My wife and I keep pet rabbits. Observe their cuteness:
We feed Jackson (he’s the black one) and Dutchess (she’s the big one) once each morning and once each night, and usually give them a few treats in between. A month or so ago, we noticed that when we open the refrigerator door they hop up from wherever they are and run right over to us, necks craned to see what we’re going to pull out. More recently, Dutchess has taken to standing up and putting her front paws on my leg whenever I’m holding a bag of treats. I think this is very cute, so I give her treats when she does it.
If you’ve ever had a pet, you’ve almost certainly seen behaviors like this. A good trainer can teach an animal to perform all sorts of tricks using food as a reward. A bad trainer can turn the same animal into a nuisance (albeit, an adorable one) who persistently begs and butts in where it’s not wanted. We are of the second type, though by some divine grace our rabbits have turned out just as sweet as can be.
Jackson and Dutchess seem to know that there is a good chance that they will get food when they see me open the refrigerator – at least, they act like it. They seem to really want the treats, and because of this we can infer that they must really like to eat the treats. This all seems very simple and intuitive, but the field of behavioral neuroscience, which studies how the brain contributes to and controls an animal’s behavior, has a long history of studying the not-so-simple ways that the brain makes animals – humans included – like and want things.
A simple question
Let’s step back for a moment.
The central question of science is: why do things happen? More specifically, why do things happen the way they do, instead of happening in another of the infinite ways they could possibly happen? For example, physicists ask questions like: "when we drop something, why does it move down instead of up, or some direction in between?" They can then do experiments and come up with theories (like the theory of gravity) to figure out just what it is about the world that makes this certain event – an object falling – happen the way it does. If we’re talking about animals, the question instantly becomes more specific: "why do animals do things?"
We are now in the realm of behavioral neuroscience. Behavioral neuroscience is a branch of psychology that is focused of describing how various arrangements of neurons (the cells that make up the brain, spinal cord, and nerves) contribute to the various things that animals do. Now we have an even tighter view of the problem: "how do animals’ nervous systems cause them to do things?"
The first step in answering this question is to look at some of our basic assumptions about behavior. We might say that behavior is simply what an animal does. This seems to be an okay definition at face value, but it misses something. Is the beating of a heart "behavior"? What about the movement of a white blood cell? A better definition might include the idea that behavior is a response to some part of an animal’s environment – it involves the animal sensing some part of their environment and reacting to it.
With this definition, we see that there is always some element of choice involved. In any situation, an animal is physically capable of doing many different things – the thing it chooses to do is its behavior. Its nervous system somehow helps it choose which behavior is best for each situation. Our question has now been refined yet again: why an animal does one behavior at a certain time instead of another behavior?
For example, let’s suppose that Jackson is pleasantly snoozing on the floor, like so:
At this point, he could do a number of things – he could keep snoozing, he could groom himself, he could move about, he could groom Dutchess, he could find and eat some food, he could drink water, he could play with toys, and so forth. Because he is snoozing, we can say that he is definitely not grooming himself, eating, running towards me, or doing anything else. When I open the refrigerator door, however, he hops right up and shuffles over to me as quickly as his tiny bunny legs will carry him. A decision has just been made – it is the job of behavioral neuroscience to explain how Jackson’s brain has decided that this is what he should do at this moment.
One can divide the answers to this question into two categories: proximate causes and ultimate causes. Proximate causes are the events that directly precede a behavior (ie. Jackson runs and hides because I have dropped a frying pan on the ground and made a very loud sound, which nerve cells in his inner ear sensed and which was then turned into a signal in his brain that told his muscles to move in a certain way.)
Ultimate causes, on the other hand, describe the reason an organism does the behavior in the first place: it’s goal (ie. Jackson hides because, if he were in a dangerous environment, it would be much safer for him to be far away from the source of a loud noise, which is likely to have come from a larger animal or some other source of danger.)
The ultimate causes of behavior are generally well-known. Sure, examples pop up every now and again that defy current theories, but we know what most behaviors do for the animal: hunting and foraging find food, fighting defends territory or resources from intruders, social interaction and bonding promote safety and opportunities to mate, and so forth.
