Introducing double bouquet cells into a modular cortical associative memory model.

Area of Science:

• Computational neuroscience
• Electrophysiology
• Cortical microcircuitry

Background:

• Established spiking attractor models for memory utilize Hebbian-Bayesian learning rules.
• Previous models faced challenges with simultaneous learning of excitation and inhibition, potentially violating Dale's principle.
• Double bouquet cells are key inhibitory interneurons in the cortex.

Purpose of the Study:

• To develop and integrate a biophysically plausible electrophysiological model of double bouquet cells into a cortical microcircuit.
• To address concerns regarding Dale's principle violation in models learning both excitation and inhibition.
• To investigate the impact of incorporating double bouquet cells on network learning and effective connectivity.

Main Methods:

• Developed an electrophysiological model of double bouquet cells.
• Integrated the double bouquet cell model into an existing spiking attractor microcircuit.
• Applied a Hebbian-Bayesian learning rule to plastic synapses onto double bouquet cells.
• Analyzed changes in network activity, learning abilities, and effective connectivity.

Main Results:

• The inclusion of double bouquet cells resolved issues with simultaneous excitation and inhibition learning, upholding Dale's principle.
• Network learning abilities and effective connectivity were preserved with plastic synapses onto double bouquet cells.
• The modified microcircuit architecture demonstrated improved biological plausibility without altering core network functions.

Conclusions:

• Biophysically realistic double bouquet cell models can be integrated into cortical microcircuits to enhance biological plausibility.
• This approach effectively solves the duplexed learning problem for excitation and inhibition in neural network models.
• The revised microcircuit architecture maintains computational function while improving adherence to biological constraints.