Incentive Sensitization - "Wanting" vs. "Liking"

The distinction between wanting and liking was first articulated by neuroscientist Kent Berridge through a series of elegant and somewhat disturbing experiments on rats in the 1990s. Berridge selectively destroyed the dopaminergic systems in rats' brains — the circuits conventionally associated with pleasure — and found something unexpected: the rats stopped seeking food, but when food was placed in their mouths, they still displayed normal pleasure responses. Destroying the opioid systems produced the opposite result: the rats still sought food compulsively but showed no pleasure upon receiving it. Wanting and liking, Berridge concluded, are distinct neurological processes, operating on different substrates, and they can come apart entirely.

The wanting system — mediated by dopamine — is fundamentally an anticipation and pursuit system. It responds to cues that predict reward, motivating approach behaviour. Critically, it is the system that learns through variable reinforcement: cues associated with unpredictable rewards become powerful motivational triggers. The liking system — mediated primarily by opioid and endocannabinoid circuits — processes the actual hedonic experience, the felt quality of receiving the reward.

Incentive sensitization occurs when the wanting system becomes hypersensitive to cues over time, particularly through repeated exposure to variable rewards. The cue — a notification badge, the pull-to-refresh gesture, the app icon on the home screen — comes to trigger powerful wanting that is disproportionate to any pleasure the reward actually delivers. The compulsion escalates even as satisfaction diminishes. Users describe this as the social media experience with unusual precision: the urgency to check, followed by the hollow vacancy of actually having checked.

This dissociation is what makes volitional restraint so unreliable as an intervention. The wanting system generates motivational states that precede conscious deliberation — the urge is felt before the decision is made. Moreover, the experience of not enjoying the behaviour does not weaken the wanting, because wanting and liking are not in feedback with each other. A user can check Instagram fifty times in a day, feel worse each time, and wake up the next morning with the wanting fully restored. The liking system's verdict is simply not heard by the wanting system.

Platform design interacts with this architecture precisely. Notification cues, streak mechanics, like counts, and the visual grammar of social feeds are engineered to function as incentive cues — stimuli that trigger dopaminergic anticipation. Over time and with repeated exposure, these cues sensitize the wanting system. The practical consequence is that willpower — which requires the prefrontal cortex to override motivational states generated by subcortical systems — is fighting a losing battle against a system designed to produce exactly this outcome.

Berridge's research suggests the intervention point is cue management rather than impulse control. If the cue triggers the wanting before the decision, removing the cue removes the trigger. This is why blocking apps at the system level is more effective than relying on in-the-moment restraint: it interrupts the cue-wanting chain before it can execute, rather than asking willpower to defeat a neurological state it was never equipped to reliably overcome.