Key facts
- What it is
- Slow, diffuse chemical signalling that changes the gain and state of whole circuits
- How it is released
- By volume transmission into the extracellular space, not into a narrow cleft
- Receptors
- Mostly metabotropic, G-protein-coupled, slow to start and long to last
- Classical systems
- Dopamine, noradrenaline, serotonin, and acetylcholine
- Source
- Small nuclei in the midbrain, brainstem, and basal forebrain with vast projections
What neuromodulation is
Consider a strange arithmetic. The locus coeruleus, the sole source of noradrenaline for the entire cortex, is a tiny nucleus in the pons containing on the order of tens of thousands of neurons in the human brain: a rounding error against 86 billion. Yet a change in its firing rate can transform the behaviour of the whole cortex, sharpening perception, narrowing attention, and shifting the balance between focused exploitation and restless exploration. A structure with a negligible share of the brain's cells exerts a global influence on almost all of them.
This is possible because neuromodulatory neurons do not obey the usual economics of connection. A cortical pyramidal cell contacts its neighbours and a few distant targets. A locus coeruleus neuron sends a single, extraordinarily branched axon that ramifies across enormous swathes of cortex, dropping release sites as it goes. One cell, many hundreds of thousands of targets. The system is built for broadcast, not for correspondence.
Neuromodulation: chemical signalling that alters the responsiveness of a neuron or circuit to other inputs, rather than driving it directly. A modulator rarely makes a neuron fire. It changes how likely that neuron is to fire when something else tells it to, how strongly its synapses will change with use, and how much of its input it treats as signal rather than noise.
The most useful metaphor, and metaphors here are risky, is the difference between the notes played and the settings on the amplifier. Fast glutamate and GABA signalling is the music: specific, temporally precise, information-bearing. Neuromodulation is the gain, the tone control, and the reverb. It does not tell the circuit what to compute. It tells it how to compute.
How it differs from fast transmission
The contrast with classical synaptic transmission is sharp on almost every axis: where the transmitter goes, how long it acts, what kind of receptor reads it, and what the effect actually is.
| Property | Fast transmission | Neuromodulation |
|---|---|---|
| Spatial reach | Point-to-point across a cleft roughly twenty nanometres wide | Diffuse; the transmitter spreads through extracellular space to many cells |
| Timescale | Onset and decay in milliseconds | Onset in hundreds of milliseconds; effects lasting seconds to minutes |
| Receptor type | Mainly ionotropic: the receptor is itself an ion channel | Mainly metabotropic: G-protein-coupled, acting through second messengers |
| Effect on target | Drives it: pushes the cell towards or away from threshold | Tunes it: changes excitability, gain, and readiness to change |
| Information carried | A specific message to a specific cell | A global state variable: alert, motivated, patient, attentive |
| Main transmitters | Glutamate and GABA | Dopamine, noradrenaline, serotonin, acetylcholine, neuropeptides |
The receptor difference is the mechanistic root of everything else. An ionotropic receptor is a channel: transmitter binds, the pore opens, ions flow, and the voltage changes within a millisecond. A metabotropic receptor opens nothing. Binding activates a G protein, which activates an enzyme, which produces a second messenger such as cyclic AMP, which activates a kinase, which phosphorylates a target protein. Every step takes time, and every step amplifies, so a small amount of transmitter can produce a large and persistent change inside the cell. Some of those changes reach the nucleus and alter gene expression, which is how a modulatory signal lasting seconds can leave a trace lasting hours.
Note the caveat that runs through all of this. The categories are not airtight. Acetylcholine acts on fast ionotropic nicotinic receptors as well as slow muscarinic ones. Glutamate has metabotropic receptors alongside its ionotropic ones. Dopamine forms genuine synapses in the striatum. Neuromodulation is a mode of signalling, not a fixed list of molecules, and the same molecule can work in both modes in different places.
Volume transmission
At a conventional synapse the architecture is exquisitely constrained. The presynaptic release site sits directly opposite a dense cluster of receptors on the postsynaptic membrane, separated by a gap so narrow that a released vesicle's contents reach a very high concentration almost instantly and are cleared just as fast. The design is for privacy and precision: this cell speaks to that cell, and no one else hears it.
Neuromodulatory axons frequently do not do this. Along their length they bear swellings called varicosities, which contain vesicles and release machinery but often face no postsynaptic specialisation at all. Transmitter released from a varicosity spills into the extracellular fluid and diffuses outward, its concentration falling with distance, reaching receptors on any cell within range that happens to express them. The signal is not addressed. It is published.
