Spike

A discrete event carrying information in spiking neural networks — the fundamental unit of flow in neuromorphic systems.

Unlike continuous activations in conventional neural networks, spikes are binary events occurring at specific times. Information is encoded in spike timing, frequency, and patterns across populations of neurons.

Systems Connection

Spikes are the flows in neural systems. They are discrete, asynchronous, and sparse — most neurons are silent most of the time. This sparsity enables event-driven processing: computation only happens when spikes arrive, dramatically reducing energy consumption compared to continuous processing.

Encoding Schemes

• Rate coding — information in spike frequency
• Temporal coding — information in precise spike timing
• Population coding — information in patterns across neurons

See Also

• Neuromorphics — parent domain
• Neuron — what generates spikes
• Temporal Coding — exploiting spike timing
• Event-Driven Processing — paradigm spikes enable