An algorithm to predict the connectome of neural microcircuits.

Area of Science:

• Neuroscience
• Computational Biology
• Systems Neuroscience

Background:

• Mapping synaptic connections (connectomics) is crucial for understanding neural circuits but experimentally intractable for even small tissue volumes.
• The neocortex has thousands of unique synaptic connection types, with only a few characterized experimentally.
• Current experimental methods face limitations in comprehensively detailing neural microcircuits.

Purpose of the Study:

• To develop a theoretical framework and algorithmic strategy for digitally reconstructing complete synaptic connectivity within a neural microcircuit.
• To predict the anatomy of all biologically viable synaptic connections using a data-driven approach.
• To create a micro-scale connectome for a defined neural tissue volume.

Main Methods:

• Developed a data-driven algorithmic strategy based on established principles of synaptic connectivity.
• Leveraged interdependencies between neural microcircuit properties to constrain connectivity reconstruction.
• The algorithm yields three parameters per connection type to predict synaptic anatomy.

Main Results:

• The algorithm successfully predicts the anatomy of biologically viable synaptic connections.
• The predictions align with a spectrum of experimental data not used during algorithm development.
• Generated a digital reconstruction of synaptic connectivity for a neural microcircuit.

Conclusions:

• An algorithmic approach to connectomics can significantly accelerate experimental mapping efforts.
• This method can identify essential datasets for accurate predictions and guide future experimental designs.
• The framework facilitates testing hypotheses about synaptic connectivity and neural circuit function.