Oscillations via Spike-Timing Dependent Plasticity in a Feed-Forward Model.

Area of Science:

• Computational Neuroscience
• Neuroplasticity
• Neural Oscillations

Background:

• Neuronal oscillations are prevalent in the central nervous system and linked to cognitive functions like attention and learning.
• The origin of neuronal oscillations remains debated: are they genetically determined or acquired through learning?

Purpose of the Study:

• To investigate the conditions under which oscillatory activity emerges via spike timing-dependent plasticity (STDP) in a feed-forward neural architecture.
• To analyze the influence of STDP rule parameters and oscillatory dynamics on synaptic weight changes.

Main Methods:

• Analysis of STDP-driven synaptic dynamics for single synapses and pre-synaptic populations onto a post-synaptic neuron.
• Development of a mean-field Fokker-Planck approximation to model synaptic dynamics.
• Investigation of conditions leading to the instability of homogeneous solutions and the emergence of oscillatory activity.

Main Results:

• Oscillatory activity emerges through spontaneous symmetry breaking, where homogeneous solutions become unstable.
• Synaptic dynamics do not converge to a fixed point but rather to a limit cycle in the general case.
• The temporal structure of the STDP rule dictates the stability of homogeneous solutions and the velocity of limit cycle drift.

Conclusions:

• Learning, specifically through STDP, can be a mechanism for generating neuronal oscillations.
• The emergence of oscillations is a result of dynamic instability and symmetry breaking in synaptic plasticity.
• STDP temporal properties are critical in shaping the stability and dynamics of emergent oscillatory network activity.