We've updated our privacy policy.

Multiscale entropy-based EEG artifact detector



Electroencephalographic (EEG) signals present a myriad of challenges to analysis, beginning with the detection of artifacts. This simple algorithm is based on a statistical physics method, multiscale entropy (MSE) analysis, which quantifies the complexity of a signal. Noise corrupted EEG signals have lower information content, and, therefore, reduced complexity compared with their noise free counterparts. The method allows to identify epochs of corrupted noise on a given EEG signal. On an epoch-by-epoch base, the complexity index (CI) of the signal is computed as the sum of the MSE on the first 5 scales. Epochs having a CI lower than a pre-set threshold are scored as containing an artifact.

The code employs the function msentropy.m, from the WFDB Matlab toolbox.


Mandatory inputs to the function are:

  • The EEG signal where noisy epochs should be recognized (nx1 vector)
  • The sampling rate of the EEG (1x1 integer) Optional inputs to the function are:
  • The length of the analysis epoch for the identification of artifacts, in seconds (integer, default = 2 s)
  • The amount of overlap between epochs, in percentage (0-100) (default = 0)
  • The artifact detection threshold. Epochs where the CI is lower than the threshold are scored as artifacts (default = 1.34) Along with the function, we provide an example that employs one of the noise-corrupted EEG signals from the the Motion Artifact Contaminated fNIRS and EEG Data Database on Physionet.


Matlab R2015a and later versions. Please note that the WFDB Matlab toolbox must be installed and in your Matlab path.


The MSE-based artifact detector software package was developed at the Wyss Institute for Biologically Inspired Engineering at Harvard University and Beth Israel Deaconess Medical Center/Harvard Medical School by Sara Mariani, Filipa Borges and Teresa S. Henriques.

When referencing this software, please cite: Sara Mariani, Ana F. T. Borges, Teresa Henriques, Ary L. Goldberger, Madalena D. Costa: Use of multiscale entropy to facilitate artifact detection in electroencephalographic signals. EMBC 2015: 7869-7872

More Questions

Please report bugs and questions at sara.mariani7@gmail.com.

Related links