Tracking Fast and Slow Changes in Synaptic Weights From Simultaneously Observed Pre- and Postsynaptic Spiking

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Synaptic plasticity underlies learning and memory, involving changes on various timescales.
• Short-term plasticity (milliseconds to seconds) and long-term plasticity (minutes to hours) dynamically alter neural communication.
• Existing models often struggle to simultaneously capture these diverse plasticity mechanisms.

Purpose of the Study:

• To develop a generalized linear model extension for inferring both short- and long-term synaptic plasticity from spiking activity.
• To provide a unified framework for analyzing dynamic changes in neural coupling.
• To investigate the necessity of tracking multiple plasticity components concurrently.

Main Methods:

• Extended a generalized linear model to incorporate additive effects for short-term plasticity based on presynaptic spike timing.
• Utilized point process adaptive smoothing to model long-term changes in synaptic weight and baseline firing rates.
• Validated the model using extensive simulations with varying synapse types, plasticity rules, firing rates, and synapse polarities (excitatory/inhibitory).

Main Results:

• The model accurately recovered time-varying synaptic weights for both depressing and facilitating synapses.
• It successfully characterized diverse long-term changes, including those induced by spike-timing-dependent plasticity (STDP).
• Simulations demonstrated robustness across different firing rates and synapse types.
• Application to experimental data highlighted the critical importance of simultaneously tracking fast and slow synaptic weight changes alongside baseline firing rate variations.

Conclusions:

• Simultaneously inferring short-term plasticity, long-term synaptic weight changes, and baseline firing rate variations is essential for accurate synaptic analysis.
• The developed model offers a flexible and powerful framework for dissecting complex, multi-timescale synaptic dynamics in neural circuits.
• Failure to account for all these factors can lead to misleading conclusions about synaptic function.