Drug addiction is a huge problem across the world, leading to large societal costs. The estimated annual cost of all drugs of abuse (including alcohol and tobacco) in terms of crime, lost productivity, and health care is $740 billion. In 2017, over 11% of those over age 12 in the US reported using illicit drugs in the past month. Alcohol and tobacco abuse are also prevalent with nearly 6% of US adults having an alcohol use disorder and 14% of US adults smoking cigarettes

Drugs of abuse are powerfully addictive because they “hijack” biological processes that  ensure we continue behaviors that promote our survival. The primary biological process all drugs of abuse have in common is that their initial use increases  dopamine in the brain, albeit via different mechanisms. 

Given these data, it is not surprising that in popular culture dopamine is thought of as the “reward” signal in the brain. But what does that mean exactly? 

 Human hand writing Dopamine chemical formula.
All drugs of abuse increases dopamine in the brain when used for the first time. Dopamine is a neurotransmitter commonly referred to as a “reward” signal in the brain.

Certain environmental stimuli elicit approach responses, especially under biologically-based “need” conditions. Put another way: Stimuli that are desirable as a result of a biological need are “rewarding”. Food is rewarding when we are hungry; water, when we are thirsty. 

Our brains are primed to learn about reward, specifically to learn what stimuli and actions in our environment will lead to survival, a process known as reinforcement learning. We learn certain behaviors are rewarding as they enable us to continue living.

Reinforcement learning involves associating a stimulus with a response that then leads to a reward. This type of learning is done by virtually all animals. For example, a lab rat can learn that when a light comes on and it presses a particular lever in a specific environment it receives a food pellet. In humans these responses are a bit more complex and abstract but follow the same general principles. For example, you may associate meeting your friend at your favorite bar with having a few of your preferred alcoholic drinks, even if you don’t feel like drinking on a particular night.

And once this learned stimulus-response association is made, the stimulus itself can be perceived as “rewarding” in a process known as incentive salience (more on this later). 

Reinforcement Learning, Reward, & Addiction

Stimulus-response learning drives most of the behaviors we think of as “drug addiction” in people. When one is addicted to a drug of abuse, stimuli associated with the use of the drug can themselves drive drug use behavior (think of the neighborhood bar for an alcoholic, or a friend you routinely smoke with when together). This is true even when the actual use of the drug is no longer “pleasurable”. In fact, most drug-addicted individuals do not find drug use  “pleasurable” any more. 

This is because addiction progresses from a binge/intoxication stage to a withdrawal/negative affect stage and finally to a preoccupation/anticipation stage which can then reactivate drug use. Thus, drugs are initially used because they are pleasurable, but over time this shifts and individuals use drugs of abuse to relieve negative withdrawal effects instead. Drug use can also be triggered by stimuli that were associated with use in and of themselves and these stimuli can promote preoccupation with drugs even in those trying to avoid them. 

Why does this happen? How can a drug that starts out as pleasurable lead to negative feelings of withdrawal when not used? Well, the brain is very adaptive and quickly modifies the biological environment such that there is less disturbance in dopamine (and other chemical) signaling after drug use. So, while drugs of abuse initially result in a large release of dopamine, this effect moderates with continued use. This is known as tolerance and explains why those addicted to drugs of abuse need to take larger and larger quantities of the drug to achieve the same effect. In fact, the continued use of addictive drugs results in notable changes in the brain dopamine system which promotes a strong biological dependence on them. 

While drugs of abuse initially result in a large release of dopamine, this effect moderates with continued use. This explains why those addicted to drugs of abuse need to take larger and larger quantities of the drug to achieve the same effect.

Dopamine signals predictors of rewards and not rewards themselves

Much research has shown that dopamine does not signal “reward” (or to be more technical, pleasure) per se but rather is used in learning the various predictors of reinforcement in the environment – reinforcement learning. 

This concept of dopamine signaling reinforcement learning was most famously demonstrated by the work of Wolfram Schultz, a professor at the University of Cambridge in the UK, who recorded the firing of dopamine-producing brain cells in the brain of primates receiving reinforcing juice rewards. 

