EEG feature engineering for epileptic seizure detection

This repo records the the MATLAB codes for the most challenging part, EEG feature engineering, for the epileptic seizure-detection task. These EEG features have been used by the published papers in Citation.

EEG feature domains

Feature No.	Feature domains	Description
1-4	basic statistics	mean amptitude, std amptitude, zero-cross times, range of amptitude
5-16	spectral analysis	power ratio and absolute power of each freq bands, alpha, beta, theta, delta, gamma, and freq centroid, total power
17-28	time-freq domain	mean and std of coefficients of discrete wavelet transform(DWT) on six freq bands
29-31	nonlinear features	ApEn, L-Z complexity, Hurst exponential
32-43	spatio-temporal domain	phase lock values on six freq bands and brain areas
44-47	synchronization measure (with fime/freq shift)	dynamic warpping in the time and freq domains
48-62	complex-network features	features extracted from the time & freq -invariant networks

Feature 1-47 refer to paper; feature 48-62 fefer to paper. The PhD thesis based on this work is at here.

EEG feature extraction
First, download this repo by:

git clone [email protected]:ieeeWang/EEG-feature-seizure-detection.git

Second, we here provide a demo for performing feature enginnering on four types of seizure EEG segments by run:

FeatureExtractDemo.m

Then, plot the EEG segments and EEG features by run:

plot-SEZseg.m
plot_Fv.m

Finally, the EEG feature mapping from EEG signals will look as below.

Citation
If you find this repository helpful, please cite our work:

@article{wang2017seizure,
  title={Seizure pattern-specific epileptic epoch detection in patients with intellectual disability},
  author={Wang, Lei and Arends, Johan BAM and Long, Xi and Cluitmans, Pierre JM and van Dijk, Johannes P},
  journal={Biomedical Signal Processing and Control},
  volume={35},
  pages={38--49},
  year={2017},
  publisher={Elsevier}
}
@article{wang2017eeg,
  title={EEG analysis of seizure patterns using visibility graphs for detection of generalized seizures},
  author={Wang, Lei and Long, Xi and Arends, Johan BAM and Aarts, Ronald M},
  journal={Journal of neuroscience methods},
  volume={290},
  pages={85--94},
  year={2017},
  publisher={Elsevier}
}
@article{wang2019broadband,
  title={A broadband method of quantifying phase synchronization for discriminating seizure EEG signals},
  author={Wang, Lei and Long, Xi and Aarts, Ronald M and van Dijk, Johannes P and Arends, Johan BAM},
  journal={Biomedical Signal Processing and Control},
  volume={52},
  pages={371--383},
  year={2019},
  publisher={Elsevier}
}
@article{wang2019eeg,
  title={EEG-based seizure detection in patients with intellectual disability: Which EEG and clinical factors are important?},
  author={Wang, Lei and Long, Xi and Aarts, Ronald M and van Dijk, Johannes P and Arends, Johan BAM},
  journal={Biomedical Signal Processing and Control},
  volume={49},
  pages={404--418},
  year={2019},
  publisher={Elsevier}
}