← Reference library

REFERENCE

Neurotransmitter Systems

The chemical messengers neurons use to talk to each other. What matters is that they split into two very different jobs: one carries content, point to point; the other changes how a whole region listens. Keep those apart and it becomes obvious where the "chemical imbalance" metaphor goes wrong.

Two modes: what is said vs how it's heard

The gap between two neurons is a synapse. The upstream cell releases molecules, receptors on the downstream cell catch them, and the signal is across. But "releasing molecules" covers two completely different uses:

Point to point (fast transmission) — the job of glutamate and GABA. The first makes the downstream cell more likely to fire (accelerator), the second less likely (brake). The reach is one synapse wide and the timescale is milliseconds. The overwhelming majority of synapses in the brain use these two, and they carry the information itself.

Diffuse (modulation) — dopamine, serotonin, noradrenaline, acetylcholine. These come from a handful of tiny nuclei in the brainstem and midbrain, tens of thousands of cells at most, whose axons spray across the entire brain like a sprinkler truck. They usually don't order a downstream cell to fire or not; they change its gain — how sensitive a region is to input, and how readily its connections can be rewritten. Releasing a molecule to spread through the extracellular space so that any nearby receptor can pick it up even has its own name: volume transmission.

The main modulatory systems

Glutamate · main excitatory

The brain's principal excitatory transmitter. Its NMDA receptor doubles as the key switch for learning: it opens only when the upstream cell is firing and the downstream cell is already depolarised — a coincidence detector by construction.

GABA · main inhibitory

The principal inhibitory transmitter. Inhibition isn't "off", it's sculpting: deciding who gets suppressed and when others get through. Brain rhythms are largely shaped by it, and many sedatives act on this line.

Dopamine · midbrain

Often called the pleasure molecule; more accurately it codes prediction error and salience — "better than expected / worth noticing" rather than pleasure itself. It serves both learning and the initiation of movement.

Serotonin (5-HT) · raphe nuclei

The receptor family is enormous (well over a dozen), which is why "serotonin levels" is close to meaningless — the same molecule can have opposite effects at different receptors. Tied to patience and waiting, to the long-run tone of mood, and to sleep.

Noradrenaline · locus coeruleus

The master control for arousal and vigilance. What it really adjusts is brain-wide signal-to-noise: a moderate rise sharpens focus, too much turns into jumpiness where everything grabs you.

Acetylcholine · basal forebrain / pons

The gate for attention and plasticity. When it arrives, cortex becomes more willing to rewrite itself according to current input — which is why it's often described as the "learning is permitted" signal. Its shifting levels across the night are also one of the things that separate sleep stages.

Neuropeptides & endocannabinoids

Slower, longer-lasting messengers. Endocannabinoids even run backwards: released by the downstream cell to regulate how much the upstream one releases — a built-in negative-feedback brake.

Why "too little of X" is almost always wrong

Three reasons, each worse than the last.

1. The effect is set by the receptor, not the molecule. The same dopamine landing on a D1-type receptor and on a D2-type receptor can do opposite things downstream. So "dopamine is high / low" is an incomplete sentence — without naming the pathway and the receptor it carries no meaning.

2. These systems tune, they don't supply. The language of modulation is timing: when it's released, onto which pathway, for how long. Compressing that into a single level is like compressing a piece of music into its average volume.

3. The system adapts. Change a transmitter's level for long enough and receptor numbers and sensitivity shift with it (up- and down-regulation). That's why many psychiatric drugs take weeks to work, and why stopping abruptly can rebound — what actually changed is the operating point of a whole regulatory loop, not an instantaneous concentration.

Which settles a common confusion: a drug working doesn't establish the cause. SSRIs act on the transporter that carries serotonin back out of the synaptic cleft, and they raise cleft serotonin within hours — yet relief takes two to three weeks. That delay is itself the evidence that the therapeutic step lives in slow downstream processes, not in the act of topping something up.