Brain Waves and Evoked Potentials as Biometric User Identification
Strategy: An Affordable Low-cost Approach
Roberto Saia, Salvatore Carta, Gianni Fenu and Livio Pompianu
Department of Mathematics and Computer Science,
University of Cagliari, Via Ospedale 72 - 09124 Cagliari, Italy
Keywords:
Security, Biometric, Brain Waves, Electroencephalography, EEG, Evoked Potentials, EP, Muse.
Abstract:
The relatively recent introduction on the market of low-cost devices able to perform an Electroencephalogra-
phy (EEG) has opened a stimulating research scenario that involves a large number of researchers previously
excluded due to the high costs of such hardware. In this regard, one of the most stimulating research fields
is focused on the use of such devices in the context of biometric systems, where the EEG data are exploited
for user identification purposes. Based on the current literature, which reports that many of these systems
are designed by combining the EEG data with a series of external stimuli (Evoked Potentials) to improve the
reliability and stability over time of the EEG patterns, this work is aimed to formalize a biometric identifica-
tion system based on low-cost EEG devices and simple stimulation instruments, such as images and sounds
generated by a computer. In other words, our objective is to design a low-cost EEG-based biometric approach
exploitable on a large number of real-world scenarios.
1 INTRODUCTION
The introduction on the market of low-cost devices
capable of detecting brain waves with relative ac-
curacy has opened numerous research paths rang-
ing from canonical domains (Xavier et al., 2021;
Prathibha et al., 2017) to completely new ones, such
as those oriented to the definition of biometric sys-
tems (Nakanishi and Maruoka, 2019). The grow-
ing number of such devices is accompanied by an
equally increasing availability of applications and li-
braries to support the development of new applica-
tions: a representative example is that of the Muse
(https://choosemuse.com) device, marketed together
with an application aimed at supporting meditation
activities, and widely used in literature for numerous
experiments, also thanks to a large number of pro-
gramming libraries available.
More formally, these devices perform an Elec-
troencephalogram (EEG) (Thakor and Sherman,
2013) by measuring the electrical activity of the brain
using a series of electrodes positioned on the scalp.
In this regard, it should be added that the number
of these electrodes is greater in professional devices,
where even the positioning requires more attention,
requiring also the use of a conductive paste, differ-
ently from the low-cost devices that use a smaller
number of electrodes, which can usually do not re-
quire any conductive paste. However, in both cases
the positioning of the electrodes follows the 10-20
System of Electrode Placement (Homan et al., 1987)
formalization shown in Figure 1.
Such a system indicates the relationship between
each electrode location and the underlying cerebral
cortex area. In order to guide in the correct position-
ing of the electrodes, each position reports the lobe
and the hemisphere indication, i.e., the C, F, P, O, and
T letters indicate, respectively, the Central, Frontal,
Parietal, Occipital, and Temporal lobe (the Central
lobe is used only for identification purposes, as it does
not actually exist). Similarly, the even numbers 2, 4,
6, 8, and 10 denote the right hemisphere, and the odd
numbers 1, 3, 5, 7, and 9 the left one; the z letter indi-
cates a median-line electrode (the smaller is a number,
the closer it is to the median line); the Nasion label
denotes the area between the forehead and nose and
Inion label denotes the area at the back of the skull.
At the time of writing, the most popular low/medium
cost EEG signal acquisition devices are those reported
in Table 1, which also provides information about the
data resolution in bits and the number of electrodes.
Based on the cost and features, as well as the
availability of development tools, among the afore-
mentioned devices we have chosen for this work the
InteraXon Muse-2, a device marketed as a tool that
helps in reaching a deep relaxed state, assisting the
614
Saia, R., Carta, S., Fenu, G. and Pompianu, L.
Brain Waves and Evoked Potentials as Biometric User Identification Strategy: An Affordable Low-cost Approach.
