Computational functions of precisely balanced neuronal microcircuits in an olfactory memory network.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Autoassociative memory network models emphasize the role of specific inhibition in information storage.
• Fast-spiking interneurons are key components of inhibitory microcircuits in cortical structures.

Purpose of the Study:

• To investigate the role of different subtypes of fast-spiking interneurons in autoassociative memory.
• To test predictions of balanced network models regarding excitation-inhibition balance and pattern correlations.
• To explore the computational functions of inhibitory microcircuits in the posterior telencephalic area Dp (pDp) of zebrafish.

Main Methods:

• Characterization and optogenetic manipulation of distinct fast-spiking interneuron subtypes in adult zebrafish pDp.
• Computational modeling of recurrent networks with neuronal assemblies to simulate network dynamics.
• Analysis of neuronal firing rates and pattern correlations under different inhibition conditions.

Main Results:

• Recurrent network models demonstrated that balanced excitation and inhibition prevent runaway activity and high pattern correlations.
• Perturbation of feedback inhibition, but not feedforward inhibition, in pDp led to the emergence of runaway correlations.
• Runaway correlations were driven by sparse, highly active neuronal subsets, not by broadened tuning curves.

Conclusions:

• Experimental results align with balanced neuronal assembly models in the zebrafish pDp.
• Specific subtypes of inhibitory microcircuits play critical computational roles in autoassociative networks.
• Feedback inhibition is essential for preventing spurious pattern correlations and maintaining network stability.