Transient responses to rapid changes in mean and variance in spiking models

Area of Science:

• Computational neuroscience
• Neuronal dynamics
• Information theory in neural systems

Background:

• Neurons receive synaptic input with independent mean and variance, suggesting dual information channels.
• Rapidly varying signals challenge neuronal responses, necessitating an understanding of temporal dynamics.

Purpose of the Study:

• To investigate how neuronal firing rate responses are affected by signals varying in mean or variance.
• To analyze the impact of different neuronal models on response onset and steady-state dynamics.

Main Methods:

• Simulating model neurons subjected to step changes in input current's mean or variance.
• Comparing response dynamics across different neuronal models, including leaky-integrate-and-fire.

Main Results:

• Response onset dynamics are model-dependent; leaky-integrate-and-fire models exhibit an instantaneous component due to a hard threshold.
• A decaying oscillatory approach to steady-state firing rate is common across models.
• The decay time constant follows a power-law relationship with noise magnitude.

Conclusions:

• Neuronal model specifics, like thresholds, critically influence response onset.
• Understanding these dynamics is key to determining neuronal population response ranges and faithful signal encoding capabilities.