DOI: 10.5220/0011297600003283
In Proceedings of the 19th International Conference on Security and Cryptography (SECRYPT 2022), pages 614-619
ISBN: 978-989-758-590-6; ISSN: 2184-7711
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
people during a meditation session, exploiting an au-
ral feedback related to the detected brain activity. It
uses five electrodes placed on a headband that, on the
basis of the placements of Figure 1, are: a reference
electrode (NZ) and four acquisition electrodes (TP9,
AF7, AF8, and TP10). All the Muse-2 functions can
be managed through an application released for the
iOS 12.2 and Android 8 or higher operating systems
but a huge number of applications and libraries are
also available on the Internet aimed to extend its fea-
tures, allowing the researchers to create customized
applications using libraries designed for several pro-
gramming languages. Some representative examples
are: Brainflow (https://brainflow.org), a Python li-
brary that offers API able to filter, parse, and ana-
lyze the EEG data; Brains@play(https://brainsatplay.
com), an open-source framework that allows the de-
velopers to create brain-responsive applications based
on the web technologies; Muse LSL (https://github.
com/alexandrebarachant/muse-lsl), a Python package
that provides functions for streaming, visualizing, and
recording the EEG data. Among those available, in
this work we have chosen to use Muse LSL, which op-
erates under the Linux operating system, allowing us
to perform a Bluetooth Low Energy (BLE) commu-
nication between the computer and the Muse device
(compatible with Muse 1, Muse 2, and Muse S), pro-
viding functions for managing and testing the device
connection, as well as a Python library (working with
both Python versions 2.7 and 3.x) for developing the
code. It should be added that the Lab Streaming Layer
(LSL) adopted by Muse LSL is a mechanism widely
used for the unified collection of time-series measure-
ments in the scientific field since it allows us the syn-
chronization of the streaming data for real-time anal-
ysis or recording. Based on the results of other stud-
Table 1: Low-cost EEG Devices.
Brand Product Resolution Electrodes Reference
name name bits number site
Emotiv Insight 14 05 https://www.emotiv.com
Emotiv Epoch X 14/16 14 https://www.emotiv.com
Emotiv Epoch Flex 14 32 https://www.emotiv.com
InteraXon Muse-2 12 04 https://choosemuse.com
InteraXon Muse S 12 04 https://choosemuse.com
Neurowsky MindWave Mobile 2 12 01 http://neurosky.com
OpenBCI Cyton Biosensing Board 24 08 https://openbci.com
ies in this regard, the idea behind this work is related
to the exploitation of external stimuli to induce reli-
able/stable EEG patterns, with the aim to exploit them
as a biometric approach for user authentication. Dif-
ferently from the solutions in the literature, which are
usually unsuitable for large-scale use in the real-world
scenarios due to a series of limitations mainly related
to the need for expensive and/or complex EEG data
acquisition and stimulation techniques, the proposed
system is designed to use simple and low-cost hard-
ware (i.e., one of those in Table 1) and stimulation
techniques (i.e., computer sounds and images).
2 BACKGROUND AND RELATED
WORK
The human brain is made up of billions of neurons,
each of which can potentially connect with thousands
of other neurons in order to establish communication
channels through low intensity electric voltages (in
the order of microvolts). This type of electrical ac-
tivity of neurons generates brain waves, which on the
basis of their frequency (in the range from 4 to 100
Hz) are classified into five categories, each of them
denoted by a Greek letter: the Delta wave less than 4
Hz, the Theta wave from 4 to 8 Hz, the Alpha wave
from 8 to 12 Hz, the Beta wave from 13 to 30 Hz, and
the Gamma wave greater than 30 Hz. Each brain area
is then characterized by different wave frequencies,
which can be generated simultaneously. Several stud-
ies demonstrated the existence of a strong correlation
between brain wave rhythms and brain state such as,
for instance, between the fastest rhythms and the brain
processing of information, and between the slowest
rhythms and the inactive brain state (Sterman, 1996).
