Biological vs artificial neurons

An artificial neuron keeps one idea from biology — sum weighted inputs and fire if the total is high — and drops spikes, timing and chemistry.

7 min read553 wordscore0 viewsupdated 2026-10-09

Definition

An artificial neuron is a function y=f(∑iwixi+b)y=f\big(\sum_i w_i x_i+b\big): multiply each input by a weight, add them up with a bias, and pass the total through a fixed nonlinearity ff. It is a deliberately crude model of a brain cell. It keeps the idea of integrating many inputs and responding when the total is high, and it throws away almost everything else.

What a real neuron does

A cortical neuron receives signals on its dendrites, branching fibres that collect input from thousands of other cells. Each contact point is a synapse. When the upstream cell fires, the synapse releases neurotransmitter, which nudges the receiving cell's voltage up (excitatory) or down (inhibitory). How big the nudge is depends on the synapse's strength.

These nudges spread to the soma (cell body) and add up. At rest, the inside of the cell sits near −70-70 mV. If the summed input pushes the voltage at the axon hillock past a threshold of about −55-55 mV, the cell fires an action potential: an all-or-none spike, about 1 ms long, that travels down the axon to the next cells. Below threshold, nothing is sent.

The artificial neuron is that sentence turned into arithmetic: weighted sum, then threshold.

BiologyArtificial neuronRole
dendritesinputs xix_ireceive signals from upstream units
synaptic strength (excitatory / inhibitory)weight wiw_i (positive / negative)how much, and in which direction, each input counts
summing of potentials in the somaz=∑iwixi+bz=\sum_i w_i x_i+bcombine all inputs into one number
resting level vs firing thresholdbias bb (a threshold θ\theta is b=−θb=-\theta)how much input is needed to respond
axon hillock: fire or notactivation f(z)f(z)turn the total into an output
axon and its terminalsoutput yy, fanned out to the next layersend the result on
The mapping the artificial neuron keeps.

McCulloch–Pitts: the first formal neuron (1943)

y={1if ∑iwixi≥θ0otherwisexi∈{0,1}y=\begin{cases}1 & \text{if } \sum_i w_i x_i \ge \theta\\ 0 & \text{otherwise}\end{cases}\qquad x_i\in\{0,1\}
xix_i
binary input: upstream neuron firing (1) or silent (0)
wiw_i
fixed connection weight, set by hand (no learning yet)
θ\theta
firing threshold
yy
binary output: fire or stay silent

Warren McCulloch and Walter Pitts showed that networks of such units can compute any Boolean function, so in principle a brain made of them could carry out any logical computation. In their model an active inhibitory input vetoes firing outright. The weights were fixed by the designer. Rosenblatt's perceptron (1958) added a rule for learning them from examples.

x1x_1x2x_2∑wixi\sum w_i x_iAND (θ=2\theta=2)OR (θ=1\theta=1)
00000
01101
10101
11211
One unit, two inputs, both weights 1. Only the threshold changes: θ=2\theta=2 gives AND, θ=1\theta=1 gives OR.

The modern artificial neuron

z=∑i=1nwixi+b,y=f(z)z=\sum_{i=1}^{n} w_i x_i + b,\qquad y=f(z)
xix_i
real-valued input (a pixel, or the output of another neuron)
wiw_i
learned weight; any real number, either sign
bb
learned bias; shifts the threshold
zz
pre-activation, the analogue of the summed membrane potential
ff
activation: step, sigmoid, ReLU, …

The hard step of McCulloch–Pitts has a derivative of zero everywhere it is defined, so gradient descent cannot train it. Modern networks swap in a smooth or piecewise-linear ff. The sigmoid σ(z)=1/(1+e−z)\sigma(z)=1/(1+e^{-z}) can be read as a firing rate between 0 and 1, which is closer to how neuroscientists summarise a spike train than a single 0/1 output.

biological neuronartificial neuronsomadendritesaxonw1=0.800.90w2=−1.200.20w3=0.500.60Σ+ b (−0.30)1.00z = 0.72 − 0.24 + 0.30 − 0.30 = 0.48y = step(z) = 1.000 → fireshover a part on either side to see its counterpartrate code: y → up to 100 Hz · net passes 1 number, neuron sends 30 spikesreal axon:100 Hz0 ms300 ms
fig.Hover any part on either side to see its counterpart. Drag the inputs xᵢ and weights wᵢ (green = excitatory, red = inhibitory) and watch the soma and Σ respond. It opens on the worked example below: z = 0.48, so the step unit fires. Raise the inhibitory input x₂ past 0.6 and it goes silent. Switch to sigmoid and the output becomes a graded 0.618, read on the bottom strip as a 62 Hz spike train. The network passes one real number; a real neuron would send 18 discrete spikes in that 300 ms window.