How about the proximate causes of behavior? We know that the brain sends electrical signals down the spinal cord to muscle fibers, which is what actually causes an animal’s muscles to move. To look at the way behavior is controlled overall, we can look at the brain. There are some areas of the brain that control specific behaviors – for example, the areas of the brain that are necessary for speaking and understanding language in humans have been studied for a long time, and some areas of the rodent brain that control repetitive behaviors like grooming and licking have been pretty extensively studied. In general, we can say that animals do certain complex behaviors because certain parts of their brain are active.
This is an imprecise answer, but it gives us a place to look. Of course, any specific behavior will be caused by specific areas of the brain, we must look at each of these in turn to understand how the brain creates behavior as a whole.
We still need some connection between the proximate causes and the ultimate causes of behavior; some process that links reason an animal does something with the parts of the animal’s brain that make its body move in a certain way. The key component of this link is motivation.
By "motivation," I don’t mean what you get from your friend who does Tai-Bo with you and tells you how good you are at it (although I totally dig that kind of motivation, and I love that friend). I’m talking about the processes in the brain that tell an animal how urgently it needs to do different things. When we say an animal is hungry or thirsty, we mean that there is something about the state of animal’s nervous system that makes it more likely to perform certain behaviors than others.
Often this can be explained by looking at what has happened to the animal in the past. For example, Dutchess is much more likely to drink when she has not had anything to drink in a while, or when it is a particularly hot and dry day and she is losing water faster than normal. Similarly, she will run to the refrigerator more quickly and excitedly if she has not eaten for a while – in psychological jargon, we would say that she is very motivated to get food; she is hungry. We would also say that the food is very rewarding to her, and so it encourages her to repeat whatever behavior preceded it (in this case, jumping up my leg.) Here, we see Dutchess in just such a strongly motivated state, trying to get a piece of food from my hand:
Now that we’ve come up with a simple answer to the question of why animals do things ("it’s the brain!"), we can work on a more complete answer. Ideally, we want to describe exactly which parts of the brain determine what is most important for an animal to do at a particular time. This turns out to be a difficult and complicated problem.
A simple answer?
To make the situation simpler, let’s move from the relatively complicated world of a pet to the well-controlled world of a scientific laboratory. A rat (like we might study in a psychological laboratory) easily learns to do a few things as it develops – he can walk, eat food, groom himself, have sex, fight with other rats, and so forth. He does not need to learn much about these behaviors – he can do them pretty well, though not perfectly, the first time he tries (provided that he is physically able to.) Other things, though, he must learn how to do.
Imagine that we have a rat who leads a pretty drab life – we keep him in an opaque box, and feed him pellets each day through a slot in the top of the box. One day, we put him in a new box, one with a button in the center of the floor. Whenever the rat steps on the button, a piece of food drops into one of the corners of the box. The first time this happens, it happens by accident – the rat has simply stumbled on to the button. Nevertheless, he eagerly eats the food, and at a specific place in his brain a set of electrical impulses fire (most theories say this happens in a set of structures referred to as the basal ganglia, including structures like the ventral tegmental area and the nucleus accumbens.) These impulses say to the rest of the brain, albeit in code, that something good has just happened.
At the same time, areas of the rat’s brain that are responsible for memory are active. Over countless millennia of evolution, brains have become very logical (at least when it comes to getting food, or anything else that we’d call "pleasurable" or "rewarding",) and so the brain tries to figure out what it can do to get more food. Since the last thing the rat did was step on the button, a set of changes happen in the rat’s brain cells that links the action of stepping on the button with the goodness of the food. A single repetition of this cycle is not usually enough for learning to take place, since the changes that take place after a single event are probably rather subtle, but after more and more cycles of stepping on the button and eating the resulting food, our rat comes to learn that one event causes the other. At this point, the brain signals that go along with the act of stepping on the button and those that go along with the pleasurable experience of eating become linked to each other – a "memory trace" has been formed, and the animal’s behavior in the future will reflect this (importantly, he will eat more tasty food.)
