← Reference library
REFERENCE LIBRARY

Basal GangliaBasal Ganglia

A set of deep-brain nuclei doing something fierce: by default they hold down every action you could take, until one option wins — and only then do they open a gate for that one. They are the gatekeeper of action selection, and the place dopamine pours its learning signal into.

In one line: a gate that's "off by default"

Intuitively, moving means "send a signal telling the muscles to go." The basal ganglia do it backwards: their output nuclei are constantly sending "brakes" downstream, pinning every candidate action in place. To let one action happen, you don't hit the gas — you specifically release the brake on that one, while the rest stay held. So "choosing" is physically "release exactly one, veto the rest."

Cortex · candidate actions Striatum · inputcaudate + putamen Output GPi/SNrbrakes by default Thalamus → release actionthalamus direct · Go release brake indirect · No-Go more brake back to cortex · loop SNcdopamine · teaching signal
Cortex→striatum→(Go/No-Go)→output→thalamus→back to cortex; dopamine modulates learning from the side

Two opposed pathways: Go and No-Go

Once the striatum receives candidate actions from the cortex, it splits two ways — one urging, one blocking:

Direct pathway (Go)
Activating it releases the brake on an action — the thalamus lets it through and the action happens. Its neurons carry D1-type dopamine receptors and get more excited when dopamine arrives.
Indirect pathway (No-Go)
Activating it presses the brake harder — suppressing actions that shouldn't happen. It carries D2-type receptors and is instead damped when dopamine arrives.
The balance
What you do or don't do depends on whether Go or No-Go is louder. Dopamine simultaneously "fuels Go and loosens No-Go," so more dopamine biases toward "move, do it"; less biases toward "hold still."

Dopamine: not just a go signal, but a "teaching signal"

The deepest layer of the basal ganglia is that it doubles as the brain's trial-and-error learning machine. The dopamine coming from the substantia nigra / ventral tegmental area carries exactly the reward prediction error — "how much better/worse this was than expected." That error pours into the striatum and rewrites the memory of "in this situation, how valuable is doing this action": surprisingly good → strengthen the Go pathway just used, so next time you want it more; worse than expected → weaken it. Over time, good habits (and bad ones) get etched bit by bit into this circuit.

When it breaks: Parkinson's and beyond

When the dopamine-making cells in the substantia nigra die off in large numbers, that's Parkinson's disease: with no dopamine to fuel Go and loosen No-Go, the scale is jammed hard toward "don't move" — hence difficulty initiating movement, slowness, rigidity, and a resting tremor. Conversely, dysregulation of this circuit reaches widely: OCD, tics (as in Tourette's), addiction, and some impulse-control problems can all be traced to this "gate" being shut too poorly or too tightly. It never governs movement alone, but also the thoughts and habits that get selected as if they were "actions."