Pre- and postsynaptically expressed spike-timing-dependent plasticity contribute differentially to neuronal learning.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Synaptic Plasticity

Background:

• Long-term synaptic plasticity, essential for learning and memory, exhibits diverse expression patterns (presynaptic vs. postsynaptic).
• The functional impact of these distinct expression sites remains incompletely understood, particularly in computational models.
• Current models often simplify plasticity to postsynaptic changes, potentially introducing bias.

Purpose of the Study:

• To investigate the functional consequences of diverse long-term synaptic plasticity expression mechanisms.
• To compare pre-, post-, and mixed-site expression of spike-timing-dependent plasticity (STDP) in computational models.
• To explore how different expression sites influence learning dynamics and neuronal output.

Main Methods:

• Adapted a spike-timing-dependent plasticity (STDP) model for independent pre-, post-, or mixed-site expression.
• Compared standard pair-based STDP with a biologically tuned triplet STDP model.
• Investigated learning outcomes under two schemes: differing input latencies and temporally correlated inputs.

Main Results:

• Presynaptic plasticity accelerated learning speed.
• Postsynaptic plasticity enhanced regulation of spike timing and frequency.
• Combined expression amplified response range and provided homeostatic control of firing rates.

Conclusions:

• Pre- and postsynaptic plasticity expression are not interchangeable; they serve complementary functions.
• Implementing plasticity solely as postsynaptic weight changes can introduce bias in computational models.
• Understanding the locus of plasticity is critical for accurate modeling of neuronal learning and network dynamics.