with the fear and panic traits separated. Fear was assessed with a tonic immobility test, and panic was measured with a moving sidewalk test. Birds with high separation distress ran longer on a moving sidewalk that moved them away from their flock mates. The researchers were able to breed either high- and low-fear or high- and low-panic animals. The panic system is tied in with the hormone oxytocin. Insel and Shapiro (1992) found that oxytocin receptors binding in subcortical areas was greater in voles who pair-bonded compared with voles who did not pair-bond. In humans, the social deficits in autism may be related to deficits in the oxytocin system (Zhang, Xu, Zhang, Han, & Han, 2017). The oxytocin system is also associated with friendliness in domestic dogs. Von Holdt, Shuldiner, and Jaowitz-Koch, et al. (2017) reported that similar genetic factors are associated with hypersociality in both domestic dogs and Williams syndrome. Wolves did not have this trait. People with Williams syndrome are super friendly, and they also have higher levels of oxytocin and decreased activation of the amygdala (Jarvinen, Lorenberg, & Bellugi, 2013).

Seek

Burgdorf and Panksepp (2006) conducted an extensive review on positive emotions and studies that support the seek emotion. Mouse breeders have bred strains of rats who are either high seekers or low seekers of novelty (Dellu, Piazza, Mago, LeMoal, & Simon, 1996). High seekers are more willing to walk around and explore their environment. Dellu et al. (1996) also discovered that high-seeking rodents had greater dopamine activity in the nucleus accumbens. New things are attractive if an animal can voluntarily approach them (Grandin, 1997). Cattle will quietly come up to a clipboard on the ground. If the paper flaps, they will back away and then cautiously approach again (Grandin, 1998). Knocking out the dopamine D4 receptor in mice reduces exploratory behavior (Dulawa, Grandy, Low, Paulus, & Geyer, 1999). More recent research clearly has shown that dopamine modulates novelty seeking in monkeys (Costa, Tran, Turochi, & Averbeck, 2014). There is an interaction between seeking and fear. If a person working with cattle makes a sudden movement, the cattle will jump back and then slowly come back when the person stays, “Still” (Grandin, 1997). I call this behavior “curiously afraid.” The nucleus accumbens has both fear and appetitive seeking circuits (Reynolds & Berridge, 2008). The nucleus accumbens has a circuit that can switch back and forth between fear and approach (Faure, Reynolds, Richard, & Burridge, 2008). There are also different circuits for wanting, which is activated by dopamine, and liking circuits, which are activated by opioids (Berridge, Robinson, & Aldridge, 2009; Smith & Berridge, 2007). A recent review showed that the core of the nucleus accumbens plays an important role in instigating approach behavior (Floresco, 2015). The outer shell may suppress nonrewarded behavior (Floresco, 2015).

Lust, nurture, and play

For the last three core emotional systems, there is less clear information. There is also overlap with the first four core emotions. Hori, Shimojo, and Tokunaga, et al. (2013) found that tickling rats and play triggered dopamine release in the nucleus accumbens. This finding would overlap with the seek emotional system. The nurture system is driven by oxytocin and involves behaviors that enable an animal to be successful at reproducing. Behaviors that are associated with oxytocin release are pair bonding, parental care of the young, and social grooming (Lee et al., 2009). The nurture system overlaps with the panic system. Animals who are strongly pair-bonded would be likely to have more intense reactions when they are separated. More solitary animals who form fewer pair bonds would have lower plasma oxytocin (Reser, 2013). Different species also vary in the degree of social inclusive chimpanzees, and lions are more group-adapted than tigers or orangutans (Reser, 2013). Lust is influenced by androgens (male hormones). There are both species differences and genetic differences within a species that have an effect on how androgens influence behavior (Luttge & Hall, 1973).

