Leveraging heterogeneity for neural computation with fading memory in layer 2/3 cortical microcircuits.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neurobiological systems exhibit inherent complexity and heterogeneity across all scales.
• Cortical microcircuit studies often simplify by assuming homogeneity, potentially limiting understanding of computational principles.

Purpose of the Study:

• To investigate the functional roles of neuronal, synaptic, and structural heterogeneities in a biophysically plausible layer 2/3 cortical microcircuit model.
• To determine how these heterogeneities contribute to the computational power and emergent behaviors of neural circuits.

Main Methods:

• Developed a layer 2/3 microcircuit model incorporating neuronal, synaptic, and structural heterogeneities.
• Constrained the model using empirical data from large-scale databases and advanced experimental methodologies.
• Analyzed the impact of individual and combined heterogeneities on circuit dynamics and computational properties.

Main Results:

• Variability in single neuron parameters was identified as the primary driver of functional specialization.
• Heterogeneous microcircuits demonstrated significantly higher computational power compared to homogeneous models.
• The response properties of fully heterogeneous circuits result from the differential contributions of various heterogeneity sources.

Conclusions:

• Acknowledging and modeling neural heterogeneity is crucial for accurately understanding cortical function and computation.
• Single neuron variability plays a dominant role in enabling sophisticated computational capabilities within neural circuits.
• Biophysically realistic models incorporating multiple sources of heterogeneity are essential for advancing our knowledge of brain function.