On robust estimation of the location parameter

Artificial neural networks (ANNs) constitute a class of flexible nonlinear models designed to mimic biological neural systems. In this entry, we introduce ANN using familiar econometric terminology and provide an overview of ANN modeling approach and its implementation methods. † Correspondence: Chung-Ming Kuan, Institute of Economics, Academia Sinica, 128 Academia Road, Sec. 2, Taipei 115, Taiwan; ckuan@econ.sinica.edu.tw. †† I would like to express my sincere gratitude to the editor, Professor Steven Durlauf, for his patience and constructive comments on early drafts of this entry. I also thank Shih-Hsun Hsu and Yu-Lieh Huang for very helpful suggestions. The remaining errors are all mine.

Artificial neural networks

New silicon probes known as Neuropixels are shown to record from hundreds of neurons simultaneously in awake and freely moving rodents. Sensory, motor and cognitive operations involve the coordinated action of large neuronal populations across multiple brain regions. Existing technologies reliably measure activity from a relatively small number of neurons with high spatial and temporal resolution, or from a large volume of neurons with low resolution. Timothy Harris and colleagues describe the design, fabrication and performance of Neuropixels, a silicon probe that can measure well-isolated neural activity from hundreds of neurons. They integrated these probes into a lightweight system that could record activity simultaneously and with high fidelity from hundreds of neurons in awake and freely moving rodents. Sensory, motor and cognitive operations involve the coordinated action of large neuronal populations across multiple brain regions in both superficial and deep structures1,2. Existing extracellular probes record neural activity with excellent spatial and temporal (sub-millisecond) resolution, but from only a few dozen neurons per shank. Optical Ca2+ imaging3,4,5 offers more coverage but lacks the temporal resolution needed to distinguish individual spikes reliably and does not measure local field potentials. Until now, no technology compatible with use in unrestrained animals has combined high spatiotemporal resolution with large volume coverage. Here we design, fabricate and test a new silicon probe known as Neuropixels to meet this need. Each probe has 384 recording channels that can programmably address 960 complementary metal–oxide–semiconductor (CMOS) processing-compatible low-impedance TiN6 sites that tile a single 10-mm long, 70 × 20-μm cross-section shank. The 6 × 9-mm probe base is fabricated with the shank on a single chip. Voltage signals are filtered, amplified, multiplexed and digitized on the base, allowing the direct transmission of noise-free digital data from the probe. The combination of dense recording sites and high channel count yielded well-isolated spiking activity from hundreds of neurons per probe implanted in mice and rats. Using two probes, more than 700 well-isolated single neurons were recorded simultaneously from five brain structures in an awake mouse. The fully integrated functionality and small size of Neuropixels probes allowed large populations of neurons from several brain structures to be recorded in freely moving animals. This combination of high-performance electrode technology and scalable chip fabrication methods opens a path towards recording of brain-wide neural activity during behaviour.

/pdf/fully-integrated-silicon-probes-for-high-density-recording-nnob1hwgn8.pdf

Fully integrated silicon probes for high-density recording of neural activity

Current neuroimaging software offer users an incredible opportunity to analyze their data in different ways, with different underlying assumptions. Several sophisticated software packages (e.g., AFNI, BrainVoyager, FSL, FreeSurfer, Nipy, R, SPM) are used to process and analyze large and often diverse (highly multi-dimensional) data. However, this heterogeneous collection of specialized applications creates several issues that hinder replicable, efficient and optimal use of neuroimaging analysis approaches: 1) No uniform access to neuroimaging analysis software and usage information; 2) No framework for comparative algorithm development and dissemination; 3) Personnel turnover in laboratories often limits methodological continuity and training new personnel takes time; 4) Neuroimaging software packages do not address computational efficiency; and 5) Methods sections in journal articles are inadequate for reproducing results. To address these issues, we present Nipype (Neuroimaging in Python: Pipelines and Interfaces; http://nipy.org/nipype), an open-source, community-developed, software package and scriptable library. Nipype solves the issues by providing Interfaces to existing neuroimaging software with uniform usage semantics and by facilitating interaction between these packages using Workflows. Nipype provides an environment that encourages interactive exploration of algorithms, eases the design of Workflows within and between packages, allows rapid comparative development of algorithms and reduces the learning curve necessary to use different packages. Nipype supports both local and remote execution on multi-core machines and clusters, without additional scripting. Nipype is BSD licensed, allowing anyone unrestricted usage. An open, community-driven development philosophy allows the software to quickly adapt and address the varied needs of the evolving neuroimaging community, especially in the context of increasing demand for reproducible research.