Nowadays, a growing number of literature works
are focused on the exploitation of EEG signals as a
biometric approach for the users identification (Yang
and Deravi, 2017), proposing approaches/strategies
able of producing unique and stable patterns over
time. In this context, it is possible to find works aimed
at formalizing these identification systems in a practi-
cal way (Thomas and Vinod, 2018), along with works
that instead deal with issues and aspects connected to
this research area (Fraschini et al., 2019). It should
be noted that to improve the stability and reliability
of the EEG patterns, an increasing number of works
adopt external stimuli of different types, although in
many cases these works do not propose a practical for-
malization of these systems usable in real-world con-
texts. This usually depends on a series of limitations
such as, for example, the long data EEG acquisition
times (Jayarathne et al., 2017).
A significant example of the above scenario is this
work (Campisi and La Rocca, 2014), where the au-
thors analyzed the brain activity for the automatic user
recognition purpose, as well as in this work (Soni
et al., 2016), where the authors instead propose a sys-
tem that allows users to set a pattern of brain waves
to perform the same task, combining eye blink, at-
tention, and the Alpha, Beta, Theta, and Delta brain
rhythms. Another significant work (Nakanishi and
Hattori, 2017) exploits the EEG activity evoked by
Brain Waves and Evoked Potentials as Biometric User Identification Strategy: An Affordable Low-cost Approach
615
Figure 1: 10-20 System of Electrode Placement.
invisible visual stimulation as biometric approach,
whereas this work (Reshmi et al., 2016) proposes an
approach to brain biometric user recognition, and in
this one (Abed and Abed, 2020) the authors formal-
ize an authentication system based on the features of
two brain waves, Gamma and Beta.
Evoked Potentials: An Evoked Potential (EP) (Walsh
et al., 2005) is an electrical potential that is mea-
sured in an area of the nervous system, such as the
brain, as a result of an external stimulus. These stim-
uli can be of different types, the most commonly
used are the Auditory Evoked Potentials (AEPs) (Seha
and Hatzinakos, 2018) (acoustical stimuli such as
a short sound) and the Visual Evoked Potentials
(VEPs) (Zhao et al., 2021) (visual stimuli such as a
light flash). The EPs are widely used to evaluate the
electrical activity related to specific areas of the brain
or the spinal cord, this in order to diagnose neurolog-
ical problems, and in the case of the brain that magni-
tude is in the order of a few microvolts at most.
An example of these approaches is the Inter-
mittent Photic Stimulation (IPS) (Coull and Pedley,
1978), where a pair of glasses capable of emitting in-
termittent light are used, or devices capable of gener-
ate sound stimuli with different frequencies, such as
in this study (Di et al., 2018), where the authors exper-
imented the effect of intermittent pure tones at 50 and
6 phon (i.e., a logarithmic unit of loudness level for
tones and complex sounds), using as frequency 125,
250, 500, 1000, and 4000 Hertz, and a duration of 10
seconds, demonstrating the relationships between the
acoustic properties of stimuli and the EEG activity.
Other studies in the literature instead ex-
perimented the effect of the so-called binaural
sounds (Colburn and Durlach, 1978), which are based
on the brain perception of interaural differences dur-
ing binaural stimulus (Blauert, 2013). In more detail,
this technique inducts the brain to interpret as a single
tone two different tones applied on the left and right
ears (formally defined carrier-tone and offset-tone),
where the detected single tone is given by the differ-
ence between the frequencies of these two ones. An
interesting literature work (Rajan et al., 2018) eval-
uates the impact of binaural sounds in terms of their
positive and negative impact in areas such as health-
care, security, education, and entertainment.
