- title: 'First Connectomics Challenge: From Imaging to Connectivity'
abstract: 'We organized a Challenge to unravel the connectivity of simulated neuronal networks. The provided data was solely based on fluorescence time series of spontaneous activity in a net- work constituted by 1000 neurons. The task of the participants was to compute the effective connectivity between neurons, with the goal to reconstruct as accurately as possible the ground truth topology of the network. The procured dataset is similar to the one measured in in vivo and in vitro recordings of calcium fluorescence imaging, and therefore the algorithms developed by the participants may largely contribute in the future to unravel major topological features of living neuronal networks from just the analysis of recorded data, and without the need of slow, painstaking experimental connectivity labeling methods. Among 143 entrants, 16 teams participated in the final round of the challenge to compete for prizes. The winners significantly outperformed the baseline method provided by the organizers. To measure influences between neurons the participants used an array of diverse methods, including transfer entropy, regression algorithms, correlation, deep learning, and network deconvolution. The development of connectivity reconstruction techniques is a major step in brain science, with many ramifications in the comprehension of neuronal computation, as well as the understanding of network dysfunctions in neuropathologies.'
volume: 46
URL: http://proceedings.mlr.press/v46/orlandi15.html
PDF: http://proceedings.mlr.press/v46/orlandi15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-orlandi15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Orlandi
given: Javier G.
- family: Ray
given: Bisakha
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Saeed
given: Mehreen
- family: Statnikov
given: Alexander
- family: Stetter
given: Olav
- family: Soriano
given: Jordi
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 1-22
id: orlandi15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 1
lastpage: 22
published: 2015-10-21 00:00:00 +0000
- title: 'Simple Connectome Inference from Partial Correlation Statistics in Calcium Imaging'
abstract: 'In this work, we propose a simple yet effective solution to the problem of connectome inference in calcium imaging data. The proposed algorithm consists of two steps. First, processing the raw signals to detect neural peak activities. Second, inferring the degree of association between neurons from partial correlation statistics. This paper summarises the methodology that led us to win the Connectomics Challenge, proposes a simplified version of our method, and finally compares our results with respect to other inference methods.'
volume: 46
URL: http://proceedings.mlr.press/v46/sutera15.html
PDF: http://proceedings.mlr.press/v46/sutera15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-sutera15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Sutera
given: Antonio
- family: Joly
given: Arnaud
- family: Francois-Lavet
given: Vincent
- family: Qiu
given: Aaron
- family: Louppe
given: Gilles
- family: Ernst
given: Damien
- family: Geurts
given: Pierre
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 23-35
id: sutera15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 23
lastpage: 35
published: 2015-10-21 00:00:00 +0000
- title: 'Supervised Neural Network Structure Recovery'
abstract: 'This paper presents our solution to the European Conference of Machine Learning Neural Connectomics Discovery Challenge. The challenge goal was to improve the performance of existing methods for recovering the neural network structure given the time series of neural activities. We propose to approximate a function able to combine several connectivity indicators between neuron pairs where each indicator is the result of running a feature engineering pipeline optimized for a particular noise level and firing synchronization rate among neurons. We proved the suitability of our solution by improving the state of the art prediction performance more than 6% and by obtaining the third best score on the test dataset out of 144 teams.'
volume: 46
URL: http://proceedings.mlr.press/v46/abril15.html
PDF: http://proceedings.mlr.press/v46/abril15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-abril15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Abril
given: Ildefons Magrans de
- family: Nowé
given: Ann
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 37-44
id: abril15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 37
lastpage: 44
published: 2015-10-21 00:00:00 +0000
- title: 'Signal Correlation Prediction Using Convolutional Neural Networks'
abstract: 'This paper focuses on analysing multiple time series relationships such as correlations between them. We develop a solution for the Connectiomics contest dataset of fluorescence imaging of neural activity recordings — the aim is reconstruction of the wiring between brain neurons. The model is implemented to achieve high evaluation score. Our model took the fourth place in this contest. The performance is similar to the other leading solutions, thus we showed that deep learning methods for time series processing are comparable to the other approaches and have wide opportunities for further improvement. We discuss a range of methods and code optimisations applied for the convolutional neural network for the time series domain.'
volume: 46
URL: http://proceedings.mlr.press/v46/romaszko15.html
PDF: http://proceedings.mlr.press/v46/romaszko15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-romaszko15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Romaszko
given: Lukasz
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 45-56
id: romaszko15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 45
lastpage: 56
published: 2015-10-21 00:00:00 +0000
- title: 'Reconstruction of Excitatory Neuronal Connectivity via Metric Score Pooling and Regularization'
abstract: 'Unravelling the causal link of neuronal pairs has considerable impacts in neuroscience, yet it still remains a major challenge. Recent investigations in the literature show that the Generalized Transfer Entropy (GTE), derived from information theory, has a great capability of reconstructing the underlying connectomics. In this work, we first generalize the GTE to a measure called Csiszar’s Transfer Entropy (CTE). With a proper choice of the convex function, the CTE outperforms the GTE in connectomic reconstruction, especially in the synchronized bursting regime where the GTE was reported to have poor sensitivity. Akin to the ensemble learning approach, we then pool various measures to achieve cutting edge neuronal network connectomic reconstruction performance. As a final step emphasize the importance of introducing regularization schemes in the network reconstruction.'
