Battaglia D., Guyon I., Lemaire V., Orlandi J., Ray B., Soriano J. (eds.) Neural Connectomics Challenge

Springer, 2017. — 122 p.Neuroscience is nowadays one of the most appealing research fields for interdisciplinary research. The rich dynamics and complexity of living neuronal networks, and the brain in particular, has long fascinated biologists, physicists and mathematicians alike. In the last decade, however, and thanks to the giant development in computational tools and scientific interconnectivity through Internet, neuroscience has experienced a new drive that seems unstoppable and more interdisciplinary than ever.
Machine learning is one of the most innovative modern computational tools. In the context of neuroscience, it has already procured extraordinary results in brain activity data analysis, artificial intelligence, and human–machine interfacing. Machine learning tools have the capacity to predict the behavior or response of a complex system given sufficient data and training. This capacity is precisely what motivated us to launch the Connectomics Challenge. The task in mind was to solve an interesting yet highly complex inverse problem: given the time series of neuronal spontaneous activity, which is the underlying connectivity between the neurons in the network?
The present volume illustrates the efforts of the scientific community to use machine learning concepts to tackle this problem and to develop tools for the advancement of neuroscience. The volume is specially oriented to the mathematical, physical and computer science community that carries out research in neuroscience problems. It may also be of great interest for the machine learning community since it exemplifies how to approach the same problem from different perspectives. Finally, a broader readership may find interesting the description and development of the Connectomics Challenge itself and get a glimpse of major open problems in current neuroscience.First Connectomics Challenge: From Imaging to Connectivity
Simple Connectome Inference from Partial Correlation Statistics in Calcium Imaging
Supervised Neural Network Structure Recovery
Signal Correlation Prediction Using Convolutional Neural Networks
Reconstruction of Excitatory Neuronal Connectivity via Metric Score Pooling and Regularization
Neural Connectivity Reconstruction from Calcium Imaging Signal Using Random Forest with Topological Features
Efficient Combination of Pairwise Feature Networks
Predicting Spiking Activities in DLS Neurons with Linear-Nonlinear-Poisson Model
SuperSlicing Frame Restoration for Anisotropic ssTEM and Video Data