spiketimingdependent

Spike-timing-dependent plasticity, often abbreviated as STDP, is a form of synaptic plasticity in which changes in synaptic strength depend on the precise relative timing of presynaptic and postsynaptic action potentials. This temporal dependence makes STDP a temporally asymmetric Hebbian learning rule that contributes to the refinement of neural circuits during development and learning. While the exact expression of STDP can vary across brain regions and synapses, the core idea is that the order and timing of spikes shape synaptic changes.

In the canonical pair-based version, if a presynaptic spike arrives shortly before a postsynaptic spike within a window of a few milliseconds, the synapse typically undergoes long-term potentiation (LTP). If the presynaptic spike follows the postsynaptic spike within a similar window, the synapse tends to undergo long-term depression (LTD). The timing windows usually span tens of milliseconds, and the magnitude of change depends on how tightly the spikes are synchronized. The underlying biological mechanisms involve NMDA receptor activation and postsynaptic calcium transients, which trigger intracellular signaling cascades that strengthen or weaken the synapse.

Variants of STDP include triplet and more complex timing rules, which can account for additional experimental findings, as well as multiplicative versus additive forms of plasticity. Neuromodulators such as dopamine can gate or modulate STDP, making the plasticity dependent on behavioral relevance or reinforcement signals. In vivo, ongoing activity and network dynamics can influence the effective learning window, leading to deviations from idealized rules in cultured tissue or intact circuits.

STDP has been demonstrated in multiple brain regions, including the hippocampus and neocortex, and remains a foundational concept in models of associative learning, memory formation, and neural development. It has also inspired computational approaches and neuromorphic hardware that imitate timing-dependent learning rules.