NEST (research platform)

NEST (research platform)
Name	NEST
Caption	NEST simulation environment
Developer	Max Planck Society; University of Cologne; Jülich Research Centre
Released	1993
Latest release	2023
Programming language	C++; Python (programming language)
Operating system	Linux; Windows; macOS
License	GNU General Public License; OSI

NEST (research platform) is a simulation environment for large-scale spiking neuronal network models used in computational neuroscience, neural engineering, and theoretical neurobiology. The platform supports experiments that connect biological data from projects such as Human Brain Project, Blue Brain Project, and Allen Institute for Brain Science to models used by researchers at institutions such as Max Planck Society, University College London, and ETH Zurich. NEST emphasizes reproducibility, high-performance computing integration, and interoperability with tools like NEURON (software), Brian (simulator), and The Virtual Brain.

Overview

NEST provides a software framework that enables construction and simulation of networks of point-neuron and conductance-based neuron models, interfacing with datasets and tools from Human Connectome Project, European Research Council, Wellcome Trust, and National Institutes of Health. The environment offers bindings to scripting languages and connector libraries used by teams at Massachusetts Institute of Technology, Harvard University, Stanford University, and California Institute of Technology to design experiments spanning single-neuron dynamics to population-level activity. It integrates with workflow managers employed by researchers at Argonne National Laboratory, Lawrence Berkeley National Laboratory, and Oak Ridge National Laboratory for large-scale runs on supercomputers like Jülich Supercomputing Centre, Forschungszentrum Jülich, and NVIDIA-accelerated clusters.

History and Development

NEST originated in the early 1990s with contributions from groups at Max Planck Society and University of Oslo, evolving through collaborations with researchers at University of Cologne, University of Freiburg, and Jülich Research Centre. Funding and coordination have involved agencies such as German Research Foundation, European Commission, and Federal Ministry of Education and Research (Germany), alongside partnerships with projects like Human Brain Project and initiatives at Karolinska Institutet. The codebase expanded through integration of advances from teams at Cajal Blue Brain and methodological exchanges with developers of NEURON (software), Brian (simulator), and PyNN.

Architecture and Design

NEST employs a modular architecture implemented in C++ with scripting interfaces in Python (programming language), designed for parallel execution on distributed-memory systems such as those at Jülich Supercomputing Centre and Barcelona Supercomputing Center. Its internal event-driven kernel coordinates neuron and synapse models contributed by groups at Ecole Polytechnique Fédérale de Lausanne, University of Edinburgh, and University of Pennsylvania, while supporting model descriptions compatible with standards from International Neuroinformatics Coordinating Facility and NeuroML. Designs draw on software engineering practices used at Google, IBM, and Microsoft Research for scalability, testing, and continuous integration.

Simulation Features and Models

NEST supports integrate-and-fire neurons, adaptive exponential models, conductance-based schemes, and plasticity rules developed and evaluated by labs at Columbia University, University of California, San Diego, and University of Cambridge. The platform includes spike-timing-dependent plasticity, short-term plasticity, and homeostatic mechanisms used in studies by researchers at Max Planck Institute for Brain Research, University of Heidelberg, and Salk Institute. Models are validated against experimental datasets from Allen Institute for Brain Science, Neurodata Without Borders, and recordings shared by consortia like BRAIN Initiative.

Software Ecosystem and Interfaces

NEST integrates with Python (programming language) ecosystem tools such as NumPy, SciPy, and Matplotlib and interfaces with model specification languages like NeuroML and libraries like PyNN, enabling workflows shared by groups at EPFL, Imperial College London, and KTH Royal Institute of Technology. It interoperates with data-management systems used by European Molecular Biology Laboratory, INCF, and Dryad (repository) and connects to visualization and analysis tools from Human Brain Project partner projects and collaborators at Ghent University.

Performance and Scalability

Optimized for MPI-based parallelism and thread-level parallelism using libraries from OpenMP and MPI (Message Passing Interface), NEST scales to millions of neurons and billions of synapses on supercomputers operated by Jülich Research Centre, Leibniz Supercomputing Centre, and Swiss National Supercomputing Centre. Performance engineering leverages profiling and optimization techniques practiced at Intel Corporation, NVIDIA, and AMD and has been benchmarked in collaborative studies with Blue Brain Project and Human Brain Project computational teams.

Applications and Use Cases

Researchers use NEST to study cortical microcircuits investigated at Max Planck Institute for Brain Research, large-scale dynamics relevant to Human Brain Project objectives, sensory systems explored at Johns Hopkins University and Columbia University, and neuromorphic computing links pursued by groups at IBM Research, Intel Labs, and ETH Zurich. Applications span theoretical studies published via Nature Neuroscience, Neuron (journal), and PLOS Computational Biology and translational work connected to projects at Karolinska Institutet, Massachusetts General Hospital, and Weizmann Institute of Science.

Community and Governance

The project is guided by a developer community and steering bodies involving researchers from Max Planck Society, University of Cologne, Jülich Research Centre, and international partners at University of Oslo, EPFL, and University of Freiburg. Governance and sustainability are supported by grants from European Research Council, German Research Foundation, and collaborations with initiatives like Human Brain Project and coordination through organizations such as International Neuroinformatics Coordinating Facility.

Category:Computational neuroscience Category:Simulation software