Stochastic resonance tuned by correlations in neural models

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Stochastic resonance (SR) is a phenomenon where a small amount of noise enhances signal detection in nonlinear systems.
• The physiological relevance of SR in neuronal models remains debated due to unclear parameter ranges.
• Previous models often assumed fixed noise levels, limiting applicability.

Purpose of the Study:

• To investigate if neurons can actively utilize stochastic resonance (SR) by modulating synaptic input correlations.
• To determine if SR in neuronal models can occur within physiologically plausible parameter ranges.
• To explore the advantages of a dynamic correlation-based SR mechanism compared to conventional approaches.

Main Methods:

• Development of a neuronal model incorporating adjustable synaptic input correlations.
• Analysis of model behavior under varying correlation levels and noise conditions.
• Comparison of the proposed mechanism with established SR theories.

Main Results:

• Demonstration that neurons can tune synaptic input correlations to exhibit SR.
• Identification of physiologically plausible parameter ranges for this SR mechanism.
• Quantification of benefits, such as improved signal-to-noise ratio, through dynamic correlation adjustment.

Conclusions:

• Neuronal SR can be achieved through the dynamic adjustment of synaptic input correlations.
• This mechanism offers a more biologically plausible pathway for SR in neurons.
• The ability to modulate correlations provides a novel mechanism for enhancing neural signal processing.