We now have a way of measuring how much the rat wants to get food at any particular time – how motivated he is. Suppose that we increase the number of times he needs to press the button to get some food – he gets food after the first press, but then it has to press the button twice for a single morsel, then four times, then eight times, then sixteen times, and so on. We can figure out how motivated our rat is to eat by how far he goes before he decides that it’s not worth it to keep trying to get the food. If we force-fed our rat, we would find him unwilling to do any button-pressing to receive food; he would simply not be hungry enough. If we starved our rat, we would find him very eager to press the button.
Variations of this technique have been used to study learning and motivation in vertebrates for many years – by this point, we have a rough understanding of which structures and chemicals do what in this process. We know that parts of the cortex, the wrinkled outermost part of the brain, are involved in making decisions, planning movements, and then sending electrical signals to muscles to actually do the movements. These parts of the cortex have lots of connections with deeper parts of the brain that deal with various parts of the decision-making process – the limbic system, which is involved in emotion, the hippocampus and amygdala, which are involved in memory and emotion, and the basal ganglia, which are involved in learning and the control of movement. All of these systems act together to come up with a plan of action in response to a particular setting, which is then sent out to the muscles and turned into a behavior.
To make the task of explaining how these decisions happen a bit easier on ourselves, let’s leave out decisions that are based on a lot of thought, like choosing a career path or deciding which insurance plan to buy. Let’s stick to simple decisions – say, ordering dinner off of a restaurant menu. If I were at dinner with you, and I asked you how you decided on a certain dish, you hopefully wouldn’t explain that uttering the phrase "I’ll have the southwestern omelet" had been closely followed by rewarding experiences enough times in the past that you had learned that this is the best thing to say to a server if you are hungry. Rather, you would say something like "It tastes good," or "I like to eat eggs." Your answer would be about pleasure, and specifically about which food you think will bring you the most pleasure.
This seems intuitive – we do things that make us feel good. In terms of the nervous system, we seem to do things because they stimulate parts of the brain which, when stimulated, make us feel some sort of pleasure. In the same way that you eat a certain food because it brings you more pleasure than another, Jackson digs through his food bowl and eats the tastiest pieces first, presumably because they cause certain parts of his brain to be active in a way that ends up being pleasurable to him. Let’s take a moment to observe this process before continuing (for the sake of science, of course):
What’s the brain got to do with it?
Behavioral neuroscientists have been working to figure out how the brain controls motivated behavior for many years. In 1957, two psychologists named James Olds and Peter Milner published a report on their discovery that, if they put electrodes in a specific part of a rat’s brain (the medial septum) so that they could stimulate it with electricity, the rat would work to receive this stimulation in the same way it would work to receive food or water. They had discovered what seemed to be a center for reward (and/or pleasure) in the brain. Stimulation here appears to signal to animals when something is good and should be sought out.
Later, they found that the same effect occurred when the electrode was placed into a group of nerve fibers in the brain called the medial forebrain bundle, which connects the ventral tegmental area (a small area in the middle of the brain) to the nucleus accumbens (another small area in the middle of the brain.) This stimulation causes signals to be sent to the nucleus accumbens, where the chemical dopamine is released. Because a rat will work for this sort of stimulation, we know that the stimulation is rewarding to the rat; it has some property that makes the rat want more of it.
To get a better handle on what we’re talking about, let’s zoom in on the relevant part of the brain first.
This is an MRI image that shows a slice through the middle of somebody’s head. That green spot is the approximate location of the ventral tegmental area, or VTA. It receives connections from many other areas of the brain. It also sends connections out to many areas of the brain, including the nucleus accumbens, or NAc – that green line represents these fibers, the ones in which stimulation can be rewarding.
The nucleus accumbens (its location is shown in pink) receives connections from the ventral tegmental area as well as the cortex and other areas of the brain, and sends connections out the basal ganglia (their approximate location is highlighted in yellow), which control movement, and these connections eventually get to the prefrontal cortex (outlined in blue), which is involved in planning and decision making. The nucleus accumbens can be divided into a core and a shell (with the shell wrapped around the core, like a boiled egg’s white is wrapped around the yolk.) We’ll come back to this in a bit.
