Wanting, Liking, and the Architecture of Reward

Kent Berridge, Terry Robinson, and colleagues, University of Michigan

Reward is usually spoken about as though it were one thing. Something feels good, so we want it. We learn where to find it, we pursue it, and the whole process gets folded into the same word. Kent Berridge and Terry Robinson spent decades pulling those pieces apart.

Their work separates reward into at least three interacting processes: liking, wanting, and learning. Liking refers to the hedonic impact of a reward. Wanting refers to incentive salience, or the motivational pull that makes something grab our attention and become worth pursuing. Learning carries the associations, predictions, and memories that tell us where the reward is and how to get it. Those processes normally overlap, but they can separate in ways that become very obvious in addiction.

One of the experiments behind this framework is hard to forget. Berridge and Robinson depleted dopamine in the nucleus accumbens and neostriatum of rats by as much as 99 percent and then measured their hedonic reactions to sweet and bitter tastes. The animals still showed normal positive reactions to sucrose and negative reactions to quinine. They could also learn new changes in palatability. The dopamine loss devastated motivated behavior while leaving much of the hedonic response intact. That finding helped separate dopamine from the simple idea of pleasure and placed it much closer to incentive salience, the process Berridge calls wanting.

That distinction gives addiction a very different shape. A person can reach a point where they barely enjoy the drug and still experience an enormous pull toward it. The cue itself can acquire motivational force. A place, a person, a smell, money in the pocket, a piece of paraphernalia, a particular emotional state, or even a time of day can suddenly change what the nervous system treats as urgent. Robinson and Berridge describe repeated drug exposure as capable of sensitizing the systems that assign incentive salience. Once sensitized, those systems can produce excessive wanting long after the pleasure associated with the drug has diminished. The effect can persist and can flare according to context, stress, cues, and internal state.

There is another piece of this that gets buried by the way we commonly talk about the “dopamine reward system.” The nucleus accumbens sits near the center of most diagrams of mesolimbic dopamine, but roughly 90 to 95 percent of its neurons are GABAergic medium spiny neurons. Dopamine reaches the accumbens largely through projections from the ventral tegmental area. The principal cells receiving and processing that information are inhibitory GABA neurons, many organized according to D1 and D2 dopamine receptor expression.

The accumbens is also receiving major glutamatergic input from cortical and limbic systems, including the prefrontal cortex, amygdala, hippocampus, and thalamus. Those inputs carry information about context, memory, goals, emotion, and the environment. Dopamine can alter how much motivational weight gets attached to that information. The GABAergic accumbens neurons integrate those signals and help organize the resulting behavioral output. Calling the whole thing a dopamine system compresses a fairly sophisticated integration problem into the name of one transmitter.

Berridge’s hedonic work adds another layer. Pleasure itself appears to depend on much more restricted circuitry. His group identified small hedonic hotspots within areas including the medial shell of the nucleus accumbens and the ventral pallidum. Opioid stimulation within those particular regions can amplify positive hedonic reactions. The territory capable of generating enhanced liking is relatively small compared with the broader systems capable of generating wanting.

That architecture has stayed in the back of my mind because it fits several pieces of the reward-system work I have been developing.

In my 2025 paper on ibogaine, I proposed reward-system recovery as a process involving dopaminergic regulation, glutamatergic signaling, GDNF, and plasticity within mesocorticolimbic circuitry. I used the term reward fidelity to describe the ability of the system to assign value more appropriately again. Berridge’s work gives me a more precise way to think about what fidelity would actually require.

Reward fidelity cannot be reduced to the amount of dopamine present in a circuit. A cue can be accurately remembered and still carry pathological motivational weight. A reward can still produce pleasure while failing to motivate pursuit. Something can become intensely wanted without being especially liked. A person can also lose interest in ordinary rewards while remaining exquisitely responsive to a narrow set of learned cues. These are different failures inside the same larger system.

The cellular organization of the nucleus accumbens also helps explain why I keep coming back to glutamate and GABA alongside dopamine. A reward-related cue arrives through networks carrying information about what the cue is, where it has appeared before, what happened the last time it was encountered, and what behavior has historically followed it. Dopamine influences the motivational significance assigned to that information. The accumbens then processes that combination through a predominantly GABAergic population whose output helps influence what happens next.

That gives us a much richer problem than trying to decide whether dopamine is simply high or low.

The preliminary UCI MIND-OUD results are interesting in this context. In the first five participants, the team reported reduced basal ganglia activation to opioid cues after treatment, alongside reductions in craving and other clinical measures. The full study is explicitly designed to examine opioid-cue reactivity, reward-circuit connectivity, spectroscopy, and EEG before and after treatment.

If that reduction in cue response survives the larger sample, I want to know exactly what changed. Did the opioid cue lose incentive salience? Did the participant still recognize and remember the cue normally while experiencing less motivational pull toward it? Did ordinary rewards become more capable of competing for attention and behavior? Did reinforcement learning change? Did effort allocation change? How quickly did the nervous system disengage once the cue disappeared?

Those questions are much closer to what reward-system recovery actually means to me.

They also shape how I think the next studies should be designed. Reward anticipation, hedonic response, effort expenditure, cue-triggered craving, reinforcement learning, prediction error, behavioral inhibition, and recovery after cue exposure can all be measured separately. Imaging can show where activity changes. EEG can show timing and dynamics. Behavioral tasks can tell us whether the person actually values and responds to reward differently. Repeated measurements can show whether those changes hold.

Berridge’s work gives me language for something I have been trying to get at for a long time. Addiction can distort the motivational value assigned to information. Recovery should eventually be visible in the way that value is assigned again.

That is what I mean by reward fidelity.

References

Berridge, K. C., & Robinson, T. E. (1998). What is the role of dopamine in reward: Hedonic impact, reward learning, or incentive salience? Brain Research Reviews, 28(3), 309–369. https://doi.org/10.1016/S0165-0173(98)00019-8

Berridge, K. C., Robinson, T. E., & Aldridge, J. W. (2009). Dissecting components of reward: “Liking,” “wanting,” and learning. Current Opinion in Pharmacology, 9(1), 65–73. https://doi.org/10.1016/j.coph.2008.12.014

Berridge, K. C., & Robinson, T. E. (2016). Liking, wanting, and the incentive-sensitization theory of addiction. American Psychologist, 71(8), 670–679. https://doi.org/10.1037/amp0000059

Peciña, S., Smith, K. S., & Berridge, K. C. (2006). Hedonic hot spots in the brain. The Neuroscientist, 12(6), 500–511. https://doi.org/10.1177/1073858406293154

Robinson, T. E., & Berridge, K. C. (2025). The incentive-sensitization theory of addiction 30 years on. Annual Review of Psychology, 76, 29–58. https://doi.org/10.1146/annurev-psych-011624-024031

Scofield, M. D., Heinsbroek, J. A., Gipson, C. D., Kupchik, Y. M., Spencer, S., Smith, A. C. W., Roberts-Wolfe, D., & Kalivas, P. W. (2016). The nucleus accumbens: Mechanisms of addiction across drug classes reflect the importance of glutamate homeostasis. Pharmacological Reviews, 68(3), 816–871. https://doi.org/10.1124/pr.116.012484

Nicolas, M. (2025). Ibogaine’s potential role in supporting reward system recovery across diagnostic boundaries. Frontiers in Pharmacology, 16, 1744383. https://doi.org/10.3389/fphar.2025.1744383