Inference of synaptic connectivity and external variability in neural microcircuits

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Estimating neural connectivity from large-scale in vivo recordings is crucial but challenged by unmeasured external synaptic input (common input problem).
• Existing functional connectivity measures often assume or ignore their direct relationship to synaptic connectivity, lacking ground truth validation for in vivo data.
• In silico simulations, while useful, rely on simplifying assumptions and numerous parameters.

Purpose of the Study:

• To investigate the conditions under which functional connectivity, synaptic connectivity, and external input variability can be distinguished.
• To clarify the relationship between measured neural activity and underlying synaptic connections.
• To develop a more robust method for inferring neural connectivity from experimental data.

Main Methods:

• Combined neuronal network simulations with mathematical analysis.
• Utilized calcium imaging data for validation.
• Investigated the precision matrix of recorded spiking activity and spatial structure of neuronal variability.

Main Results:

• The precision matrix of spiking activity, while not uniquely determining synaptic connectivity, is practically closely related to it.
• This relationship is strengthened when the spatial structure of neuronal variability is jointly considered.
• Demonstrated the interplay between functional connectivity, synaptic connectivity, and external input.

Conclusions:

• Functional connectivity can serve as a reliable proxy for synaptic connectivity under specific conditions.
• Incorporating spatial information of neuronal variability enhances the accuracy of connectivity estimations.
• This work provides a framework for better understanding neural network organization from large-scale recordings.