Still resistance in accepting animal emotions

In his later work, Joseph LeDoux questioned whether or not animals are really conscious and feel fear (LeDoux, 2014; LeDoux & Brown, 2017). Maybe the fear circuits are simply “survival circuits.” After reading these articles, I found two holes in LeDoux’s new views of his own research. He totally left out all the studies on pharmaceutical effects on animal behavior. Veterinarians often use human psychiatric medicine in dogs (Cannar et al., 2014; King, Simpson, & Overall, et al., 2000;Landsberg et al., 2008; Seksel & Lindeman, 1998). There is a significant amount of literature on how psychiatric medications such as Prozac (fluoxetine) have similar effects on animals and people (Overall, 1997). Research on individual differences and genetic influences on animal behavior has also been omitted. Rats can be bred to be either reactive to a stimulus or nonreactive (Broadhurst, 1975), and mice have been bred to be either high or low in anxiety (Muigg et al., 2009). Dellu et al. (1996) reported that differences in novelty seeking in rats may be similar to those of people. Lawrence, Terlouw, and Illius (1991) also documented many individual differences in pig behavior. Animals can also be bred to be either high or low in fear or high or low in separation distress (Faure & Mills, 1998; Mills & Faure, 1986). Pharmaceuticals that are used on people are often tested on animals to determine if a new substance may be effective for use in psychiatry. If the nervous system of a rat was totally different from that humans, this research would be useless. It has proven to be very useful. Hatherall, Sanchez, and Morilak (2017) found that an antidepressant reduced fear responses and enabled active coping in rats. Research with brain scans in humans showed that taking a single dose of an antidepressant reduced the response of the amygdala to pictures of angry faces (Murphy, Norburg, & O’Sullivan, et al., 2009). It also reduced the amygdala’s responses to angry-face pictures that were shown for a fraction of a second that the person did not consciously see (Harmer, Mackay, & Reid, et al., 2006;Sheline, Barch, & Donnelly, et al., 2001). This finding may have been due to the low-road fear circuit described in LeDoux’s older articles and the “survival circuit” discussed in his most recent articles (LeDoux, 2014; Rogan & LeDoux, 1996). I take antidepressants, and they stopped a constant feeling of fear (Grandin, 1995). It was like having my fear system volume control turned down, similar to turning down the heat on a thermostat.

The early experiments with brain-stimulating electrodes clearly showed that either stimulation or destruction of a subcortical area would turn on or turn off fear, rage, or seeking. More recent research that identified finer details of brain circuitry showed that the amygdala contains circuits that are also involved in nonfear behavior (Reynolds & Berridge, 2008). When I looked at a number of studies, the amygdala seemed to be more biased toward fear and the nucleus accumbens was biased toward seek. Modern methods have shown that the brain circuitry between the amygdala and nucleus accumbens is complex. Some people may question whether the old experiments are still valid. It is my opinion that they are. When large parts of subcortical structures were either stimulated or destroyed, obvious behaviors occurred such as fear, rage, or seeking exploration.

A music mixing board is a useful model for zoos

Panksepp’s seven emotional systems provide an easy-to-use framework that zoo professionals can use to provide better conditions for their animals. Each one of the seven emotional systems is influenced by both genetics and experience. I like to use the visually conceptual model of a music mixing board. The volume of each of the seven emotional systems can be either raised or lowered. In my work with animals, I have observed both individual differences and differences between cattle breeds in behavior. The music mixing board model makes it easy for me to understand these differences. It is also a model that most people who work hands-on with animals can easily understand.

In all species, there are differences in fearfulness and seeking. Some cattle become highly agitated during handling or will run fast when released from a restrainer (Curley, Paschal, Welsh, & Randal, 2006; Voisinet, Grandin, Tatum, O’Connor, & Struthers, 2007). Other cattle who have been raised under exactly the same conditions and handling will have a lower reaction. The more fearful reactive cattle had lower weight gain (Voisinet et al., 1997). Recently, I was conducting a cattle-handling lab with beef calves from our Angus herd at Colorado State University. One animal was really wild and agitated compared with his calm herd mates. His behavior was definitely influenced by genetics because all the cattle came from the same herd with the same previous experiences. Both the animal’s genetics and his or her previous experiences have an effect on behavior (Grandin, 1997; Grandin & Shively, 2015). Research with global positioning system collars has shown that some cattle are more active grazers and will walk farther than others (Goodman, Cibils, & Weley, et al., 2016). Researchers called them “go getters” and “laid-back” types of grazing patterns (Goodman et al., 2016). It is likely that the “go getters” are high in the seek emotion. Lions also have individual differences. Some lions are bold and are more likely to confront a possible intruder (Heinsohn & Packer, 1995).

Species differences in emotions

Behavior differences between different species can also be understood using a music mixing board as a conceptual model of emotions. The species of animals who pace in a zoo are often the animals who walk great distances in the wild. I have observed that polar bears are some of the worst pacers. More inactive animals or small animals may adapt better in a zoo environment. Smaller animals require less space to provide more natural habitats. In dogs, I have observed a big difference in the seek trait between Labrador Retriever dogs. One dog never stopped chasing a ball and repeatedly brought the ball up and dropped it to get people to throw it. The other dog, who grew up in the same neighborhood and played with the same kids, had no interest in the ball. It was obvious to me when I was a young child that the two Labradors, Tucky and Hunter, had different personalities. Until recently, as a scientist, I was not allowed to use words such as personality or emotions.

