A library with some tools and functions for EEG signal analysis
The module eeglib is a library for Python that provides tools to analyse electroencephalography (EEG) signals. This library is mainly a feature extraction tool that includes lots of frequently used algorithms in EEG processing with using a sliding window approach. eeglib provides a friendly interface that allows data scientists who work with EEG signals to extract lots of features with just a few lines.
- Different types of processings
- Band Power
- Synchronization Likelihood
- Petrosian and Higuchi Fractal Dimensions
- Hjorth Parameters
- Detrended Fluctuation Analysis
- Sample Entropy
- Lempel-Ziv Complexity
- Cross Correlation Coeficient
- Load data from
- CSV files
- EDF files
- numpy arrays
- Feature extraction oriented
- Sliding window oriented
- Flexible and easy
Installation using pip:
$ pip install eeglib
Bellow there is a Quickstart Guide to eeglib. If you are interested in the API, you can find it here.
The next example shows a basic usage of the library. In it is shown how to load a file and apply a processing (Petrosian Fractal Dimension) to the data in windows of all the data.
from eeglib.helpers import CSVHelper helper= CSVHelper("fake_EEG_signal.csv") for eeg in helper: print(eeg.PFD())
This will show this:
[ 1.03089233 1.03229887 1.03181488 1.03123267 1.03069761]
This returns an array of the same size of the channels of the data (5) and each position of the array correspond with each channel.
The previous example applies the PFD to all the data in the file, but you may want to segment the data in different windows and that can be done in the next way:
helper= CSVHelper("fake_EEG_signal.csv",windowSize=256) for eeg in helper: print(eeg.PFD())
This will show this:
[ 1.03922468 1.03897773 1.03971798 1.03674636 1.03873059] [ 1.03848326 1.04168343 1.04094783 1.04168343 1.03699509] [ 1.03996434 1.04045647 1.03996434 1.03774006 1.03947143] [ 1.03749194 1.04045647 1.03897773 1.0402105 1.03873059]
Now the function has been called 4 times, this is because of the data has a lenght of 1024 samples and the window selected has a size of 256, so the windows contained in the data are 1024/256=4.
Now you may want to move the windows in another ways, like the ones that are shown in the next image:
So, if you want to make the windows overlap between them you can do it this way:
helper= CSVHelper("fake_EEG_signal.csv",windowSize=256) for eeg in helper[::128]: print(eeg.PFD())
Maybe you want to preprocess the signals stored in the window before extracting features from them. Currently this library allows the next Preprocessings:
- Bandpass filtering
- Z-Scores normalization
- Independent Component Analysis
These preprocessings can be applied at the load of the data by the Helpers:
helper = CSVHelper("fake_EEG_signal.csv", lowpass=30, highpass=1, normalize=True, ICA=True)
A Wrapper is an object that envelops a helper and simplifies the proccess of computing features that can be later be used, for example, in machine learning algorithms. The next example shows an example of how wrappers can be used:
from eeglib import wrapper, helpers helper = helpers.CSVHelper("fake_EEG_signal.csv", windowSize=128) wrap = wrapper.Wrapper(helper) wrap.addFeature.HFD() wrap.addFeature.getFourierTransform() wrap.addFeature.synchronizationLikelihood() features=wrap.getAllFeatures()
So, the scheme to follow with wrappers is the next:
- Create the Helper object.
- Create the wrapper object.
- Select the desired features to compute. They can be parameterized by adding the parameters just behind the name.
- Call the method "getAllFeatures()" in order to compute every feature from every window at once or iterate over the Wrapper object for obtaining the features of each window. They are returned as a pandas.DataFrame or a pandas.Series.
If eeglib has been useful in your research, please, consider citing the next article.
This library was initialy a Final Degree Project and you can find the documentation of the development in the next link:
Later it was extented as part of a Master's thesis that can be found in the next link:
There are also some papers related to this library that can be seen bellow:
- Computational EEG Analysis Techniques When Playing Video Games: A Systematic Review
- Analysis of Cognitive Load Using EEG when Interacting with Mobile Devices
- eeglib: A Python module for EEG feature extraction
Not open access
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size eeglib-0.4.1-py3-none-any.whl (25.6 kB)||File type Wheel||Python version py3||Upload date||Hashes View|
|Filename, size eeglib-0.4.1.tar.gz (25.2 kB)||File type Source||Python version None||Upload date||Hashes View|