Abstract— It is an open-ended challenge to accurately detect
the epileptic seizures through electroencephalogram (EEG)
signals. Recently published studies have made elaborate
attempts to distinguish between the normal and epileptic EEG
signals by advanced nonlinear entropy methods, such as the
approximate entropy, sample entropy, fuzzy entropy, and
permutation entropy, etc. Most recently, a novel distribution
entropy (DistEn) has been reported to have superior
performance compared with the conventional entropy methods
for especially short length data. We thus aimed, in the present
study, to show the potential of DistEn in the analysis of epileptic
EEG signals. The publicly-accessible Bonn database which
consisted of normal, interictal, and ictal EEG signals was used in
this study. Three different measurement protocols were set for
better understanding the performance of DistEn, which are: i)
calculate the DistEn of a specific EEG signal using the full
recording; ii) calculate the DistEn by averaging the results for
all its possible non-overlapped 5 second segments; and iii)
calculate it by averaging the DistEn values for all the possible
non-overlapped segments of 1 second length, respectively.
Results for all three protocols indicated a statistically
significantly increased DistEn for the ictal class compared with
both the normal and interictal classes. Besides, the results
obtained under the third protocol, which only used very short
segments (1 s) of EEG recordings showed a significantly (p <
0.05) increased DistEn for the interictal class in compassion with
the normal class, whereas both analyses using relatively long
EEG signals failed in tracking this difference between them,
which may be due to a nonstationarity effect on entropy
algorithm. The capability of discriminating between the normal
and interictal EEG signals is of great clinical relevance since it
may provide helpful tools for the detection of a seizure onset.
Therefore, our study suggests that the DistEn analysis of EEG
signals is very promising for clinical and even portable EEG
monitoring.
*Research supported by the China Postdoctoral Science Foundation
under Grant 2014M561933 and the National Natural Science Foundation of
China under Grant 61471223.
P. Li is with the School of Control Science and Engineering, Shandong
University, Jinan 250061, China (corresponding author to provide phone:
+86-15969688120; e-mail: pli@sdu.edu.cn).
C. Yan and C. Liu are with the School of Control Science and Engineering,
Shandong University, Jinan 250061, China (e-mail: cyan1210@gmail.com;
changchunliu@sdu.edu.cn)
C. Karmakar is with the centre of Pattern Recognition and Data Analytics
(PRaDA), Deakin Univeristy, Geelong, VIC 3220, Australia (e-mail:
karmakar@deakin.edu.au). He is also with Electrical & Electronic
Engineering Department, University of Melbourne, Parkville, Melbourne,
VIC 3010, Australia (email: karmakar@unimelb.edu.au).
I. INTRODUCTION
Epilepsy is a chronic brain disorder in humans that can
have adverse effects on social and psychological well-being.
About 1% of the world’s population have epilepsy [1], nearly
85% of which occur in developing countries [2]. People with
epilepsy are surrounded by possible lower educational
achievement, worse employment outcomes, and even a lot of
social stigma, suffering from paroxysmal occurrence of
seizures. These recurrent and sudden occurring of seizures are
dangerous and can lead to life-threatening situations.
Detection of epilepsy, especially at its onset, is thus very
beneficial for the safety of epileptic people and for improving
their quality of life.
Electroencephalogram (EEG) measures the electrical
activity of the brain. Abnormal epileptic waveforms can be
identified in EEG signals. Owing to the non-invasive and
low-cost nature of its measurement, EEG analysis is nearly the
most common effective diagnostic method for the detection of
epilepsy [3]. Since EEG signals are nonlinear, non-stationary
and complex in nature, to accurately describe the epileptic
EEGs’ intrinsic features are still now challenging.
Recently, several nonlinear entropy measures developed
from the concepts of chaos and nonlinear dynamics have been
introduced to the recognition of epileptic EEG signals. These
measures include approximate entropy (ApEn) [4, 5], sample
entropy (SampEn) [4, 6], fuzzy entropy (FuzzyEn) [7], and
permutation entropy (PE) [8, 9], etc. The published results
demonstrated that an entropy analysis of EEG signals may be
a promising prospect in the field of epilepsy detection [4-9].
Distribution entropy (DistEn), which has been developed
based upon the probability density of vector-to-vector
distances in state space, is a new measure, published most
recently by Li et al. in 2015, for assessing the complexity, or
chaotic behavior, of time-series [10]. DistEn has been shown
superior, as compared with SampEn and FuzzyEn, in
analyzing series with extremely short sampling points. It also
precludes the dependence upon input parameters of those
ApEn-based measures. In the present study, we aimed to show
the performance of DistEn in the analysis of epileptic EEG
signals.
II. DATA AND METHODS
A. The Open-Source EEG Data
The publicly-available Bonn database [11] was used in
this study. The whole dataset consisted of five sets (denoted as
Z, O, N, F, and S) and each of them contains 100
single-channel EEG segments of 23.6 seconds duration with
sampling frequency of 173.61 Hz. Sets Z and O were taken
from five healthy subjects. The subjects were awake and