A loop running from the midbrain into the ventral striatum and back out to the prefrontal cortex. It's often called the "pleasure centre," and that name is almost entirely wrong — what it computes is "is this worth approaching," not "how good does this feel."
The circuit starts with a small cluster of midbrain cells (the ventral tegmental area, VTA) that manufacture dopamine and spray it forward onto a few key stations. The main line lands in the nucleus accumbens (the core of the ventral striatum), then runs through the ventral pallidum and thalamus back up to the prefrontal cortex — which in turn sends a heavy set of descending connections back down onto the accumbens. It's a closed loop, not a one-way street.
VTA supplies dopamine, the accumbens integrates, the loop returns to prefrontal cortex — which pushes back down
VTA (ventral tegmental area)
One of the brain's dopamine factories. What its dopamine carries isn't content but reward prediction error — how much better or worse than expected. The branch to the accumbens is the mesolimbic pathway; the branch to the frontal cortex is the mesocortical pathway.
Nucleus accumbens
The junction box. It stacks three streams — how much it's worth (dopamine), what situation this is (hippocampus), what it means emotionally (amygdala) — and outputs "approach or not." It divides into core (action and cue learning) and shell, which houses the small patches that handle actual pleasure.
Ventral pallidum
The output gate. It's an obligatory waypoint for wanting, and it also contains hedonic hotspots — lesion it and sweet taste can turn into an aversive reaction.
Prefrontal cortex (medial / orbitofrontal)
The upper storey: it wires reward valuation into planning and inhibition. Its descending glutamate connections press down on the accumbens — and when that line weakens, you have the circuit basis of impulsivity and loss of control.
Amygdala · hippocampus · interoception
Three inputs: emotional significance, where you are (context), and what your body is currently doing. Which is why the same stimulus is worth wildly different amounts depending on mood, place, and whether you're hungry or full.
Wanting ≠ liking: two systems in one building
Two mechanisms live inside this circuit and are constantly conflated:
Wanting (incentive salience) — dopamine-driven, broad, easy to amplify, and about going after things. Liking (hedonic reactions) — handled by a few very small hedonic hotspots in the accumbens shell and ventral pallidum, running on the brain's own opioid and cannabinoid molecules: small, fragile, and hard to turn up.
That accounts for two everyday facts: pleasure has a ceiling and craving doesn't, and it is entirely possible to want something intensely while enjoying it not at all — which is exactly the shape of addiction, the two curves coming apart.
How it relates to the basal ganglia
Same architecture, different floor. The basal ganglia diagram — cortex → striatum → output nuclei → thalamus → back to cortex — has a ventral version here: the ventral loop computes "is it worth it, should I approach," the dorsal loop handles which action to run and which habits to keep. And they aren't parallel: the connections between dopamine cells and striatum form a spiral, letting influence climb step by step from ventral to dorsal. That spiral is the anatomical route for the progression from wanting, to habit, to compulsion.
Three common misreadings
❌ Dopamine = pleasure
No. Dopamine-depleted animals still show normal "this tastes good" reactions; they just stop pursuing. It's a motivation and learning signal, not the pleasure itself.
❌ This is the "pleasure centre"
The early experiments where rats self-stimulated relentlessly were read as finding a pleasure centre. Later analysis favours a different reading: what was being ignited was wanting — the animal chasing an anticipation that never arrives.
❌ Only drugs use this pathway
Food, sex, social approval, money, information, and the swipe that promises a better next clip all run through the same circuit. What makes drugs special isn't that they use it — it's that they pry it open pharmacologically, bypassing the prediction and saturation every natural reward has to pass through.