Why volume transmission changes the logic: in a point-to-point synapse the wiring diagram determines who hears what. In volume transmission the wiring diagram is almost irrelevant; what matters is which cells express the receptor. A single release event can act on excitatory neurons, inhibitory interneurons, and astrocytes at once, and can produce opposite effects in each, because dopamine acting on a D1 receptor and dopamine acting on a D2 receptor do different things. The consequence is that a neuromodulator does not simply turn a circuit up or down. It reconfigures the relationships within it.
This also explains the timescale. Clearance from a synaptic cleft is nearly instantaneous, because the volume is minute and transporters ring it. Clearance from the extracellular space is slow: the transmitter must diffuse back to a transporter or be broken down by an enzyme such as monoamine oxidase, and in the meantime it lingers. Modulatory signals therefore have a natural persistence built into their geometry.
The four classical systems
Four neuromodulatory systems dominate the literature, and they share a common architecture that is worth stating explicitly, because it is unusual and it is the source of their power. Each originates in a small, anatomically discrete nucleus. Each sends highly branched axons to a very large fraction of the forebrain. Each acts primarily on metabotropic receptors. And each, when its firing changes, changes the operating state of the brain as a whole.
Dopamine
Two midbrain nuclei, two broad jobs. The substantia nigra pars compacta projects to the basal ganglia and is essential for the initiation and smoothness of movement; its degeneration produces Parkinson's disease. The ventral tegmental area projects to the striatum and prefrontal cortex and carries a learning signal. Recordings by Schultz and colleagues showed that these neurons fire not to reward itself but to reward that is better than predicted, and pause when a predicted reward fails to arrive. Dopamine thus encodes a reward prediction error: the teaching signal that tells a circuit to update its expectations. It also underlies motivation, the willingness to expend effort in pursuit of something.
Noradrenaline
A small pigmented nucleus in the pons that supplies essentially all of the cortex's noradrenaline. It is the brain's alerting system. Tonic firing tracks arousal, from near-silence in deep sleep to high rates in stress; phasic bursts fire to salient, unexpected, or behaviourally relevant events. Aston-Jones and Cohen proposed that the system implements adaptive gain: moderate tonic firing with strong phasic bursts favours focused exploitation of the current task, while high tonic firing with weak phasic responses favours distractible exploration. Noradrenaline is also central to the stress response and to the vividness with which emotional events are remembered.
Serotonin
A midline column of nuclei running the length of the brainstem, projecting almost everywhere. Serotonin's functions are the hardest of the four to summarise, partly because there are at least fourteen receptor subtypes doing different things. It participates in the regulation of mood, sleep and wakefulness, appetite, and pain. A recurring theme in modern work is patience and the tolerance of delay: serotonergic activity is associated with the willingness to wait for a better outcome rather than take an immediate lesser one. Serotonin neurons fire fastest during quiet waking, slow during non-REM sleep, and fall almost silent in REM.
Acetylcholine
Two sources with two roles. Basal forebrain cholinergic neurons project to the cortex and hippocampus, where acetylcholine promotes cortical activation, sharpens attention by improving the signal-to-noise ratio of sensory responses, and gates the encoding of new memories. Their loss is a prominent feature of Alzheimer's disease. Brainstem cholinergic nuclei drive the arousal of both waking and REM sleep, which is why the cortex during REM shows an activated, waking-like pattern even though the body is paralysed and serotonergic and noradrenergic firing has ceased.
Notice what these four have in common beyond anatomy. None of them is carrying content. Dopamine is not the signal that a chocolate biscuit exists; it is the signal that the biscuit was better than you expected. Noradrenaline is not the signal that a door slammed; it is the signal that something unexpected and important just happened, so attend. These are meta-signals, statements about the informational situation rather than about the world. That is precisely why they can be broadcast indiscriminately: a message about how to weight incoming information is useful to every circuit at once.
Histamine, orexin, and wakefulness
Two further systems deserve mention because of their outsized role in the wake-sleep switch. Histamine, familiar from allergy, is also a neuromodulator: it is released by neurons of the tuberomammillary nucleus in the posterior hypothalamus, which project widely and fire during waking, slowing in non-REM sleep and stopping in REM. Their contribution to wakefulness is easy to demonstrate in everyday life: older antihistamines that cross the blood-brain barrier are notoriously sedating, precisely because they block the same histamine receptors that help keep the cortex awake.
Orexin, also called hypocretin, is a pair of neuropeptides made by a small population of neurons in the lateral hypothalamus. These neurons excite and stabilise all the other arousal systems: the locus coeruleus, the raphe, the histaminergic and cholinergic nuclei. Orexin does not so much produce wakefulness as hold it steady, preventing rapid, uncontrolled flipping between states. The clinical evidence is unusually clean: narcolepsy with cataplexy is associated with the loss of orexin neurons, and the resulting condition is precisely a failure of state stability, with sleep intruding into waking and features of REM sleep, such as muscle paralysis, appearing during it.
