Section: Software and Platforms

Software and Platform

• The NeuralFieldSimulator (https://gforge.inria.fr/projects/nfsimulator/ ) computes numerically activity in two-dimensional neural fields by solving integral-differential equations involving transmission delays and visualizes the spatio-temporal activity. The tool includes a GUI that allows the user to choose field parameters. It is written in Python, open-source and is aimed to be promoted to become a major graphical visualization tool in the domain of neural field theory.

• AnaesthesiaSimulator (https://gforge.inria.fr/projects/anasim/ ) simulates the activity of networks of spiking neurons subject to specific receptor dynamics. The tool is a platform to test effects of anaesthetics on neural activity and is still in its first stage of development. The neural activity is planned to be visualized in a 2D and 3D-plot evolving in time. It is written in Python, open-source and involves heavily the simulation package BRIAN (http://briansimulator.org/ ).

OpenViBE (http://openvibe.inria.fr/ ) is a C++ open-source software devoted to the design, test and use of Brain-Computer Interfaces. The OpenViBE platform consists of a set of software modules that can be integrated easily and efficiently to design BCI applications. Key features of the platform are its modularity, high-performance, portability, its multiple-users facilities and its connection with high-end/Virtual Reality displays. The designer tool of the platform enables to build complete scenarios based on existing software modules using a dedicated graphical language and a simple Graphical User Interface (GUI). This software is available on the Inria Forge (https://gforge.inria.fr/projects/openvibe/ ) under the terms of the LGPL-V2 license. The development of OpenVibe is done in association with other Inria research teams (Hybrid, Athena, Potioc) for the national Inria project: ADT OpenViBE-NT. Neurosys is in charge of machine learning techniques and the interoperability with other tools such as Matlab, BCI2000, or TOBI.

The package DEvariants (https://sites.google.com/site/laurebuhry/publications/optimization-algorithms ) includes Matlab routines which implements new variants of the Differential Evolution (an evolutionary algorithm) strategies. The novelty lies in the selection process where we proposed to use a multinomial law to recombine the individuals/vectors. Compared to the standard strategies, our variants allow a faster convergence and a better avoidance of local minima. The different variants are provided with a test sample of functions, the DeJong benchmark. The audience is any scientific user familiar with evolutionary optimization.