So we know that we do things because they tend to cause a certain kind of activity in certain areas of the brain. It seems obvious that we do things because they make us feel good. It seems reasonable that the brain activity that encourages us to do things also plays a role in making us feel good. In fact, the circuits that the nucleus accumbens is part of seem to be central to both the pleasure we feel and the motivated behavior that it persuades us to do (like getting up to cut the evening’s third piece of pie while you’re in the middle of writing a blog post, because it just tastes so good.)
This strikes me as a rather profound area of scientific research for a few of reasons. For starters, it represents a sort of triumph in science’s quest to address big, important questions (like "What makes us happy?", and "Why do we do the things we do?") While the many questions raised in this field aren’t all answered yet, the reward system has been and continues to be a fertile ground for coming up with new ways to study and think about the brain.
Secondly, studying pleasure, reward and motivation gives us a language with which to talk about not only the behavior but the experiences of animals. It’s important, if we are to share a world with organisms that we cannot talk to, to be able to understand how they feel things and why they feel them as well as why they behave the way that we do.
Lastly, pleasure and motivation are central to living, and especially central to living a good life. Certainly pleasure is not equivalent to happiness, but it is hard to imagine happiness existing without pleasure. It is pleasure (and displeasure) that gives texture and humanity to what would otherwise be a machine-like, purely functional existence. To imagine a world where people lacked the ability to like and dislike things is, in fact, imagining a world devoid of people as we know them – the abilities to like and to want are what give life meaning. It is uniquely humbling to realize that something so central to human experience as pleasure arises from the activity of neurons, from a great, intricate, mostly-undiscovered dance of energy and molecules inside each of our heads.
All that sentimental stuff aside, we’ve come to a problem. If "pleasure" (a subjective feeling) and "reward" (a signal that teaches us to do things) both cause the same behaviors, and both arise from the same sort of activity in the same part of the brain, are they different at all? Why do we need to talk about "pleasure" at all, especially when we’re talking about animals (like my pet rabbits, who can’t talk to us and tell us what they feel, no matter how much I may wish they could)? Why talk about animals "experiencing" things when it’s so much simpler to talk about animals "responding"?
Psychology has a long history of regarding pleasure and reward as the same thing, accidentally or not. For psychologists who were (and are) concerned primarily with measuring behavior, it often seems unimportant or even absurd to talk about pleasure, when "reward" describes the same thing but doesn’t imply that the behaviors you are studying rely on anything as ethereal as a "feeling". Even if you recognize that the organism probably feels something that goes along with reward, "pleasure" and "reward" appear to be so close to the same thing that there is no clear way to separate them – they have the same effects, they are caused by the same stimuli, and they even (using some simple measurements) appear to arise from the same structures in the brain.
The story, it turns out, is not so simple.
Liking, wanting, and the nucleus accumbens
Three researchers (Kent Berridge, Terry Robinson, and J. Wayne Aldridge) recently published a paper summarizing a line of research they and many others have been working on since at least the early 1980′s. Their findings argue that liking something and wanting something are actually dissociable in the brain – that they rely on the activity of different neurons, so you can have one without the other. As they say in the 2009 research report: "Usually a brain ‘likes’ the rewards that it ‘wants’. But sometimes it may just ‘want’ them."
A corollary to this is: "Sometimes the brain may just ‘like’ things," a notion which has led some to propose that the world would be better if we all had electrodes inside of our head that would stimulate our brains in a pleasurable way. It turns out, as Berridge and Morten Kringelbach mention in a later publication, that the few ethically questionable experiments that have been done using brain self-stimulation in humans did not seem to produce happy people – sometimes they produced people who were very motivated to push a button to get brain stimulation, and sometimes (as is the case with more recent studies of brain stimulation for depression) they have helped people to not feel quite as bad as they did.
Besides being interesting to philosophers and utopian dreamers, this area of research is of interest to healthcare practitioners and their patients because figuring out what makes somebody "like" doing something and what makes them "want" to do something may help us treat disorders in which these processes go wrong – for example, in depression, obsessive compulsive disorder, and drug addiction.
Before we get too worried about the implications of this idea, though, let’s see what sort of evidence there is to back it up.
