Spiking

Spiking generally refers to the generation of action potentials in neurons and, more broadly, to any sudden transient electrical event in an excitable cell. In neuroscience, spikes are the dominant means of neuronal signaling.

An action potential is initiated when the membrane potential reaches a threshold due to excitatory input. Voltage-gated sodium channels open, causing rapid depolarization; sodium channels inactivate, potassium channels open, repolarization occurs; the cell experiences a brief refractory period. Spikes propagate along axons, often with saltatory conduction.

Information is carried by sequences of spikes called spike trains. Coding schemes include rate coding (information carried by average firing rate) and temporal coding (precise timing). Firing patterns vary: regular spiking, irregular spiking, bursting, and adaptation.

In computing, artificial spiking neural networks use neurons that emit discrete spikes and communicate through spike timing. They leverage temporal coding and event-driven processing; training methods include spike-timing-dependent plasticity and surrogate-gradient approaches; hardware implementations aim to reduce energy consumption via neuromorphic chips.

In data analysis, "spikes" describe brief, sharp transients in signals, such as spike detection in neural or electrophysiological data. In pharmacology or physiology, the term is occasionally used to describe sudden transient discharges or action potentials observed in various cell types.