Neural connectomics challenge /
Saved in:
| Imprint: | Cham : Springer, 2017. |
|---|---|
| Description: | 1 online resource |
| Language: | English |
| Series: | The Springer series on challenges in machine learning, 2520-131X Springer series on challenges in machine learning, |
| Subject: | |
| Format: | E-Resource Book |
| URL for this record: | http://pi.lib.uchicago.edu/1001/cat/bib/11307081 |
Table of Contents:
- Foreword; Preface; Contents; First Connectomics Challenge: From Imaging to Connectivity; 1 Introduction; 2 Challenge Design; 3 Results; 3.1 Challenge Duration; 3.2 Overfitting; 3.3 PR Curves; 3.4 Edge Orientation; 3.5 Subnetworks; 4 Methods; 5 Conclusions; References; Simple Connectome Inference from Partial Correlation Statistics in Calcium Imaging; 1 Introduction; 2 Signal Processing; 3 Connectome Inference from Partial Correlation Statistics; 4 Experiments; 5 Conclusions; A.1 Signal Processing; A.2 Weighted Average of Partial Correlation Statistics; A.3 Prediction of Edge Orientation.
- A.4 ExperimentsReferences; Supervised Neural Network Structure Recovery; 1 Introduction; 2 Model; 2.1 Spike Inference; 2.2 Connectivity Indicators; 2.3 Network Deconvolution; 2.4 Modeling Approach; 3 Evaluation; 4 Conclusions and Future Work; References; Signal Correlation Prediction Using Convolutional Neural Networks; 1 Introduction; 2 Dataset and Evaluation; 3 CNN Model; 3.1 The First Solution: Basic Approach; 3.2 Background on Convolutional Neural Networks; 3.3 Introduction to CNN Filter and CNN Model; 3.4 CNN Filter Key Time Series Processing Methods; 3.5 CNN Filter Structure.
- 3.6 On the CNN Model Implementation and Development4 Results; 5 Conclusions; References; Reconstruction of Excitatory Neuronal Connectivity via Metric Score Pooling and Regularization; 1 Introduction; 2 Methods; 2.1 Preprocessing of Calcium Imaging; 2.2 Csiszár's Transfer Entropy; 2.3 Correlation Metrics; 2.4 Pooling of Different Metric Scores; 2.5 Regularization on the Recovered Network; 2.6 Evaluation of the Reconstruction Performance; 3 Results; 3.1 CTE; 3.2 Pooling Metrics Scores; 3.3 Network Regularization; 3.4 Challenge Results; 4 Discussion; References.
- Neural Connectivity Reconstruction from Calcium Imaging Signal Using Random Forest with Topological Features1 Introduction; 2 Methods; 2.1 Efficient Features Extraction; 2.2 Random Forest; 2.3 Random Forest with Topological Features; 2.4 Random Forest Training with Constant Representation Changes; 3 Evaluation; 4 Conclusions; References; Efficient Combination of Pairwise Feature Networks; 1 Introduction; 2 Typical Methods; 2.1 Correlation with Discretization; 2.2 Generalized Transfer Entropy; 3 Our Proposal: Unsupervised Ensemble of CLRed Pairwise Features; 3.1 Feature 1: Symmetrized GTE.
- 3.2 Feature 2: Correlation of the Extrema of the Signals3.3 Feature 3: Mean Squared Error of Difference Signal; 3.4 Feature 4: Range of Difference Signal; 4 Experiments; 5 Conclusion; References; Predicting Spiking Activities in DLS Neurons with Linear-Nonlinear-Poisson Model; 1 Introduction; 1.1 Dorsolateral Striatum Single Body Part Neurons; 1.2 The Linear-Nonlinear-Poisson Model; 2 Methods; 2.1 Data Collection and Preprocessing; 2.2 Experimental Design; 3 Results; 3.1 Predicting Neural Activity with Features from All Modalities.