Initially, these neurons fire to unexpected reward (juice) delivery. If a cue (tone or light) perfectly predicts the juice delivery over repeated trials,  the dopamine neurons fired in the presence of the cue and not the reward. And when a reward is not followed by a stimulus previously paired with it, there is an observable dopamine “dip” locked to the time when the reward was expected to occur. 

Follow-up studies have also shown the amazing ability for dopamine-producing neurons to encode reward prediction in a scaled manner (stronger dopamine response to higher probability predictors of reward) and has resulted in perhaps the most well-accepted computational model for a biological process: temporal difference model for reinforcement learning. 

Wanting vs Liking and Dopamine’s Role in Perpetuating Addiction

Work by Kent Berridge at the University of Michigan has demonstrated that dopamine motivates behavior and affects how hard animals are willing to work for rewards (“wanting”)

The processes of reinforcement learning described above naturally occur but the extra boost of dopamine release associated with taking an addictive drug further strengthens stimulus-response associations. Cues or stimuli that predict drug use can then themselves become “rewarding” and trigger wanting/craving responses in the brain as it anticipates drug use. And via other dopamine-related processes, drug use behaviors can become habitual, being guided by stimuli and the environment more than one’s active choice to use drugs. 

Glass of whiskey with ice on colorful Christmas lights bokeh background
Cues or stimuli, such as an old neighborhood or bar, can be enough to trigger drug cravings.

While many may conflate liking with wanting and the role of “reward” in all of this, the implications around the role dopamine plays in these processes are critical, especially if one is working to develop treatments to combat drug addiction. Compulsive drug use despite negative consequences is what results in addictive drugs negatively interfering in someone’s life NOT the pleasure the drug provides. So, a better understanding of what processes mediate wanting and craving for drugs of abuse is essential as we seek to combat drug addiction. 

My own work has sought to understand how the release of dopamine after oral d-amphetamine affects the human brain. We found that dopamine release correlates with participants “wanting more” (NOT “liking”) d-amphetamine in three core brain regions often associated with reward and drug-related effects: ventral striatum (VS), ventromedial prefrontal cortex (vmPFC), and insula. The VS is a core brain hub of reward valuation along with the vmPFC. Others have also found a relationship between VS dopamine release and “wanting”. The insula is a region of the brain often associated with drug craving/wanting and, in fact, damage to this part of the brain results in a loss of craving for cigarettes in smokers. Future efforts to modulate these craving-related systems and their associated dopamine signals through interventions such as transcranial magnetic stimulation may ultimately help drug-addicted individuals effectively stop their problematic drug use. 

We are just beginning to understand the neurobiological bases of drug addictive processes but continued research into them promises the development of better treatments in the future. 

Concluding Thoughts

Hopefully this piece has illustrated the complex role dopamine plays in signaling reward. Research that has emerged over the last few decades using sophisticated techniques to measure brain signalling in animals and humans has implicated dopamine in reinforcement learning processes and, by extensive the incentive salience of cues associated with rewards. The role of dopamine in signaling what stimuli predict reward is hijacked and pushed into overdrive by drugs of abuse that themselves release dopamine. Thus, after repeated pairings of stimuli and drug rewards, the brain adapts to respond powerfully to drug-related stimuli and cues, prompting craving in addicted individuals. 

Not everyone is as susceptible to these dopamine-mediated learning processes, though. How individual differences in biology ultimately map onto risk for drug addiction is a matter of intense interest in the field of neuroscience but is beyond the scope of this current post. For the time being, I encourage you to explore the references below for more on the complex and nuanced role dopamine plays in reward and learning processes.   



A Neural Substrate of Prediction and Reward


Neurobiology of addiction: A neurocircuitry analysis


Liking, Wanting and the Incentive-Sensitization Theory of Addiction


Pleasure Systems in the Brain


Learning, Reward, and Decision Making


Neural mechanisms underlying the vulnerability to develop compulsive drug-seeking habits and addiction


Dopaminergic Mechanisms in Actions and Habits


Imaging genetics and the neurobiological basis of individual differences in vulnerability to addiction