← Reference library

REFERENCE

Motor System & Cerebellum

Between "I want to move" and a muscle actually contracting sit one trunk line, two side loops and one return path — plus a counter-intuitive fact: there is no clean body map on motor cortex.

The trunk line: cortex → spinal cord → muscle

Every voluntary movement squeezes through the same exit. The stations along the way are not messengers; each one rewrites the command:

Primary motor cortex (M1)

The strip just in front of the central sulcus. Its neurons are not one-cell-per-muscle; each is broadly tuned around a preferred direction, and direction, speed and force live in the distribution across a population. That property is exactly what makes motor intent decodable by a BCI.

Corticospinal tract

About a million axons descending from cortex, crossing to the other side in the medulla — which is why the left hemisphere drives the right body. In primates some fibres synapse directly onto spinal motor neurons; the stronger that direct link, the more independently the fingers move.

Spinal motor neurons

Sherrington's final common path: cortex, brainstem, reflex arcs and the cord's own rhythm generators all converge to vote on this one cell. Nothing contracts until it fires.

The cord computes too

Pulling your hand from a hot pan, the automatic recovery step after a stumble, the basic rhythm of walking — all can be produced in the spinal cord without cortex. Cortex handles whether and how skilfully, not every step's detail.

Cerebellum: a machine for predicting how it should feel

The cerebellum occupies a modest corner of the brain yet holds more than half of all its neurons. Its job is best read like this: every time cortex issues a command, the cerebellum receives a copy and uses it to compute in advance what the body should feel like once the movement lands — a forward (internal) model.

When the real sensation returns, it is compared against that prediction. Match, and nothing happens; mismatch, and the difference becomes an error signal (delivered by climbing fibres) that revises the next prediction. People with cerebellar damage are not paralysed, but their movements wobble, overshoot and miss — they can only correct by watching, having lost the head start. The same mechanism explains why you cannot tickle yourself: the prediction cancels self-generated sensation precisely.

Basal ganglia: they don't produce movement, they release it

The basal ganglia are a set of deep nuclei whose main business is not how to move but whether, and which one: by default they hold the brake on every candidate action, and the selected one gets its brake released. Dopamine sets how tight that brake is — which is why Parkinson's (loss of dopamine neurons) shows up as difficulty starting and slowness, not as weak muscles. → basal ganglia

Counter-intuitive: the homunculus is wrong

The textbook body map laid along motor cortex — Penfield's homunculus — has been drawn for eighty years as a continuous band running from foot to face. Recent high-precision imaging finds that it is not continuous: the band is chopped into segments by three regions that control no specific muscle at all, and those regions instead connect to networks governing whole-body posture, arousal and internal state.

So motor cortex interleaves two kinds of thing: effector-specific patches and patches that treat the whole body as one event. That accounts for something long overdue an explanation — why preparing to move always comes bundled with shifts in heart rate, breathing and posture. Movement was never a matter of one muscle.