How does a neuron generate an electrical signal?

For: A reader who knows that neurons communicate but wants a physical explanation.

A nerve signal begins with a physical arrangement: a cell membrane separates solutions containing charged particles, and the membrane does not let every kind of particle cross equally freely. A difference in concentration encourages movement in one direction; an electrical difference can encourage or oppose it. The useful question is therefore not where a neuron stores an idea, but how its membrane changes the movement of charge. MIT’s opening neural-computation lecture starts with those two influences. The Noba neurons module supplies the cell vocabulary.

An electrical circuit is a model of this behaviour. Resistance represents difficulty in passing current; capacitance represents the membrane’s ability to separate charge. This helps explain why the voltage need not jump immediately to a new value when current arrives. MIT’s RC-circuit lecture connects the circuit to injected-current experiments. Its components should be read as representations of measurable properties, rather than as miniature manufactured parts inside the cell.

Different questions need different amounts of biological detail. A leaky integrate-and-fire model follows a voltage-like variable, allows it to leak towards a resting value and records a spike when a specified threshold is crossed. It then resets that variable. This can make the timing of spikes easier to study, but the reset is a modelling rule: it does not explain the full shape of a biological action potential. The Neuromatch tutorial makes this reduced model available for exploration.

The next level asks how membrane currents change as voltage and time change. The Hodgkin–Huxley introduction brings in voltage clamp, an experimental method for studying current while controlling membrane voltage. This is a useful connection between an equation and an observation: model components become more convincing when an experiment can distinguish their effects.

Keep three things separate as you learn: a recorded voltage trace, a model that reproduces part of that trace, and an explanation of the underlying biological machinery. Agreement on one output does not show that every mechanism in the model is correct. A simple model may be excellent for one timing question and unsuitable for another question about ion channels. Conversely, adding more detail does not automatically make a model easier to test.

Your first goal is to explain why input, membrane properties and time jointly matter. Leave the detailed equations for the deeper route if necessary. Being able to name what a model leaves out is already an important part of understanding what it explains.

Next step

Continue to synapses and ask how several inputs combine before a neuron produces an output.

Limits of this route

MIT neural computation is 2018 teaching. Publisher descriptions, the stated Neuromatch sections and an opening portion of the Noba module were inspected; no complete viewing or notebook execution is claimed. The mathematical continuations need more preparation than the opening route.