Neurotransmitters 101

Two transmitters do almost all of the brain's raw signalling. Glutamate is the main excitatory transmitter, the go signal that drives most fast communication and is essential for learning and memory; see NMDA, glutamate and memory. GABA is its opposite, the main inhibitory transmitter, the brake that keeps the whole system from over-firing.

The two are not symmetrical, and the asymmetry has consequences. Glutamate has no extracellular enzyme that destroys it; the only significant way to remove it from the space between cells is for transporter proteins in glial cells and neurons to pull it back in. That single fact makes those transporters responsible for keeping glutamate concentrations non-toxic, and their failure is central to what goes wrong in stroke, hypoglycaemia and several degenerative diseases ^[6]. GABA is cleared the same way, but with a much smaller margin for disaster.

Signalling with these two is also startlingly expensive. In grey matter, action potentials account for roughly 47% of the energy spent on signalling and the postsynaptic effects of glutamate for another 34%, with the resting potential taking about 13% and recycling glutamate itself only 3%. Raising the average firing rate by a single action potential per cortical neuron per second increases oxygen consumption by about 145 mL per 100 g of grey matter per hour ^[7]. The brain cannot simply run everything harder, and that constraint shapes what any drug can plausibly do.

Nearly everything else is a neuromodulator: a transmitter that does not carry the raw signal itself but changes the tone, the gain and the flavour of it. The best summary of what that means comes from the invertebrate work where it was first made precise: the wiring diagram provides a minimal structure, and the neuromodulatory environment then constructs and specifies the functional circuits that actually produce behaviour ^[15]. The same anatomy, differently modulated, is a different machine.

Dopamine is the transmitter of motivation, effort and drive; see the dopaminergic system. Serotonin shapes mood, satiety and patience; see the serotonergic system. Acetylcholine runs attention, encoding and signal-to-noise in the cortex. Norepinephrine drives alertness and the fight-or-flight response. Most stimulants and antidepressants raise one or more of these; most calming agents raise GABA or lower glutamate.

The table is the map. Read the receptor column as the important one: it is the receptors, not the transmitter, that decide what a compound does.

Two of these deserve a correction that is repeated often enough to be worth stating plainly here. Dopamine is not the pleasure molecule. Recordings from midbrain dopamine neurons show they are activated by rewards that are better than predicted, unaffected by rewards that are exactly as predicted, and depressed by rewards that are worse than predicted; that is a teaching signal about surprise, not a signal of enjoyment ^[9]. The related theory separating wanting from liking makes the same point from the behavioural side: the systems that make something feel worth pursuing can be sensitised independently of the systems that make it feel good, which is how craving can grow while pleasure shrinks ^[10].

Norepinephrine is not simply arousal either. All of the brain's noradrenaline comes from one small brainstem nucleus, the locus coeruleus, and its neurons run in two modes: phasic bursts tied to the outcome of a decision, which favour sticking with the current task, and a tonic mode associated with disengaging and looking for something better ^[13][14]. That is a much more specific job than turning the volume up, and it explains why noradrenergic drugs can improve focus and increase distractibility depending on where a person starts.

Every psychoactive compound in this wiki acts at one of six steps. Learning the steps is more useful than learning drug names, because it tells you in advance what a new compound's side effects and interactions are likely to look like.

The release step is worth a moment on its own, because its speed is the reason brains work at all. When an action potential reaches a terminal, calcium enters and binds synaptotagmin, which triggers a core fusion machinery of SNARE and SM proteins to merge the vesicle with the membrane in under a millisecond. RIM proteins hold primed vesicles at the active zone and simultaneously recruit the calcium channels, so that calcium ions flow more or less directly from the channel onto the sensor ^[3]. That architecture is what turns an electrical event into a chemical one without losing timing.

Two general rules fall out of the table. Blocking clearance is the most common drug action in psychiatry, because a transporter is an easier target than a receptor and raising an existing signal is gentler than imposing a new one. And acting early in the cycle produces weaker, more diffuse effects than acting late: a precursor raises the ceiling on supply but does not decide when or where the transmitter is used, which is a large part of why amino acid supplements underperform the drugs that act on receptors and transporters.

Receptors come in two architectures, and the difference decides almost everything about how a drug feels.

Ionotropic receptors are the channel. Transmitter binds, the pore opens, ions move, and the whole event is over in microseconds to milliseconds. AMPA, NMDA and kainate receptors for glutamate work this way, as do GABA-A receptors and nicotinic acetylcholine receptors. They are what fast, point-to-point signalling is made of, and a drug acting on them tends to have an immediate and unmistakable effect.