volume: 46
URL: http://proceedings.mlr.press/v46/tao15.html
PDF: http://proceedings.mlr.press/v46/tao15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-tao15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Tao
given: Chenyang
- family: Lin
given: Wei
- family: Feng
given: Jianfeng
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 57-66
id: tao15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 57
lastpage: 66
published: 2015-10-21 00:00:00 +0000
- title: 'Neural Connectivity Reconstruction from Calcium Imaging Signal using Random Forest with Topological Features'
abstract: 'Connectomics is becoming an increasingly popular area of research. With the recent advances in optical imaging of the neural activity tens of thousands of neurons can be monitored simultaneously. In this paper we present a method of incorporating topological knowledge inside data representation for Random Forest classifier in order to reconstruct the neural connections from patterns of their activities. Proposed technique leads to the model competitive with state-of-the art methods like Deep Convolutional Neural Networks and Graph Decomposition techniques. This claim is supported by the results (5th place with 0.003 in terms of AUC ROC loss to the top contestant) obtained in the connectomics competition organized on the Kaggle platform.'
volume: 46
URL: http://proceedings.mlr.press/v46/czarnecki15.html
PDF: http://proceedings.mlr.press/v46/czarnecki15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-czarnecki15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Czarnecki
given: Wojciech M.
- family: Jozefowicz
given: Rafal
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 67-76
id: czarnecki15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 67
lastpage: 76
published: 2015-10-21 00:00:00 +0000
- title: 'Effcient combination of pairwise feature networks'
abstract: 'This paper presents a novel method for the reconstruction of a neural network connectivity using calcium fluorescence data. We introduce a fast unsupervised method to integrate different networks that reconstructs structural connectivity from neuron activity. Our method improves the state-of-the-art reconstruction method General Transfer Entropy (GTE). We are able to better eliminate indirect links, improving therefore the quality of the network via a normalization and ensemble process of GTE and three new informative features. The approach is based on a simple combination of networks, which is remarkably fast. The performance of our approach is benchmarked on simulated time series provided at the connectomics challenge and also submitted at the public competition.'
volume: 46
URL: http://proceedings.mlr.press/v46/bellot15.html
PDF: http://proceedings.mlr.press/v46/bellot15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-bellot15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Bellot
given: Pau
- family: Meyer
given: Patrick E.
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 77-84
id: bellot15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 77
lastpage: 84
published: 2015-10-21 00:00:00 +0000
- title: 'Predicting Spiking Activities in DLS Neurons with Linear-Nonlinear-Poisson Model'
abstract: 'Spike train generation in primary motor cortex (M1) and somatosensory cortex (S1) has been studied extensively and is relatively well understood. On the contrary, the functionality and physiology of the dorsolateral striatum (DLS), the immediate downstream region of M1 and S1 and a critical link in the motor circuit, still requires intensive investigation. In the current study, spike trains of individual DLS neurons were reconstructed using a Linear-Nonlinear-Poisson model with features from two modalities: (1) the head position modality, which contains information regarding head movement and proprioception of the animal’s head; (2) the spike history modality, which contains information regarding the intrinsic physiological properties of the neuron. For the majority of the neurons examined, viable reconstruction accuracy was achieved when the neural activity was modeled with either feature modality or the two feature modalities combined. Subpopulations of neurons were also identiﬁied that had better reconstruction accuracy when modeled with features from single modalities. This study demonstrates the feasibility of spike train reconstruction in DLS neurons and provides insights into the physiology of DLS neurons. '
volume: 46
URL: http://proceedings.mlr.press/v46/ma15.html
PDF: http://proceedings.mlr.press/v46/ma15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-ma15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Ma
given: Sisi
- family: Barker
given: David J.
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 85-92
id: ma15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 85
lastpage: 92
published: 2015-10-21 00:00:00 +0000
- title: 'SuperSlicing Frame Restoration for Anisotropic ssTEM and Video Data'
abstract: 'In biological imaging the data is often represented by a sequence of anisotropic frames — the resolution in one dimension is significantly lower than in the other dimensions. E.g. in electron microscopy it arises from the thickness of a scanned section. This leads to blurred images and raises problems in tasks like neuronal image segmentation. We present the details and additional evaluation of an approach originally introduced in Laptev et al. (2014) called SuperSlicing to decompose the observed frame into a sequence of plausible hidden sub-frames. Based on sub-frame decomposition by SuperSlicing we propose a novel automated method to perform neuronal structure segmentation. We test our approach on a popular connectomics benchmark, where SuperSlicing preserves topological structures significantly better than other algorithms. We also generalize the approach for video anisotropicity that comes from the long exposure time and show that our method outperforms baseline methods on a reconstruction of low frame rate videos of natural scenes. '
volume: 46
URL: http://proceedings.mlr.press/v46/laptev15.html
PDF: http://proceedings.mlr.press/v46/laptev15.pdf
edit: https://github.com/mlresearch/v46/edit/gh-pages/_posts/2015-10-21-laptev15.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of the Neural Connectomics Workshop at ECML 2014'
publisher: 'PMLR'
author:
- family: Laptev
given: Dmitry
- family: Buhmann
given: Joachim M.
editor:
- family: Battaglia
given: Demian
- family: Guyon
given: Isabelle
- family: Lemaire
given: Vincent
- family: Soriano
given: Jordi
page: 93-103
id: laptev15
issued:
date-parts:
- 2015
- 10
- 21
firstpage: 93
lastpage: 103
published: 2015-10-21 00:00:00 +0000