A recent work in literature (Zhao et al., 2021)
compares the performance of three different VEP sig-
nals in the context of a VEP-based biometric user
identification system, whereas another work (Rosli
et al., 2021) takes into account the use of EEG data
with visual stimuli for the same purpose, using the
Wavelet Packet Decomposition (WPD) technique in
the feature extraction process. With regard to the bio-
metric user identification systems based on EEG data
and auditory stimuli, in this work (Mukai and Nakan-
ishi, 2020), the authors use ultrasound stimuli for the
EEG stimulation, with the aim to avoid that the users
can be distracted from their current activity during
the identification process. For the sake of complete-
ness, it should be added that although visual and audi-
tory stimuli are the most commonly used in literature,
some works use other type of stimuli such as, for in-
stance, those based on vibrations (Nakashima et al.,
2021).
However, it should be added that all the aforemen-
tioned approaches, as well as most of the others in
the literature, do not have characteristics of perfor-
mance, simplicity, and cost that make them suitable
for widespread use as a biometric system.
Open Problems: Regardless of the technique/strat-
egy used to define a biometric system based on EEG
data, there are some well-known problems that re-
duce its practical feasibility, such as: the Data Com-
plexity, which is given by the fact that the EEG data
are complex, non-linear, and non-stationary (Mahato
and Paul, 2019); the Data Heterogeneity, which de-
pends on the fact that considerable differences exist
in the EEG data related to different users (Jausovec,
1997); the Data Calibration, which is related to the
need of a system calibration before the EEG data ac-
quisition (Jayarathne et al., 2017); the Data Diversity,
which depends on the difficulty of obtaining the same
EEG patterns over time, due to multiple factors that
influence users (Kaur et al., 2017).
Evaluation Metrics: The evaluation metrics largely
adopted in the literature to evaluate the performance
of the biometric systems based on the EEG data are
mainly based on two rates, the False Acceptance Rate
(FAR) and the False Rejection Rate (FRR) (Dahel and
Xiao, 2003), which express, respectively, how many
times a user is erroneously allowed access, and how
many times a legitimate user is erroneously denied
SECRYPT 2022 - 19th International Conference on Security and Cryptography
616
access. Another widely used metric based on the
aforementioned two ones is the Half Total Error Rate
(HTER), which is calculated as HT ER =
(FAR+FRR)
2
,
whereas the point at which we have the intersection
of the FAR and FRR values is named Equal Error
Rate (EER). The percentage of users correctly identi-
fied over their total number is the Correct Recognition
Rate (CRR) metric is calculated as TAR=1-FRR.
Research Motivation: This work stems from the ob-
servation that, unlike other biometric user identifica-
tion approaches (e.g., facial recognition, fingerprint
recognition, retinal recognition, etc.), those based on
EEG data are characterized in literature by a more the-
oretical than practical formalization, due to some in-
trinsic limitations. The most important of these are
certainly the difficulty of obtaining EEG patterns that
well characterize the user and that are stable over
time, as well as a series of problems related to the
preliminary configuration of the biometric system and
the related acquisition time. In light of this obser-
vation, this work is aimed at formalizing a low-cost
biometric identification system based on EEG data
detected during an external stimulation of the users,
which is practically usable in the real-world scenar-
ios, although with some limitations, in order to pro-
vide a starting point for subsequent improvements.
3 APPROACH FORMALIZATION
The idea behind the proposed approach is mainly
based on the differential analysis of EEG data
recorded before and after the application on the sub-
ject of a series of stimuli: practically, we capture
the brain waves activity in a defined time-frame, op-
erating without the external stimuli for the half of
time, and with them in the remaining time. We take
into consideration all the brain waves types, read
data from the Muse LSL stream, compute the av-
erage power of a signal in each specific frequency
range in order to reduce the noise, according to the
most widely used approach for the EEG data analy-
sis, where the sensors data is decomposed into dis-
tinct frequency bands, i.e., δ = [0.5, 4] Hz, θ = [4, 8]
Hz, α = [8, 12] Hz, β = [12, 30] Hz, and γ = [30, 100]
Hz. The band power is calculated computing the one-
dimensional Discrete Fourier Transform (DFT) using
the Fast Fourier Transform (FFT) algorithm, which
returns, for each frequency, a complex number that
allows us to extract amplitude and phase of the signal
at the desired frequency. The experimental environ-
ment in terms of the software and hardware will be
composed by the elements reported in Table 2.