Worked example: three inputs

  1. 1Inputs x=(0.9, 0.2, 0.6)x=(0.9,\,0.2,\,0.6), weights w=(0.8, −1.2, 0.5)w=(0.8,\,-1.2,\,0.5), bias b=−0.3b=-0.3. Input 2 is inhibitory.
  2. 2Weighted terms: 0.8⋅0.9=0.720.8\cdot0.9=0.72,   −1.2⋅0.2=−0.24\;-1.2\cdot0.2=-0.24,   0.5⋅0.6=0.30\;0.5\cdot0.6=0.30.
  3. 3Sum: z=0.72−0.24+0.30−0.30=0.48z=0.72-0.24+0.30-0.30=0.48.
  4. 4Step: z≥0z\ge 0, so y=1y=1 (fires). ReLU: y=max⁡(0, 0.48)=0.48y=\max(0,\,0.48)=0.48.
  5. 5Sigmoid: y=1/(1+e−0.48)=0.618y=1/(1+e^{-0.48})=0.618. Read as a rate with a 100 Hz ceiling, that is about 62 Hz.
  6. 6To silence the step unit, strengthen the inhibitory input: z=0.72−1.2 x2z=0.72-1.2\,x_2, which is negative once x2>0.6x_2>0.6. At x2=0.65x_2=0.65, z=−0.06z=-0.06 and y=0y=0. (Lowering w2w_2 alone would need w2<−3.6w_2<-3.6.)
python
1import torch
2
3x = torch.tensor([0.9, 0.2, 0.6])
4w = torch.tensor([0.8, -1.2, 0.5])
5b = torch.tensor(-0.3)
6
7z = w @ x + b # tensor(0.4800)
8print((z >= 0).float()) # step: 1.0
9print(torch.sigmoid(z)) # sigmoid: 0.6177
10print(torch.relu(z)) # ReLU: 0.4800
11
12# A layer is many such neurons sharing the same inputs:
13layer = torch.nn.Linear(3, 128) # 128 neurons, each with 3 weights + 1 bias

Where the analogy breaks

AspectBiological neuronArtificial neuron
Signaldiscrete spikes; information in rate and precise timingone real number per forward pass
Timecontinuous dynamics; potentials leak and decay over millisecondsno time at all (outside recurrent models); one synchronous step
Connectionsroughly 10410^4 synapses per cortical neuron; ∼ ⁣86\sim\!86 billion neurons in a human brainas many as the layer has inputs; a large model has billions of weights in total
Dendritescompute nonlinearly on their own; single human cortical neurons can compute XOR in their dendrites (Gidon et al., 2020)a plain linear sum
Sign of outputa given neuron is excitatory or inhibitory, not both (Dale's principle)any weight can take either sign and flip during training
Learninglocal rules: synapses change from the activity of the two cells they join (Hebbian learning, spike-timing plasticity)backpropagation: a global error signal sent backwards through exact copies of the forward weights
Energywhole brain runs on about 20 Wtraining a large model takes megawatt-hours

Check yourself

In the artificial neuron, which quantity plays the role of synaptic strength?

Check yourself

A McCulloch–Pitts unit has two binary inputs with weights 1 and 1. Which threshold makes it compute OR?

Remember

  • ✓Dendrites → inputs, synapses → weights, soma → weighted sum, axon hillock → activation, axon → output.
  • ✓McCulloch–Pitts (1943): binary inputs, fixed weights, hard threshold; enough for any Boolean function.
  • ✓Modern neurons use real values and smooth activations so they can be trained by gradient descent.
  • ✓Real neurons spike, integrate over time, have ~10⁴ synapses with nonlinear dendrites, and learn with local rules.
  • ✓Treat the brain as the source of the metaphor, not as a specification.
← back to Deep Learning roadmap