From Neurotransmitter Receptor Occupancy to Circuit Response

From Binding to Circuit Output

A measured or modeled dopamine or serotonin concentration does not by itself specify a neural response. The fraction of receptors occupied depends on affinity and competing ligands; the signal produced by occupied receptors depends on efficacy and cellular coupling; and circuit output can depend nonlinearly on signals from several receptor subtypes. Keeping these stages distinct helps prevent a common modeling error: treating receptor occupancy as if it were already a measure of behavioral or circuit effect.

The framework here is a compact, illustrative mapping. It is useful for connecting transmitter exposure to a hypothesized receptor-mediated response, but its parameters must be interpreted in the context of the receptor, cell type, and circuit being modeled. Dopamine D1-like and D2-like receptors, and the many serotonin receptor subtypes, do not have universal circuit effects independent of their anatomical and cellular context.

Occupancy and Receptor-Specific Signal

These expressions assume a single class of independent binding sites, equilibrium, and known free concentrations at the receptor. They do not use total tissue concentration as a substitute for free concentration. When there is only one ligand, occupancy rises with concentration and approaches 1 as \(C\) becomes large relative to \(K_D\). For example, if \(C=K_D\), then \(\theta=1/2\).

Occupancy is not efficacy. Let \(\epsilon_l\) denote the efficacy of ligand \(l\), scaled so that a reference full agonist has \(\epsilon=1\), a partial agonist has \(0<\epsilon<1\), and a neutral antagonist has \(\epsilon=0\). In a simple model, efficacy-weighted occupancy is \(x=\sum_l \epsilon_l\theta_l\). This is a simplified measure of receptor stimulus, not a universal molecular quantity. An antagonist can occupy receptors while contributing no positive stimulus in this representation; by competing with an endogenous agonist, it can nevertheless reduce that agonist's occupancy and signal.

A saturating transduction function maps receptor stimulus \(x\) to a normalized receptor-specific signal \(r\):

Here, \(r\) ranges from 0 toward 1, and \(\tau\) is a dimensionless transduction-gain parameter that summarizes receptor abundance and downstream coupling in this model. A larger \(\tau\) represents greater effective amplification, often described as receptor reserve: a substantial response can occur without occupying every receptor. The model is an operational approximation, not a claim that receptor number and coupling can always be reduced to one identifiable parameter.

Why the Transduction Function Saturates

For \(r=\tau x/(1+\tau x)\), differentiation with respect to \(x\) gives

So increasing stimulus cannot decrease this model's receptor signal. The slope is strictly positive when \(\tau>0\), and is zero when \(\tau=0\). Also, \(r=0\) at \(x=0\), while \(r\) approaches 1 as \(x\) increases without bound when \(\tau>0\). The diminishing slope at larger \(x\) is the model's representation of saturation. These properties are built into the chosen function; they should not be mistaken for empirical proof that a particular receptor pathway follows this exact curve.

Receptor Reserve and Partial Agonism

Receptor reserve means that the system can produce a substantial response without requiring occupancy of all available receptors. In the model, a higher \(\tau\) increases \(r\) at a given positive stimulus \(x\). Consequently, a partial agonist may produce a considerable signal in a system with high reserve, while producing a smaller signal in a system with low reserve. The term “partial agonist” describes efficacy relative to a reference agonist in a specified assay or system; it does not guarantee one fixed response across tissues.

Combining Opposing Subtype Signals

A circuit model can combine receptor-specific signals with signed weights. A positive weight represents a contribution that increases the chosen circuit input; a negative weight represents a contribution that decreases it. These signs are model-specific. They should be justified for the relevant neuronal populations and circuit endpoint rather than assigned from the transmitter name alone.

One possible nonlinear mapping from a net input \(z\) to a bounded output \(y\) is the logistic function:

Here \(b\) is a baseline input, \(w_j\) is the signed weight for receptor subtype \(j\), and \(r_j\) is its normalized signal. The logistic output lies between 0 and 1. Unless the model explicitly defines \(y\) as a probability, it is better interpreted as a normalized circuit output, not as the probability of an event. A weighted sum is one assumption about integration; real circuits may include additional interactions, delays, and state dependence.

Assumptions, Diagnostics, and Alternatives

Before applying this mapping, check the assumptions that carry its interpretation. The occupancy calculation assumes receptor-level free concentrations and binding equilibrium. The simple receptor signal assumes that efficacy-weighted occupancy is a useful stimulus summary and that a saturating function with a single gain parameter is adequate. The circuit equation assumes that subtype signals combine through specified weights and that the chosen output transformation is appropriate for the endpoint. These are modeling assumptions to test, not consequences of measuring occupancy.

Useful diagnostics include plotting predicted receptor signal against stimulus and predicted circuit output against net input; checking that the functions behave as intended over the observed range; and comparing predicted responses with independent measures of receptor-proximal signaling or circuit activity. Residual patterns, systematic underprediction at high exposure, or different response curves across cell types may indicate missing mechanisms, a wrong transduction shape, or an inappropriate shared \(\tau\). If the model estimates occupancy and efficacy only through their product, those quantities may be difficult to distinguish from response data alone.

The main alternatives answer different questions. An occupancy-only model stops at binding and is appropriate only when occupancy itself is the endpoint of interest. A linear occupancy-response model is simpler and may be adequate over a narrow range, but it does not represent saturation and can predict unbounded signal if extrapolated. The operational saturating model represents reserve compactly but has limited mechanistic detail. More detailed receptor-state or signaling-cascade models can represent multiple steps, but require more parameters and data. A circuit model with signed subtype weights adds integration assumptions; it should be compared with a model omitting selected pathways or using a different justified integration rule.

A frequent mistake is to read “60% occupancy” as “60% receptor response,” then carry that number directly into a circuit endpoint. The correction is to calculate occupancy, apply efficacy and transduction, and then combine subtype signals using an explicit circuit model. A second pitfall is to label a subtype intrinsically excitatory or inhibitory. The sign of its effect depends on receptor location, cell identity, and the measured circuit output.