Design and implementation of a brain computer interface system

In this dissertation we present the results of our effort to provide a modular, completely freeand open source brain-computer interface (BCI) system in Python. While other BCI systems exist, they are usually monolithic and written in a low level programming language like C++. Monolithic BCI systems cover the whole BCI processing chain from data acquisition, to signal processing, and feedbackand stimulus presentation, but they also force the user to use the whole system. This is problematic, as a researcher who is specialized, for example, in methods development will probably already have a toolbox of algorithms in a programming language he is familiar with. Being forced to use a monolithic system, means he has to migrate the tools and algorithms to the monolithic system first, and then work in an unfamiliar environment. Being written in C++, those BCI systems also make it unnecessary hard for non-computer scientists to add new functionality to those systems. Even among computer scientists, C++ is notorious for the skill level required to write decent C++ code. And in the field of BCI, only a small part of researchers are actually computer scientists. Those BCI systems try to minimize the need to write C++ code, by providing a large set of amplifier drivers, toolbox methods and paradigms, but since a large part in BCI research is in developing newor improving existing methods or paradigms, writing code is inevitable. With our BCI system we want to solve those problems altogether. We present a modular BCI system that is written in Python, an easy to learn yet powerful programming language fit for scientific computing. Our BCI system consists of three independent parts that can be used together as a complete BCI system, or independently, as a component in existing BCI systems. In this thesis we will cover for each component the requirements, the design aspects that lead to the implementation of the component, and demonstrate how to use the component individually in their domain. At the end of this thesis, we will demonstrate how to combine all three components to a complete BCI system and perform a real-time online experiment. Or system is freeand open source software licensed under free software licenses. The source code is freely available.