https://doi.org/10.1177/1550059420905724
Clinical EEG and Neuroscience
1–9
© EEG and Clinical Neuroscience
Society (ECNS) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/1550059420905724
journals.sagepub.com/home/eeg
Original Article
Introduction
In the past decades, machine learning (ML) that consists of
multiple techniques has become one of the most common
approaches for both prognostic and diagnostic purposes.
Classification techniques of ML like artificial neural networks
(ANNs) are extensively used in methods proposing solutions
to prognostic and diagnostic problems in medical domains by
predicting categorical variables for a specific disease.
1
Substance use disorder (SUD) is a crucial and serious health
problem to be taken into consideration causing impairments in
cognition and behavior. There were 183 million cannabis, 35
million opioid, and 37 million amphetamine users by 2015.
2
.
Opioids are a category of substances that have a morphine-
type action in the body and include heroin, which has a highly
addictive potential and results in mortality and morbidity.
3
Gomes et al
4
reported that the percentage of deaths attributable
to opioids increased and in the 24- to 35-year age group, 20%
of deaths were attributable to opioids in 2016. Moreover, opi-
oid treatment is costly; according to Kirson et al,
5
opioid abus-
ers had significantly higher annual health care resource
utilization, leading to $14 810 in per-patient cost compared
with nonabusers.
905724EEG XX X 10.1177/1550059420905724Clinical EEG and NeuroscienceErguzel et al
research-article 2020
1
Department of Software Engineering, Faculty of Engineering and Natural
Sciences, Uskudar University, Istanbul, Turkey
2
Department of Mechatronics, Faculty of Engineering, Bulent Evevit
University, Zonguldak, Turkey
3
Department of Psychology, Faculty of Humanities and Social Sciences,
Uskudar University, Istanbul, Turkey
4
NP Istanbul Brain Hospital, Istanbul, Turkey
Corresponding Author:
Turker Tekin Erguzel, Department of Software Engineering, Faculty of
Engineering and Natural Sciences, Uskudar University, Altunizade Mah. Haluk
Turksoy Sk. No. 14, Uskudar, Istanbul, PK 34662, Turkey.
Email: turker.erguzel@uskudar.edu.tr
Entropy: A Promising EEG Biomarker
Dichotomizing Subjects With Opioid
Use Disorder and Healthy Controls
Turker Tekin Erguzel
1
, Caglar Uyulan
2
, Baris Unsalver
3,4
, Alper Evrensel
3,4
,
Merve Cebi
3
, Cemal Onur Noyan
3,4
, Baris Metin
3,4
, Gul Eryilmaz
3,4
,
Gokben Hizli Sayar
3,4
, and Nevzat Tarhan
3,4
Abstract
Electroencephalography (EEG) signals are known to be nonstationary and often multicomponential signals containing information
about the condition of the brain. Since the EEG signal has complex, nonlinear, nonstationary, and highly random behaviour,
numerous linear feature extraction methods related to the short-time windowing technique do not satisfy higher classification
accuracy. Since biosignals are highly subjective, the symptoms may appear at random in the time scale and very small variations
in EEG signals may depict a definite type of brain abnormality it is valuable and vital to extract and analyze the EEG signal
parameters using computers. The challenge is to design and develop signal processing algorithms that extract this subtle
information and use it for diagnosis, monitoring, and treatment of subjects suffering from psychiatric disorders. For this purpose,
finite impulse response–based filtering process was employed rather than traditional time and frequency domain methods. Finite
impulse response subbands were analyzed further to obtain feature vectors of different entropy markers and these features
were fed into a classifier namely multilayer perceptron. The performances of the classifiers were finally compared considering
overall classification accuracies, area under receiver operating characteristic curve scores. Our results underline the potential
benefit of the introduced methodology is promising and is to be treated as a clinical interface in dichotomizing substance use
disorders subjects and for other medical data analysis studies. The results also indicate that entropy estimators can distinguish
normal and opioid use disorder subjects. EEG data and theta frequency band have distinctive capability for almost all types of
entropies while nonextensive Tsallis entropy outperforms compared with other types of entropies.
Keywords
entropy, opioid use disorder, artificial neural network (ANN), signal processing, EEG
Received March 13, 2019; revised November 18, 2019; accepted January 16, 2020.