FARCI: Fast and Robust Connectome Inference.

Area of Science:

• Computational Neuroscience
• Neuroimaging Analysis
• Network Science

Background:

• Neuronal connectome inference from large-scale activity recordings is crucial in computational neuroscience.
• Two-photon Calcium imaging generates high-dimensional data requiring advanced analytical tools.
• Existing methods face challenges in accuracy, scalability, and computational efficiency.

Purpose of the Study:

• To develop FARCI (Fast and Robust Connectome Inference), a MATLAB package for accurate neuronal connectome inference.
• To evaluate FARCI's performance on in silico datasets from the Neural Connectomics Challenge and NAOMi simulator.
• To compare FARCI against established methods like Generalized Transfer Entropy (GTE) and FluoroSNNAP.

Main Methods:

• Utilized partial correlations to quantify functional association strength between neurons.
• Reconstructed neuronal connectomes using the developed FARCI package.
• Validated performance using diverse in silico datasets, varying network sizes, noise levels, and missing neuron data.

Main Results:

• FARCI demonstrated high accuracy in reconstructing neuronal connectomes across various network sizes.
• The method proved robust to missing neurons and noise, maintaining performance.
• FARCI significantly outperformed GTE and FluoroSNNAP in accuracy, computational speed, and scalability.

Conclusions:

• FARCI offers a computationally efficient and highly scalable solution for neuronal connectome inference.
• The package provides accurate and robust connectome reconstruction from two-photon Calcium imaging data.
• FARCI represents a significant advancement for large-scale neural network analysis in computational neuroscience.