Spontaneous dynamics of synaptic weights in stochastic models with pair-based spike-timing-dependent plasticity

Area of Science:

• Computational Neuroscience
• Theoretical Neuroscience
• Systems Neuroscience

Background:

• Spike-timing dependent plasticity (STDP) is a fundamental mechanism for synaptic plasticity in neuronal networks.
• Understanding STDP dynamics is crucial for comprehending learning and memory.
• Existing models may not fully capture the complex interplay of timescales in synaptic updates.

Purpose of the Study:

• To theoretically analyze several STDP rules using Markovian theory.
• To investigate the long-time limits of synaptic weight dynamics under STDP.
• To explore the influence of different STDP rules and synapse types (excitatory/inhibitory) on synaptic stability.

Main Methods:

• Development of a theoretical framework based on Markovian theory.
• Exploitation of timescale separation between neuronal activity and synaptic weight updates.
• Derivation of asymptotic regimes for synaptic weight dynamics under various STDP rules.

Main Results:

• Presynaptic and postsynaptic spike pairing dictates synaptic weight dynamics in excitatory synapses, especially under low input.
• Mean-field analyses might overlook critical STDP properties.
• Anti-Hebbian STDP promotes stable synaptic weights, while Hebbian STDP is required for non-null weight convergence in inhibitory synapses.

Conclusions:

• The specific pairing schemes of spikes significantly influence synaptic plasticity, particularly in excitatory connections.
• Different STDP rules exhibit distinct behaviors on excitatory versus inhibitory synapses.
• This work provides insights into adaptive neuronal systems and general adaptive systems, suggesting avenues for future research.