Supercomputer framework for reverse engineering firing patterns of neuron populations to identify their synaptic...

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Neural Engineering

Background:

• Understanding neural circuit organization is crucial for deciphering brain function.
• Spinal motoneurons integrate diverse synaptic inputs to control motor commands.
• Reverse engineering (RE) approaches aim to infer neural circuit structure from activity patterns.

Purpose of the Study:

• To develop and validate novel RE techniques for identifying synaptic input organization in neuronal populations.
• To assess the efficacy of these RE methods in reconstructing complex excitatory, inhibitory, and neuromodulatory input patterns.
• To evaluate the impact of neural system non-uniqueness and neuromodulation on RE accuracy.

Main Methods:

• In silico simulations of spinal motoneurons subjected to extensive parameter searches using supercomputing resources.
• Development of RE algorithms to analyze simulated neuronal firing patterns and infer input characteristics.
• Quantitative analysis of RE performance, measuring variance accounted for in reconstructed input patterns.

Main Results:

• Simulated neuronal firing patterns contained sufficient information to significantly restrict possible input combinations.
• RE techniques successfully estimated simulated patterns of excitation, inhibition, and neuromodulation with 75-90% variance accounted for.
• Neuromodulation-induced nonlinearities in firing patterns aided, rather than hindered, the RE process.

Conclusions:

• Novel RE techniques can effectively identify synaptic input organization from neuronal firing patterns.
• Computational power advancements will likely enhance the accuracy and applicability of RE in neuroscience.
• This approach holds promise for understanding neural computation across various neural systems.