The steps that compose the proposed approach,
Table 2: Experimental Environment.
Type Version Details Reference
Experimental Intel CPUs 3.40 GHz×8
Workstation i7-6700 RAM 16 GB
EEG Muse Version 2
device Headband https://choosemuse.com
Operating Linux 64 bit https://www.debian.org
system Debian 11 Kernel 5.10.0-9 https://www.kernel.org
Programming Python Version 3.9.2 https://www.python.org
language/IDE Eclipse PyDev Version 9.2.0 https://www.pydev.org
Python AccelBrainBeat Version 1.0.5 pypi.org/project/AccelBrainBeat
libraries for Scikit-learn Version 1.0.1 https://scikit-learn.org
EEG and eeglib Version 0.4.1 https://github.com/Xiul109/eeglib
external Muse LSL Version 2.1.0 https://github.com/
stimuli alexandrebarachant/muse-lsl
whose high-level architecture is shown in Figure 2,
are reported in the following:
- Acquisition System Calibration: verification of the
correct positioning of the Muse-2 device after the
user wore the headband; this is performed by mea-
suring the signal activity on each headband elec-
trode using the Muse LSL library functionalities;
- Unstimulated EEG Data Collection: collection of
the EEG data for half of a experimentally defined
time-frame, without applying any external stimulus
during the data acquisition; in this phase the signals
relating to all the brain waves will be acquired, the
most characterized of which will be selected later;
- Stimulated EEG Data Collection: collection of the
EEG data for the remaining period of the time-
frame, applying a series of external stimuli; the na-
ture and type of stimuli will be defined experimen-
tally using both visual (VEP) and auditory (AEP)
stimuli at different intensities and frequencies;
- Data Transform and Analysis: elaboration of the
collected data through the FFT algorithm, prepro-
cessing techniques, and differential analysis of the
EEG patterns; the EEG data collected with and
without the external stimuli are compared in order
to extract information that characterizes the way in
which the user responds to external stimuli;
- Data Comparing and Classification: the obtained
patterns are compared to the database of existing
ones with the aim to identify the users; a tolerance
margin is experimentally defined in this phase.
Future Direction: According to the experimental
steps previously mentioned, we want to investigate
the influence on the EEG signal of simple stimuli gen-
erated through flashing of the computer screen and
environmental sounds. In other words, differently
from the work in the literature we want to explore the
feasibility of a system realized using only low-cost
EEG device and simple stimuli that do not require ad-
ditional hardware.
In this regard, although this work proposes only a
theoretical formalization, a series of preliminary ex-
periments was carried out on a small number of users,
the results of which indicate the feasibility of the idea
Brain Waves and Evoked Potentials as Biometric User Identification Strategy: An Affordable Low-cost Approach
617
Figure 2: High-level Approach Architecture.
behind the proposed system, showing a greater char-
acterization of the user and a better stability over time
of the EEG patterns obtained through a comparative
analysis of the EEG data collected before and after
the application of an external stimulus (flashing of
the computer screen), then without using dedicated
hardware for its generation. These results must sub-
sequently be verified and validated in depth using a
significant number of users, as well as concerning
different stimul and data analysis and classification
techniques. Besides defining a low-cost affordable
EEG-based biometric approach to user identification,
our work will be facing the open issues previously
described by experimenting with techniques already
used in some previous works, such as, the feature
engineering (Carta et al., 2020; Heaton, 2016), the
feature space transformation (Saia et al., 2020), the
training data decomposition (Saia et al., 2021), and
the data enrichment/discretization (Saia et al., 2019a;
Saia et al., 2019b). An automatic data calibration al-
gorithm will be also developed to reduce the time of
the initial process of data recognition, in the context
of the Muse-2 device functionalities will be exploited.