How modulators set states
Put the systems together and a picture emerges of the brain as an instrument with a small number of global controls. The wiring, the pattern of excitatory and inhibitory synapses that carries the actual information, stays essentially fixed from one minute to the next. What changes, continuously, is the setting of these controls, and that is enough to make the same wiring behave in categorically different ways.
Alert and focused
Moderate tonic noradrenaline with strong phasic bursts, high acetylcholine. Sensory cortex responds sharply to relevant input and is relatively deaf to the rest; the signal-to-noise ratio is high; the current task is exploited rather than abandoned.
Restless and exploratory
High tonic noradrenaline with blunted phasic responses. The system disengages from the current task and becomes responsive to alternatives. This looks like distractibility, and in a stable environment it is; in a changing one it is the correct policy.
Motivated to pursue
Dopamine release in the striatum and prefrontal cortex, driven by cues that predict reward. Effort feels worth expending; behaviour is energised towards a goal. Deplete this and the animal still enjoys the reward, but will not work for it.
Drowsy and descending
Falling noradrenaline, serotonin, histamine, and orexin. Cortical neurons drift towards synchronised, slow, rhythmic firing, visible as the large slow waves of an EEG in deep sleep. Information transfer through the cortex collapses.
REM sleep
A peculiar and revealing state. Acetylcholine is high, and the cortex looks activated, close to waking. But noradrenaline and serotonin have fallen essentially to zero. The result is a brain running an activated cortex without its usual alerting and mood-setting chemistry, which is one plausible reason dreams are vivid, emotionally strange, and poorly remembered.
REM is the most instructive case, because it isolates the variables. Cortical activation and noradrenergic arousal, which normally rise and fall together, come apart. What that dissociation shows is that the neuromodulators are not a single dial labelled arousal. They are several independent dials, and consciousness looks very different depending on the combination, not just the sum.
Two myths worth killing
Neuromodulators suffer more than any other part of neuroscience from a persistent popular shorthand: one molecule, one feeling. The shorthand is memorable, it sells supplements, and it is wrong in a way that actively obstructs understanding.
Claim: dopamine is the pleasure chemical.
Truth: dopamine is far more about wanting than about liking. Recordings from midbrain dopamine neurons show they do not fire in proportion to how good a reward is; they fire in proportion to how much better it is than expected, and they pause below baseline when an expected reward does not arrive. That is a prediction error, a teaching signal, not a hedonic one. Behavioural work reinforces the point: animals with dopamine depleted will not work for food, but still show every sign of enjoying food placed in their mouths. Wanting collapses; liking survives. And this is before we consider that dopamine's other great job, in the nigrostriatal pathway, has nothing to do with reward at all: it is the control of movement, and its loss causes Parkinson's disease.
Claim: serotonin is the happiness chemical, and depression is a serotonin deficiency.
Truth: no one has demonstrated a simple deficit of serotonin as the cause of depression, and the chemical-imbalance framing is not what the evidence supports. Serotonin has at least fourteen receptor subtypes distributed across the brain and body, and the great majority of the body's serotonin is not in the brain at all but in the gut, where it regulates digestion. Within the brain, serotonergic function has been linked to sleep-wake regulation, appetite, pain, and, in more recent work, to patience and the willingness to tolerate delay for a better outcome. That SSRIs act on serotonin and can help some people is not evidence that low serotonin caused the problem, any more than aspirin relieving a headache shows the headache was an aspirin deficiency.
The deeper error behind both myths is a category mistake. A transmitter is not a feeling. It is a molecule, released at particular places, read by particular receptors, and its meaning is made by the circuit it lands in, not by its own chemistry. The same dopamine molecule in the striatum smooths a movement and in the ventral striatum energises a pursuit. Chemistry supplies the alphabet; the circuits write the words.
Sources
- Schultz W. Predictive reward signal of dopamine neurons. Journal of Neurophysiology. 1998;80(1):1-27.
- Aston-Jones G, Cohen JD. An integrative theory of locus coeruleus-norepinephrine function: adaptive gain and optimal performance. Annual Review of Neuroscience. 2005;28:403-450.
- Kandel ER, Koester JD, Mack SH, Siegelbaum SA. Principles of Neural Science. 6th ed. McGraw-Hill; 2021.
This page is an educational reference. It is not medical advice, does not concern any medication, and does not diagnose or treat any condition.