First, the researchers needed a way to measure how much pleasure an animal gets from something. More importantly, they needed to measure how much an animal likes something without measuring motivation – that is, in a way that doesn’t rely on the animal seeking out or consuming that thing. To illustrate the problem, think of our button pressing rat – in that model, we are measuring how much the animal wants the food. It may seem that an animal would only work for things it likes, but if we want to show that wanting and liking can be separated from each other, we need an independent measurement of each.
It turns out that it’s easy to measure how much an individual likes something, if you use the right kind of thing. Tastes turn out to be well-suited for this. Think about it: a good way to identify whether somebody likes or dislikes the taste of something is to watch its face as it tastes the substance. It turns out that this technique can be used (more or less successfully, depending on who you ask) to determine whether human infants, primate infants, and rats like or dislike a taste. Of course, the Youtube-watching public is already familiar with the use of this technique in humans:
Somehow, I doubt this would be a useful approach to use with my rabbits, whose faces seem to be permanently fixed in a single expression: grumpy.
Berridge, Robinson, and Aldridge argue that human infants, some primates, and rats make similar faces when they like or dislike a taste, and that these faces can quantified in a way that produces useful scientific data. Using this measurement, as well as more traditional measurements of how hard animals will work for a reward, they and other researchers have shown that there are situations in which an animal’s apparent ‘like’ of a food item can be separated from its ‘want’ for that food.
For example, in one experiment Berridge and his collaborator Elliot Valenstein investigated why a certain type of electrical brain stimulation causes animals to eat. When they gave rats this type of stimulation (of the lateral hypothalamus) it caused some of the rats to seek food more strongly than rats who didn’t get the stimulation. After selecting a group of rats in whom this stimulation caused feeding, the researchers tested whether that stimulation affected the faces they made in response to different tastes.
If pleasure and motivation always go together, one would expect that as the animals were more motivated to eat food, good-tasting things would also taste better to them. In fact, the rats’ facial reactions to sweet tastes did not markedly change while they were receiving brain stimulation. Their responses to sour tastes did change in some cases, indicating that bad tastes actually tasted worse during the brain stimulation. It appeared that the stimulation made the rats ‘want’ the food, but it did not make them ‘like’ the food any more.
In another study Berridge and his coworker Susana Pecina wanted to know if liking and wanting were caused by different groups of brain cells. Opiate drugs (like morphine, which Berridge and Pecina used in their study) both feel good and are rewarding – they affect both pleasure and motivation. Prior to this study, it had been shown that when opiate drugs are put into any part of the nucleus accumbens, they encourage rats to eat – they made the rats more motivated to eat. The experimenters wanted to find out, though, what sort of effect the same drugs had on the pleasure (or the "hedonic impact") of eating when injected into the nucleus accumbens.
Their procedure was elegant: they implanted needles into the nucleus accumbens of rats, then gave them injections of morphine directly into the nucleus accumbens just before giving them either sugar or quinine, a bitter chemical found in tonic water. By varying the location of the implanted needle, they built up a map of nucleus accumbens showing the places where morphine affected how much the rats liked or disliked the tastes. Using a procedure where rats got morphine injections just before being allowed to press a lever to receive food, they built up another map showing where in the nucleus accumbens the morphine made the rats want the food more.
When morphine was injected anywhere in the nucleus accumbens, it made the rats spend more time finding, manipulating, and eating food – it increased their motivation to obtain and eat food. It was only when morphine was injected into a smaller area in the center of the nucleus that the rats showed more positive reactions to sweet tastes, though. This means that in a certain area in the shell of the nucleus accumbens, morphine injections made the rats want food more, but didn’t make them like the taste of it any more.
These results make the case that liking and wanting can be separated from each other both in behavior and in the anatomy of the brain. It’s interesting to note that there was no region of the nucleus accumbens in Berridge and Pecina’s map that was solely responsible for the pleasure-enhancing effects of morphine – they could get wanting without liking, but not the other way around.
