Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals

Various information can be inferred from electrical signals generated by the brain using the electroencephalogram (EEG). This information can be used for the detection or classification of several diseases using many signal processing methods. Multifractal analysis has been widely used in the last t...

Full description

Saved in:
Bibliographic Details
Main Authors: Bechir Hbibi, Lamine Mili, Kamel Baccar, Abdelkader Mami
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11095672/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849768049337434112
author Bechir Hbibi
Lamine Mili
Kamel Baccar
Abdelkader Mami
author_facet Bechir Hbibi
Lamine Mili
Kamel Baccar
Abdelkader Mami
author_sort Bechir Hbibi
collection DOAJ
description Various information can be inferred from electrical signals generated by the brain using the electroencephalogram (EEG). This information can be used for the detection or classification of several diseases using many signal processing methods. Multifractal analysis has been widely used in the last two decades to examine various electrical signals generated by the human body. In this study, we analyze the multifractal characteristics of a data set comprising 27 EEG signals where the objective is to demonstrate the efficacy of this method in differentiating the states of vigilance in patients. These signals are obtained from comatose patients in the intensive care unit of the National Institute of Neurology of Tunis, each of whom had a documented Glasgow coma scale score. The purpose of analyzing these data is to identify the connection between the coma scales and the signals generated by some regions of the cerebral cortex in the brain and how this could lead to determining the state of the coma of these patients based on their electrical brain activities, especially due to the number of limitations of the behavioral and physiological coma scales currently used. Because EEG signals are multifractal, we applied the wavelet leader method to extract features from the singularity spectra of the recorded signals with a known Glasgow coma scale and classify them. We found that the singularity spectrum curves of patients in deep, mild, and light coma are distant from each other. Furthermore, we found that the spectra of patients in a light coma are shifted to the right compared to those of patients in a deep coma. On the other hand, curves calculated from the sedated patients are on the left of the curves of those in deep coma and are shrinking in shape compared to non-sedated patients. The lighter the coma, the more distinct the curves, especially the frontal channel curves. We also calculated the h values at the peak of the singularity spectrum curves of three different channels from the three other areas of the brain for the three groups. We found that the results from the back areas of the brain are aligned with those from the frontal lobe area. The h value averages of sedated patients, deep coma patients, and light coma patients are 0.116, 0.196, and 0.317, respectively. These results provided strong evidence that the multifractal analysis of EEG signals can be utilized to classify the vigilance states of comatose patients. However, although this analysis overcomes the limited physiological scales currently used, it still needs more investigation with a large sample of comatose patients.
format Article
id doaj-art-e76c969a842a4d4ba53fcaf39096a596
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-e76c969a842a4d4ba53fcaf39096a5962025-08-20T03:03:58ZengIEEEIEEE Access2169-35362025-01-011313768413769410.1109/ACCESS.2025.359251811095672Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG SignalsBechir Hbibi0https://orcid.org/0009-0000-7876-0577Lamine Mili1https://orcid.org/0000-0001-6134-3945Kamel Baccar2Abdelkader Mami3Application Laboratory of Energy Efficiency and Renewable Energies, Faculty of Sciences of Tunis, University of Tunis El Manar, Tunis, TunisiaArlington Research Center, Virginia Tech, Arlington, VA, USAIntensive Care and Anesthesia Department, National Institute of Neurology, Tunis, TunisiaApplication Laboratory of Energy Efficiency and Renewable Energies, Faculty of Sciences of Tunis, University of Tunis El Manar, Tunis, TunisiaVarious information can be inferred from electrical signals generated by the brain using the electroencephalogram (EEG). This information can be used for the detection or classification of several diseases using many signal processing methods. Multifractal analysis has been widely used in the last two decades to examine various electrical signals generated by the human body. In this study, we analyze the multifractal characteristics of a data set comprising 27 EEG signals where the objective is to demonstrate the efficacy of this method in differentiating the states of vigilance in patients. These signals are obtained from comatose patients in the intensive care unit of the National Institute of Neurology of Tunis, each of whom had a documented Glasgow coma scale score. The purpose of analyzing these data is to identify the connection between the coma scales and the signals generated by some regions of the cerebral cortex in the brain and how this could lead to determining the state of the coma of these patients based on their electrical brain activities, especially due to the number of limitations of the behavioral and physiological coma scales currently used. Because EEG signals are multifractal, we applied the wavelet leader method to extract features from the singularity spectra of the recorded signals with a known Glasgow coma scale and classify them. We found that the singularity spectrum curves of patients in deep, mild, and light coma are distant from each other. Furthermore, we found that the spectra of patients in a light coma are shifted to the right compared to those of patients in a deep coma. On the other hand, curves calculated from the sedated patients are on the left of the curves of those in deep coma and are shrinking in shape compared to non-sedated patients. The lighter the coma, the more distinct the curves, especially the frontal channel curves. We also calculated the h values at the peak of the singularity spectrum curves of three different channels from the three other areas of the brain for the three groups. We found that the results from the back areas of the brain are aligned with those from the frontal lobe area. The h value averages of sedated patients, deep coma patients, and light coma patients are 0.116, 0.196, and 0.317, respectively. These results provided strong evidence that the multifractal analysis of EEG signals can be utilized to classify the vigilance states of comatose patients. However, although this analysis overcomes the limited physiological scales currently used, it still needs more investigation with a large sample of comatose patients.https://ieeexplore.ieee.org/document/11095672/Comamultifractal analysissingularity spectrum analysisfeature extractionelectroencephalographyGlasgow coma scale
spellingShingle Bechir Hbibi
Lamine Mili
Kamel Baccar
Abdelkader Mami
Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
IEEE Access
Coma
multifractal analysis
singularity spectrum analysis
feature extraction
electroencephalography
Glasgow coma scale
title Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
title_full Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
title_fullStr Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
title_full_unstemmed Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
title_short Vigilance State Classification of Comatose Patients Based on Multifractal Analysis of EEG Signals
title_sort vigilance state classification of comatose patients based on multifractal analysis of eeg signals
topic Coma
multifractal analysis
singularity spectrum analysis
feature extraction
electroencephalography
Glasgow coma scale
url https://ieeexplore.ieee.org/document/11095672/
work_keys_str_mv AT bechirhbibi vigilancestateclassificationofcomatosepatientsbasedonmultifractalanalysisofeegsignals
AT laminemili vigilancestateclassificationofcomatosepatientsbasedonmultifractalanalysisofeegsignals
AT kamelbaccar vigilancestateclassificationofcomatosepatientsbasedonmultifractalanalysisofeegsignals
AT abdelkadermami vigilancestateclassificationofcomatosepatientsbasedonmultifractalanalysisofeegsignals