The data diversity will be instead faced by the adop-
tion of differential data patterns given by the compar-
ison of the data collected before and after the applica-
tion of external stimuli.
4 CONCLUSIONS
This work aimed to verify the existence of conditions
that justified subsequent experimentation for practi-
cally formalizing a biometric user identification sys-
tem based on EEG data acquired under the effect of
external visual stimulation, to face some well-known
problems related to the difficulty to obtain stable char-
acterizing EEG patterns. The performed study of
the current literature and the results obtained in our
preliminary experimentation indicate the existence of
such conditions, albeit with limitations whose impact
will be quantified through future experiments.
Unlike most of the works in the literature, this
work proposes an affordable biometric system based
on low-cost hardware for the detection of EEG sig-
nals, and on an approach of generating external stim-
uli based on existing devices (i.e., the computer
screen), therefore without the need to use dedicated
devices for this purpose (e.g., flash-emitting glasses).
This combination was designed to allow us a wide
use of the system that will be used in many real-
world contexts, perhaps combined with other differ-
ent widespread biometric approaches in order to face
the EEG limitations, improving the reliability of the
biometric system.
ACKNOWLEDGEMENTS
This research is partially funded and supported by:
project “Studio per l’adeguamento di aree portale
per tematismo - BRIC INAIL 2019 - FENU” CUP
F24G20000100001”; “PON R&I 2014-2020 Action
IV.6 - CUP F25F21002270003”.
REFERENCES
Abed, S. S. and Abed, Z. F. (2020). User authentica-
tion system based specified brain waves. Journal of
Discrete Mathematical Sciences and Cryptography,
23(5):1021–1024.
Blauert, J. (2013). The technology of binaural listening.
Springer.
Campisi, P. and La Rocca, D. (2014). Brain waves for
automatic biometric-based user recognition. IEEE
transactions on information forensics and security,
9(5):782–800.
Carta, S., Podda, A. S., Recupero, D. R., and Saia, R.
(2020). A local feature engineering strategy to im-
prove network anomaly detection. Future Internet,
12(10):177.
Colburn, H. S. and Durlach, N. I. (1978). Models of binau-
ral interaction. Handbook of perception, 4:467–518.
Coull, B. M. and Pedley, T. A. (1978). Intermittent photic
stimulation. clinical usefulness of non-convulsive re-
sponses. Electroencephalography and Clinical Neu-
rophysiology, 44(3):353–363.
Dahel, S. and Xiao, Q. (2003). Accuracy performance anal-
ysis of multimodal biometrics. In IEEE Systems, Man
and Cybernetics SocietyInformation Assurance Work-
shop, 2003., pages 170–173. IEEE.
Di, G.-Q., Fan, M.-C., and Lin, Q.-H. (2018). An experi-
mental study on eeg characteristics induced by inter-
mittent pure tone stimuli at different frequencies. Ap-
plied Acoustics, 141:46–53.
Fraschini, M., Pani, S. M., Didaci, L., and Marcialis, G. L.
(2019). Robustness of functional connectivity met-
rics for eeg-based personal identification over task-
SECRYPT 2022 - 19th International Conference on Security and Cryptography
618
induced intra-class and inter-class variations. Pattern
Recognition Letters, 125:49–54.
Heaton, J. (2016). An empirical analysis of feature en-
gineering for predictive modeling. In SoutheastCon
2016, pages 1–6. IEEE.
Homan, R. W., Herman, J., and Purdy, P. (1987). Cere-
bral location of international 10–20 system electrode
placement. Electroencephalography and clinical neu-
rophysiology, 66(4):376–382.
Jausovec, N. (1997). Differences in eeg activity during the
solution of closed and open problems. Creativity Re-
search Journal, 10(4):317–324.