The path forward for zoos

When I was invited by Ron Kagan, chief executive officer and director of the Detroit Zoological Society, to write this article, I was asked to give my perspective on the path forward for zoo animal welfare. Some animal advocates recommend eliminating zoos. I do not agree with this recommendation. A full discussion of the pros and cons of zoos is beyond the scope of this article. In today’s electronic world, kids need to see real animals. Some animals are much easier to keep than others. Rose and Riley (2017) maintained that abnormal repetitive behavior may be part of a coping mechanism. This suggestion may add weight to discussions that some species may not be suitable for the zoo (Rose & Riley, 2017).

I provide some thoughts from my visits to zoos. I have observed wonderful chimpanzee and ape exhibits where the animals are housed in social groups. In my work with antelopes, I also observed that the antelopes, giraffes, and other grazing or herbivore animals often adapted well. Many small animals can have a good life in well-designed and well-managed zoos. Elephants are probably the most problematic. I recently visited an exhibit where the elephants were well treated and had lots of attention from their keepers. However, I could not stop thinking as I stood in front of this small exhibit that it was a well-run prison. In my book Animals Make Us Human, I wrote that the “homebody” animals do well in zoos and species who are “nomads” and roam large territories do more stereotypies. Sometimes, the solution to providing an animal in captivity with a better life is to do something that does not look “natural.” I watched Gus the polar bear in New York jump and play on numerous plastic barrels that were put in his pond (Grandin & Johnson, 2009). The exhibit looked like a child’s playpen, but Gus appeared to be having fun. It was definitely not natural, but it was probably effective. The polar bears in the Detroit Zoo are kept occupied by a variety of behavioral opportunities throughout each day, including catching fish thrown by the keepers. My personal reaction to this exhibit was very positive. Zoo enrichment programs are reducing stereotypies by 50% to 60% (Grandin & Johnson, 2009).

It may seem anthropomorphic to say that an animal is having fun or looks bored or anxious. A new science, quantitative behavioral analysis, may help prove that the normal viewer’s reaction to animal behavior is accurate (Saint-Anne et al., 2013). Some studies have shown that when numerous observers rated an animal’s behavior with common words such as anxious, relaxed, agitated, happy, or sad, their responses correlated with physiological measures. The responses from more than one observer were analyzed with statistical methods to determine an average response. Two studies showed that when videos of sheep and cattle during transport were described with common words such as “calm” or “relaxed,” physiological measures of stress were lower (Stockman et al., 2012; Wickham et al., 2010).

What is natural?

Another philosophical area is: What is natural? People may argue that Gus the polar bear is not living in a natural environment. A question one might ask is: What is natural? At a recent conference on range management, a speaker reminded the audience that in the United States, people have always had an effect on animal habitats. Native Americans had been living on the land long before the Europeans arrived.

In my travels, I have seen many signs of wildlife adapting to manmade structures. In Denver, CO, the sparrows have chosen to live in the glass roof of the central concourse at the airport. Rabbits inhabit the parking lot, and I saw a coyote in my doctor’s parking lot. The pigeons really like the concrete cliffs that people built for them in the airport parking garage. During a trip to the Chinese countryside, I saw many things. Many of the birds have been eaten by hungry people, but a few lucky birds found safe living spaces. When I flew into a Chinese airport, I saw beautiful pheasants who found a refuge within the airport grounds. Airport security had provided them with protection. They also had sparrows in the terminal building. While driving in the countryside, I saw numerous nests of a huge bird who nests in a safe location on electrical towers. The birds learned that people will not climb to the top. These individual animals had excellent welfare, but their habitat was definitely not natural. It is likely that some animal species may no longer exist in China because they were killed by people looking for food. There is a need for better wildlife conservation programs in China and many other countries.

Conclusions

Animal welfare and being “natural” are two different things. These two topics need to be discussed as separate issues. There is much evidence that animals have emotions. People working with animals will be able to better provide for the animals for whom they care if they can determine the emotion that motivates a behavior. The framework of the Panksepp emotional systems should be very helpful. For example, animals who are highly fearful may need secluded places to hide, and extra care must be taken to avoid frightening them when they are conditioned to cooperate with veterinary procedures. A species who forms strong pair bonds would be more likely to become highly stressed when separated from their mate than would a species with low pair bonding. An active animal who is a high seeker may need more novel activities compared with a low-seeking animal such as a koala bear. Sometimes a behavior can have an unexpected motivation. When gerbils frantically dig, they are not exploring, but they are trying to hide to avoid predators (Wiedenmayer, 1997). They are probably motivated by fear instead of seeking. Figuring out the correct motivation will facilitate the development of more effective environmental enrichments.