Synaptic depression increases the selectivity of a neuron to its preferred pattern and binarizes the neural code

Area of Science:

• Computational Neuroscience
• Neural Coding

Background:

• The Leaky Integrate-and-Fire (LIF) model exhibits poor selectivity in recognizing preferred input patterns.
• Neuron response in LIF models is dictated by total injected current, leading to similar responses for complete low-activity patterns and incomplete high-activity patterns.

Purpose of the Study:

• To propose a theoretical model that enhances neural pattern recognition selectivity.
• To investigate the role of synaptic dynamics in improving neuron response specificity.

Main Methods:

• Development of a theoretical model for depressing synapses with linear recovery.
• Simulations using biological models of strong synaptic depression.
• Analysis of neuron selectivity based on somatic decay and depression recovery time constants.

Main Results:

• The proposed model makes time-averaged synaptic current independent of input spike train frequency.
• This results in a binary neural code where neuron response strength depends solely on the number of active inputs.
• Optimal selectivity is achieved with long somatic decay time constants (>50 ms) and recovery time constants >= somatic decay time constant.

Conclusions:

• Synaptic depression can significantly improve a neuron's pattern recognition capabilities.
• By decoupling response from input firing rate, neurons can achieve more precise pattern detection.
• The findings suggest a mechanism for enhancing neural information processing through specific synaptic properties.