STDP (Spike-Timing-Dependent Plasticity) is a biologically plausible unsupervised learning rule for SNNs — adjusting synaptic weights based on the relative timing of pre-synaptic and post-synaptic spikes.

What Is STDP?

• The Rule: "Neurons that fire together, wire together" (Hebb).
• If input spike (Pre) comes before output spike (Post) -> Strengthen weight (LTP). "I caused you to fire."
• If input spike (Pre) comes after output spike (Post) -> Weaken weight (LTD). "I was late/irrelevant."
• Causality: STDP inherently captures causal relationships.

Why It Matters

• Unsupervised: Allows networks to learn features from data streams locally without global error backpropagation.
• Hardware Friendly: Extremely easy to implement on local neuromorphic circuits (memristors).
• Adaptation: Enables continuous online learning and adaptation to drifting signals.

STDP is the mechanism of memory — the fundamental synaptic algorithm that allows biological brains to wire themselves based on experience.