Self-organization of memory activity through spike-timing-dependent plasticity

Area of Science:

• Computational neuroscience
• Neural network dynamics
• Synaptic plasticity

Background:

• Working memory relies on sustained neural activity.
• Spike-timing-dependent plasticity (STDP) is a key mechanism for synaptic modification.
• Previous studies often focused on single neurons, limiting understanding of network-level phenomena.

Purpose of the Study:

• To investigate how STDP self-organizes neural network activity for memory.
• To explore the impact of different STDP time windows on network dynamics.
• To elucidate the relationship between asynchronous and synchronous neural activity patterns.

Main Methods:

• Computational modeling of neural networks.
• Simulations of STDP with varying time window parameters (approx. 10 ms and 20 ms).
• Analysis of emergent network activity patterns, including firing rates and synchrony.

Main Results:

• Short STDP time windows (approx. 10 ms) self-organized networks to exhibit low-rate (20-30 Hz) asynchronous persistent activity, characteristic of working memory delay periods.
• Longer STDP time windows (approx. 20 ms) led to the self-organization of layered cell assemblies supporting synchronized spike propagation (synfire chains).
• Network-level effects on activity regulation were observed that are not apparent from single-neuron analysis.

Conclusions:

• STDP is a unifying principle for self-organizing distinct neural network architectures for sustained activity.
• Asynchronous activity and synfire chains represent different network realizations of the same synaptic wiring principle.
• These findings offer insights into the neural basis of working memory and precisely timed neural processing.