Metabotropic receptors are a switch that starts a cascade. Transmitter binds a G-protein-coupled receptor, which activates a G protein, which changes the concentration of a second messenger, which changes the behaviour of enzymes and channels elsewhere in the cell. That takes tens of milliseconds to minutes, amplifies the original signal enormously, and can alter gene transcription. Dopamine, serotonin, adrenergic, muscarinic, opioid, cannabinoid and adenosine receptors are all of this type; see how to read a chemical class for why that shared architecture makes their pharmacology rhyme.

The consequence worth carrying away is about subtypes. Ionotropic glutamate receptors are built from eighteen gene products that coassemble in different combinations ^[4]. GABA-A receptors are pentamers drawn from nineteen possible subunits, and a careful catalogue of the ones that actually exist in tissue lists at least twenty-six ^[5]. Serotonin has around fourteen receptors, and even after decades of medicinal chemistry several of them have no selective ligand, which means their function is largely unknown ^[11]. Dopamine has five, and their signalling runs through at least two independent pathways ^[8].

So a sentence like raises serotonin is close to meaningless on its own. Extra serotonin reaching a 5-HT2A receptor on a cortical pyramidal cell and extra serotonin reaching a 5-HT1A autoreceptor on a raphe neuron produce opposite results, and a drug that raises the transmitter everywhere does both at once. The same objection applies to raising dopamine, raising GABA and raising acetylcholine. Whenever a source makes a claim at the level of a transmitter rather than a receptor, it has skipped the step that determines the answer.

It is tempting to think that more of a good transmitter is better, but the brain runs on balance, not maximums, and each system has a specific failure mode when pushed.

Too much glutamate is toxic to neurons. Excitotoxicity is not a metaphor; sustained receptor activation floods cells with calcium and kills them, which is why the transporters that clear glutamate are so central to the pathology of ischaemia and hypoglycaemia ^[6]. Too much GABA suppresses breathing, which is the mechanism behind fatal overdoses of sedatives, and the reason combining two GABAergic compounds is one of the few genuinely dangerous interactions in this space; see anxiolytics and GABA. Chronically forced dopamine flattens motivation, because a system built to signal surprise adapts to a constant signal by treating it as the new baseline ^[9].

Every one of these systems also adapts to being pushed, by removing receptors from the surface or making them less responsive. That is the shared mechanism behind tolerance, and it is why the effect of a compound on day sixty is a different question from its effect on day one; see tolerance, dependence and down-regulation.

This is also the honest frame for interaction warnings. Most of the serious ones in this wiki are not exotic chemistry. They are two compounds pushing the same system too far in the same direction: two serotonergics, two GABAergics, two things that raise heart rate. Knowing which lever a compound pulls tells you in advance which other compounds are the ones to be careful with, which is the entire logic behind interactions and stacks.

This field is often presented as settled, and a great deal of it is not. Four gaps are worth knowing about because they change how to read almost any claim about a transmitter.

The chemical imbalance account of mood is not supported. A systematic umbrella review of the main lines of serotonin research in depression, covering metabolite concentrations, receptor and transporter imaging, tryptophan depletion, and genetic association studies including one analysis of 115,257 people, found no consistent evidence of an association between serotonin and depression and no support for the idea that depression is caused by lowered serotonin activity ^[16]. Antidepressants can work without that theory being right; the theory itself did not survive the evidence. Treat any explanation of a mood effect that rests on a transmitter being low or high as a hypothesis rather than a finding.

Human measurement is almost entirely indirect. There is no way to watch transmitter release in a living human brain. What exists is receptor and transporter imaging with radioligands, metabolite concentrations in cerebrospinal fluid, and microdialysis in animals, each of which measures something adjacent to the quantity of interest. Most confident statements about a compound raising a transmitter in humans are inferences from animal work.

Neurons release more than one transmitter. Co-transmission is common rather than exceptional, and the same cell can release a fast transmitter and a slow modulator at different firing rates, which means labelling a pathway by a single chemical is a simplification that hides part of what it does.

Context decides the effect. The neuromodulation literature is clear that the same modulator can reconfigure the same circuit into functionally different machines depending on state ^[15], and that acetylcholine's effects depend on where it is released, which receptor subtype is present, and which population of neurons is targeted ^[12]. That is why the same compound helps one person concentrate and makes another anxious, and why an average effect size in a trial can be small while individual responses are large in both directions.

None of this makes the map useless. It makes it a map rather than the territory: good enough to predict which compounds will interact, which side effects to expect, and which claims to distrust, and not good enough to predict what any one substance will do to any one person.