In the last years Python has gained more and more traction in the scientific community. Projects like Numpy, SciPy, and Matplotlib have created a strong foundation for scientific computing in Python and machine learning packages like Scikit-learn or packages for data analysis like Pandas are building on top of it. Yet, in the brain-computer interfacing (BCI) community Matlab is still the dominant programming language.
We present Wyrm, an open source BCI toolbox in Python. Wyrm is applicable to a wide range of neuroscientific problems. It can be used as a toolbox for analysis and visualization of neurophysiological data (e.g. EEG, ECoG, fMRI, or NIRS) and it is suitable for real-time online experiments. In Wyrm we implemented dozens of methods, covering a broad range of aspects for off-line analysis and online experiments. The list of algorithms includes: channel selection, IIR filters, sub-sampling, spectrograms, spectra, baseline removal for signal processing; Common Spatial Patterns (CSP), Source Power Co-modulation (SPoC), classwise average, jumping means, signed $r^2$-values for feature extraction; Linear Discriminant Analyis (LDA) with and without shrinkage for machine learning; various plotting methods and many more. It is worth mentioning that with scikit-learn you have a wide range of machine learning algorithms readily at your disposal. Our data format is very compatible with scikit-learn and one can usually apply the algorithms without any data conversion step at all.
Since the correctness of its methods is crucial for a toolbox, we used unit testing to ensure all methods work as intended. In our toolbox each method is tested respectively by at least a handful of test cases to ensure that the methods calculate the correct results, throw the expected errors if necessary, etc. The total amount of code for all tests is roughly 2-3 times bigger than the amount code for the toolbox methods.
As a software toolbox would be hard to use without proper documentation, we provide documentation that consists of readable prose and extensive API documentation. Each method of the toolbox is thoroughly documented and has usually a short summary, a detailed description of the algorithm, a list of expected inputs, return values and exceptions, as well as cross references to related methods in- or outside the toolbox and example code to demonstrate how to use the method.
The ongoing transition from Python 2 to Python 3 was also considered and we decided to support both Python versions. Wyrm is mainly developed under Python 2.7, but written in a forward compatible way to support Python 3 as well. Our unit tests ensure that the methods provide the expected results in Python 2 and Python 3.
To show how to use the toolbox realistic scenarios we provide two off-line analysis scripts, where we demonstrate how to use the toolbox to complete two tasks from the BCI Competition III. The data sets from the competition are freely available and one can reproduce our results using the scripts and the data. We also provide a simulated online BCI experiment using a data set from the same competition.
Together with Mushu our signal acquisition library and Pyff our Framework for Feedback and Stimulus Presentation, Wyrm adds the final piece to our ongoing effort to provide a complete, free and open source BCI system in Python.