/pdf/nipype-a-flexible-lightweight-and-extensible-neuroimaging-ksf6bnxjjz.pdf

Nipype: A Flexible, Lightweight and Extensible Neuroimaging Data Processing Framework in Python

In order to understand how the mammalian neocortex is performing computations, two things are necessary; we need to have a good understanding of the available neuronal processing units and mechanisms, and we need to gain a better understanding of how those mechanisms are combined to build functioning systems. Therefore, in recent years there is an increasing interest in how spiking neural networks (SNN) can be used to perform complex computations or solve pattern recognition tasks. However, it remains a challenging task to design SNNs which use biologically plausible mechanisms (especially for learning new patterns), since most of such SNN architectures rely on training in a rate-based network and subsequent conversion to a SNN. We present a SNN for digit recognition which is based on mechanisms with increased biological plausibility, i.e. conductance-based instead of current-based synapses, spike-timing-dependent plasticity with time-dependent weight change, lateral inhibition, and an adaptive spiking threshold. Unlike most other systems, we do not use a teaching signal and do not present any class labels to the network. Using this unsupervised learning scheme, our architecture achieves 95% accuracy on the MNIST benchmark, which is better than previous SNN implementations without supervision. The fact that we used no domain-specific knowledge points toward the general applicability of our network design. Also, the performance of our network scales well with the number of neurons used and shows similar performance for four different learning rules, indicating robustness of the full combination of mechanisms, which suggests applicability in heterogeneous biological neural networks.

/pdf/unsupervised-learning-of-digit-recognition-using-spike-4aclr1f742.pdf

Unsupervised learning of digit recognition using spike-timing-dependent plasticity.

Developments in microfabrication technology have enabled the production of neural electrode arrays with hundreds of closely spaced recording sites, and electrodes with thousands of sites are under development. These probes in principle allow the simultaneous recording of very large numbers of neurons. However, use of this technology requires the development of techniques for decoding the spike times of the recorded neurons from the raw data captured from the probes. Here we present a set of tools to solve this problem, implemented in a suite of practical, user-friendly, open-source software. We validate these methods on data from the cortex, hippocampus and thalamus of rat, mouse, macaque and marmoset, demonstrating error rates as low as 5%.

/pdf/spike-sorting-for-large-dense-electrode-arrays-56e8lxdbd3.pdf

Spike sorting for large, dense electrode arrays

"Brian" is a simulator for spiking neural networks (http://www.briansimulator.org). The focus is on making the writing of simulation code as quick and easy as possible for the user, and on flexibility: new and non-standard models are no more difficult to define than standard ones. This allows scientists to spend more time on the details of their models, and less on their implementation. Neuron models are defined by writing differential equations in standard mathematical notation, facilitating scientific communication. Brian is written in the Python programming language, and uses vector-based computation to allow for efficient simulations. It is particularly useful for neuroscientific modelling at the systems level, and for teaching computational neuroscience.

The brian simulator.

"Brian" is a new simulator for spiking neural networks, written in Python (http://briandiensfr) It is an intuitive and highly flexible tool for rapidly developing new models, especially networks of single-compartment neurons In addition to using standard types of neuron models, users can define models by writing arbitrary differential equations in ordinary mathematical notation Python scientific libraries can also be used for defining models and analysing data Vectorisation techniques allow efficient simulations despite the overheads of an interpreted language Brian will be especially valuable for working on non-standard neuron models not easily covered by existing software, and as an alternative to using Matlab or C for simulations With its easy and intuitive syntax, Brian is also very well suited for teaching computational neuroscience

/pdf/brian-a-simulator-for-spiking-neural-networks-in-python-1la8h75g7t.pdf

Brian: A Simulator for Spiking Neural Networks in Python

/pdf/brian-a-simulator-for-spiking-neural-networks-in-python-1vx36bghdr.pdf

Brian: a simulator for spiking neural networks in Python

Simulating the brain starts with understanding the activity of a single neuron. From there, it quickly gets very complicated. To reconstruct the brain with computers, neuroscientists have to first understand how one brain cell communicates with another using electrical and chemical signals, and then describe these events using code. At this point, neuroscientists can begin to build digital copies of complex neural networks to learn more about how those networks interpret and process information. To do this, computational neuroscientists have developed simulators that take models for how the brain works to simulate neural networks. These simulators need to be able to express many different models, simulate these models accurately, and be relatively easy to use. Unfortunately, simulators that can express a wide range of models tend to require technical expertise from users, or perform poorly; while those capable of simulating models efficiently can only do so for a limited number of models. An approach to increase the range of models simulators can express is to use so-called ‘model description languages’. These languages describe each element within a model and the relationships between them, but only among a limited set of possibilities, which does not include the environment. This is a problem when attempting to simulate the brain, because a brain is precisely supposed to interact with the outside world. Stimberg et al. set out to develop a simulator that allows neuroscientists to express several neural models in a simple way, while preserving high performance, without using model description languages. Instead of describing each element within a specific model, the simulator generates code derived from equations provided in the model. This code is then inserted into the computational experiments. This means that the simulator generates code specific to each model, allowing it to perform well across a range of models. The result, Brian 2, is a neural simulator designed to overcome the rigidity of other simulators while maintaining performance. Stimberg et al. illustrate the performance of Brian 2 with a series of computational experiments, showing how Brian 2 can test unconventional models, and demonstrating how users can extend the code to use Brian 2 beyond its built-in capabilities.

Dan F. M. Goodman

Papers

Spike sorting for large, dense electrode arrays

The brian simulator.

Brian: A Simulator for Spiking Neural Networks in Python

Brian: a simulator for spiking neural networks in Python

Brian 2, an intuitive and efficient neural simulator