A blog of language, neuroscience, and deep learning

An exploration of some biological neuron models, in Javascript

- Different types of biological neuron models are presented below, with controls for the various parameters they use
- The models are arranged (roughly) in order from least detailed to most detailed
- The vertical axis is the voltage of the neuron in millivolts, and the horizontal axis is the current time in milliseconds
- For each model, the initial input voltage is zero, and stepped to the specified voltage after 20 milliseconds

- This is the simplest neuron model out there
- Considers the neuron’s membrane to be an RC circuit, hence the characteristic RC curve
- Each drop represents a spike; because the circuit is so stereotyped, spikes can be predicted exactly given knowledge of the neuron’s input voltages

- This model is able to capture a wide variety of firing dynamics with an exceptionally small number of parameters
- Check out this website for some parameters that do cool stuff

- This model is based on detailed patch clamp experiments, and models spiking at an ionic level
- This model was created in 1952

- The plot above compares two of the Hogdkin Huxley model’s gating variables at various times
- The first choice is the variable that is displayed on the X axis, and the second choice is the variable this is displayed on the Y axis