Schizophrenia diagnosis based on diverse epoch size resting-state EEG using machine learning

View article
PeerJ Computer Science

Main article text

 

Introduction

  1. To propose a highly effective method for classifying epoched EEG signals of SZ and healthy controls using machine learning algorithms in terms of decreased complexity.

  2. To investigate the effects of augmentation techniques using two methods SNR and stretch on classification accuracy.

  3. To compare the classification performance of less number of electrode usage with varying epoch lengths of 1-, 2-, and 5-s.

Materials and Methods

Dataset

Processing and feature extraction

Fast Fourier transform

Approximate entropy

Shannon entropy

Log energy entropy

where (wn2i,j) are the computed WPD coefficients.

Kurtosis

Normalization

Classification

Support vector machine

K-nearest neighbors

Quadratic discriminant analysis

Ensemble classifier

Evaluation metrics

Results

One-second epoch size

Two-second epoch size

Five-second epoch size

Discussion

  1. The data set is open access instead of a collected private dataset or one gained from cohort studies. Even though the number of subjects was low, we used the epoching approach that enabled us to increase the size of the input data for machine learning.

  2. The dataset in this article aimed to identify the condition rather than to assess its degree of severity. Thus, it does not provide the stage of schizophrenia (prodromal, active, and residual).

Conclusion

  1. Using these feature extraction methods with this type of signal can reach high results.

  2. Window epoch sizes could enhance classification precision.

  3. The outcomes fluctuate depending on the window size. As a result, the signal decomposition window-size epoch is important for identifying tiny changes in the EEG recording.

Supplemental Information

One-Second Epoch Size Confusion Matrix Results.

DOI: 10.7717/peerj-cs.2170/supp-1

One-Second Epoch Size Confusion Matrix Results with SNR.

DOI: 10.7717/peerj-cs.2170/supp-2

One-Second Epoch Size Confusion Matrix Results with Stretch.

DOI: 10.7717/peerj-cs.2170/supp-3

One-Second Epoch Size Confusion Matrix Results with 5 electrodes.

DOI: 10.7717/peerj-cs.2170/supp-4

One-Second Epoch Size Confusion Matrix Results with 8 electrodes.

DOI: 10.7717/peerj-cs.2170/supp-5

Two-Second Epoch Size Confusion Matrix Results.

DOI: 10.7717/peerj-cs.2170/supp-6

Two-Second Epoch Size Confusion Matrix Results with SNR.

DOI: 10.7717/peerj-cs.2170/supp-7

Two-Second Epoch Size Confusion Matrix Results with Stretch.

DOI: 10.7717/peerj-cs.2170/supp-8

Two-Second Epoch Size Confusion Matrix Results with 5 electrodes.

DOI: 10.7717/peerj-cs.2170/supp-9

Two-Second Epoch Size Confusion Matrix Results with 8 electrodes.

DOI: 10.7717/peerj-cs.2170/supp-10

Five-Second Epoch Size Confusion Matrix Results.

DOI: 10.7717/peerj-cs.2170/supp-11

Five-Second Epoch Size Confusion Matrix Results with SNR augmented method.

DOI: 10.7717/peerj-cs.2170/supp-12

Five-Second Epoch Size Confusion Matrix Results with Stretch signals.

DOI: 10.7717/peerj-cs.2170/supp-13

Five-Second Epoch Size Confusion Matrix Results with 5 electrodes.

Five-Second Epoch Size Confusion Matrix Results with 5 electrodes

DOI: 10.7717/peerj-cs.2170/supp-14

Five-Second Epoch Size Confusion Matrix Results with 8 electrodes.

DOI: 10.7717/peerj-cs.2170/supp-15

One-second epoch size receiver operating characteristics curves (ROC) for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-16

One-second epoch size receiver operating characteristics curves (ROC) with SNR for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-17

One-second epoch size receiver operating characteristics curves (ROC) with Stretch for the four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-18

One-second epoch size receiver operating characteristics curves (ROC) with 5 electrodes for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-19

One-second epoch size receiver operating characteristics curves (ROC) with 8 electrodes for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-20

Two-second epoch size receiver operating characteristics curves (ROC) for the four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-21

Two-second epoch size receiver operating characteristics curves (ROC) with SNR for the four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-22

Two-second epoch size receiver operating characteristics curves (ROC) with Stretch for the four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-23

Two-second epoch size receiver operating characteristics curves (ROC) for the 5 electrodes with four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-24

Two-second epoch size receiver operating characteristics curves (ROC) for the 8 electrodes with four classifiers.

DOI: 10.7717/peerj-cs.2170/supp-25

Five-second epoch size receiver operating characteristics curves (ROC) for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-26

Five-second epoch size receiver operating characteristics curves (ROC) with SNR for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-27

Five-second epoch size receiver operating characteristics curves (ROC) with Stretch for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-28

Five-second epoch size receiver operating characteristics curves (ROC) with 5 Electrodes for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-29

Five-second epoch size receiver operating characteristics curves (ROC) with 8 Electrodes for the four classifiers with six features extraction types.

DOI: 10.7717/peerj-cs.2170/supp-30

Additional Information and Declarations

Competing Interests

Ivan Miguel Pires and Paulo Jorge Coelho are Academic Editors for PeerJ Computer Science.

Author Contributions

Athar Alazzawı conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Saif Aljumaili conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Adil Deniz Duru conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Osman Nuri Uçan conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Oğuz Bayat conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Paulo Jorge Coelho conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Ivan Miguel Pires conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Data Availability

The following information was supplied regarding data availability:

The EEG in schizophrenia dataset is available at: Olejarczyk, Elzbieta; Jernajczyk, Wojciech, 2017, “EEG in schizophrenia”, https://doi.org/10.18150/repod.0107441, RepOD, V1.

The code is available at Zenodo: Alazzawi, A., Aljumaili, S., Duru, A. D., Uçan, O. N., Bayat, O., Coelho, P. J., & Pires, I. M. (2024). Schizophrenia diagnosis based on diverse epoch size using EEG signal on resting-state. Zenodo. https://doi.org/10.5281/zenodo.11075184.

Funding

This work was supported by FCT-Fundação para a Ciência e Tecnologia, I.P. by project reference UIDB/50008/2020, and DOI identifier https://doi.org/10.54499/UIDB/50008/2020. This work was also funded by FCT/MEC through national funds and co-funded by the FEDER-PT2020 partnership agreement under the project UIDB/00308/2020 (DOI 10.54499/UIDB/00308/2020). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

1 Citation 1,186 Views 69 Downloads

Your institution may have Open Access funds available for qualifying authors. See if you qualify

Publish for free

Comment on Articles or Preprints and we'll waive your author fee
Learn more

Five new journals in Chemistry

Free to publish • Peer-reviewed • From PeerJ
Find out more