Dendrites: Why Biological Neurons Are Deep Neural Networks

Early artificial neural networks, inspired by McCulloch and Pitts’ perceptron, treated neurons as simple summing units with a threshold, but modern neurophysiology shows that real cortical neurons violate this simplification in crucial ways.

Cortical pyramidal neurons’ dendrites are not just passive cables; they contain voltage-gated sodium, calcium, and NMDA channels that turn them into active nonlinear processors of synaptic input.

Human layer 2/3 cortical pyramidal neurons exhibit dendritic calcium spikes with properties that enable them to perform logic-gate-like computations previously thought to require multi-layer networks.

A detailed biophysical model of a single cortical neuron—with realistic morphology and ion-channel dynamics—can be approximated by a multi-layer deep convolutional neural network trained to match its input–output behavior.