In fact, pleasure and motivation appear to be distinct processes in many more ways than this, even if they are intertwined with each other both in everyday life and in the anatomy of the brain. Berridge’s works describe "hotspots" of pleasure in the brain, which are specific areas where certain kinds of activity or drugs produce pleasure in ways that may or may not be associated with changes in motivation – these include the nucleus accumbens and VTA, as I mentioned above, but also parts of the cortex and other structures in the brain. By performing many experiments investigating the functions of many areas of the brain, researchers are building maps that describe where, why, and how our brains determine our behavior. These maps represent the first steps in answering that deceptively simple question: why do we do what we do?
Filling in the holes
It is good to keep in mind that the findings I’ve discussed are incomplete – it turns out that many of the brain’s “pleasure centers” have a variety of functions, none of which are as simple as I’ve made out the role of nucleus accumbens in pleasure to be. In addition, pleasure and reward, like most functions of the brain, appear to arise from incredibly complicated arrangements of cells spread throughout the brain. It is impossible to take very complex behaviors (like pursuing goals and enjoying reaching them) and collapse them down into a certain kind of activity of a small structure in the brain.
There is so much research going on in this area of neuroscience that it’s impossible for me to tell the whole story – I hope I have conveyed some of the excitement of it, though. These studies stand as a proof-of-concept that wanting something and liking it may be very different processes. If nothing else, the problem of distinguishing pleasure from reward serves as an example that even the most intuitively sensible ideas about behavior are likely to be wrong in some way, or are at least worth looking at critically.
As more studies have been (and continue to be) published, such results are disputed, argued over, expanded, and corrected. The first hundred-some years of neuroscience have left us with a fragmented and rapidly changing picture of how motivation works, and I suspect that the research done in the next hundred years will continue to fill in the holes in our map of the brain, even if it’s never completed.
Nevertheless, I still have that profound, romantic feeling about the study of motivation and pleasure. It gives me many small moments of awe to think about all the complex things that go into motivation, whether it’s a toddler standing on her tip-toes to reach the top shelf of the refrigerator where the chocolate milk is, a shopper deciding which car to buy, a rabbit begging for a treat, a rat pressing a button, or myself deciding what to eat. Most of all, it reminds me of the reason that I was drawn to study neuroscience in the first place: to understand humanity, to understand why we do what we do, we must understand the body, and most of all the brain.
OLDS J, & MILNER P (1954). Positive reinforcement produced by electrical stimulation of septal area and other regions of rat brain. Journal of comparative and physiological psychology, 47 (6), 419-27 PMID: 13233369
Berridge KC, Robinson TE, & Aldridge JW (2009). Dissecting components of reward: ‘liking’, ‘wanting’, and learning. Current opinion in pharmacology, 9 (1), 65-73 PMID: 19162544
Berridge, K. (1996). Food reward: Brain substrates of wanting and liking Neuroscience & Biobehavioral Reviews, 20 (1), 1-25 DOI: 10.1016/0149-7634(95)00033-B
Berridge, K., & Valenstein, E. (1991). What psychological process mediates feeding evoked by electrical stimulation of the lateral hypothalamus? Behavioral Neuroscience, 105 (1), 3-14 DOI: 10.1037//0735-7044.105.1.3
Peciña S, & Berridge KC (2005). Hedonic hot spot in nucleus accumbens shell: where do mu-opioids cause increased hedonic impact of sweetness? The Journal of neuroscience : the official journal of the Society for Neuroscience, 25 (50), 11777-86 PMID: 16354936
PECINA, S. (2008). Opioid reward ‘liking’ and ‘wanting’ in the nucleus accumbens Physiology & Behavior, 94 (5), 675-680 DOI: 10.1016/j.physbeh.2008.04.006
Image credits: All photos except for the MRI image (which was released under an Creative Commons Attribution-Sharealike 2.0 generic license by Flickr user everyone’s idle ) were taken by Michael Lisieski, and are licensed under a Creative Commons Attribution-NonCommercial0ShareAlike 3.0 Unported License.
About the Author: Michael Lisieski is an undergraduate student pursuing degrees in Pharmacology and Psychology in Buffalo, New York, where he lives with his wife, three rabbits, and a leopard gecko. His major professional interests are in studying and treating drug addiction. He writes about cephalopods in science and culture at Cephalove, and tweets about everything else as @cephalover.
Secrets of the Universe: Past, Present, FutureX