Jayarathne, I., Cohen, M., and Amarakeerthi, S. (2017).
Survey of eeg-based biometric authentication. In 2017
IEEE 8th International Conference on Awareness Sci-
ence and Technology (iCAST), pages 324–329. IEEE.
Kaur, B., Kumar, P., Roy, P. P., and Singh, D. (2017). Im-
pact of ageing on eeg based biometric systems. In
2017 4th IAPR Asian Conference on Pattern Recogni-
tion (ACPR), pages 459–464. IEEE.
Mahato, S. and Paul, S. (2019). Detection of major depres-
sive disorder using linear and non-linear features from
eeg signals. Microsystem Technologies, 25(3):1065–
1076.
Mukai, K. and Nakanishi, I. (2020). Introduction of frac-
tal dimension feature and reduction of calculation
amount in person authentication using evoked eeg by
ultrasound. In 2020 IEEE REGION 10 CONFER-
ENCE (TENCON), pages 567–572. IEEE.
Nakanishi, I. and Hattori, M. (2017). Biometric potential
of brain waves evoked by invisible visual stimulation.
In 2017 International Conference on Biometrics and
Kansei Engineering (ICBAKE), pages 94–99. IEEE.
Nakanishi, I. and Maruoka, T. (2019). Biometric authenti-
cation using evoked potentials stimulated by personal
ultrasound. In 2019 42nd International Conference
on Telecommunications and Signal Processing (TSP),
pages 365–368. IEEE.
Nakashima, H., Shindo, Y., and Nakanishi, I. (2021).
Performance improvement in user verification using
evoked electroencephalogram by imperceptible vibra-
tion stimuli. In 2021 20th International Sympo-
sium on Communications and Information Technolo-
gies (ISCIT), pages 109–113. IEEE.
Prathibha, R., Swetha, L., and Shobha, K. (2017). Brain
computer interface: Design and development of a
smart robotic gripper for a prosthesis environment.
In 2017 International conference on networks &
advances in computational technologies (NetACT),
pages 278–283. IEEE.
Rajan, A., Hashim, A., Akre, V., Walid, H., Nassiri, N., and
Ahmed, M. (2018). The impacts of binaural beats. In
2018 Fifth HCT Information Technology Trends (ITT),
pages 353–357. IEEE.
Reshmi, K., Muhammed, P. I., Priya, V., and Akhila, V.
(2016). A novel approach to brain biometric user
recognition. Procedia Technology, 25:240–247.
Rosli, F. A., Saidatul, A., Abdullah, A. A., and Navea, R. F.
(2021). The wavelet packet decomposition features
applied in eeg based authentication system. In Jour-
nal of Physics: Conference Series, volume 1997, page
012035. IOP Publishing.
Saia, R., Carta, S., Recupero, D. R., and Fenu, G. (2020).
A feature space transformation to intrusion detection
systems. In KDIR, pages 137–144.
Saia, R., Carta, S., Recupero, D. R., Fenu, G., and Saia, M.
(2019a). A discretized enriched technique to enhance
machine learning performance in credit scoring. In
KDIR, pages 202–213.
Saia, R., Carta, S., Recupero, D. R., Fenu, G., and Stan-
ciu, M. (2019b). A discretized extended feature space
(defs) model to improve the anomaly detection per-
formance in network intrusion detection systems. In
KDIR, pages 322–329.
Saia, R., Podda, A. S., Fenu, G., and Balia, R. (2021).
Decomposing training data to improve network intru-
sion detection performance. In KDIR, pages 241–248.
SCITEPRESS.
Seha, S. N. A. and Hatzinakos, D. (2018). Human recogni-
tion using transient auditory evoked potentials: a pre-
liminary study. IET Biometrics, 7(3):242–250.
Soni, Y. S., Somani, S., and Shete, V. (2016). Biometric
user authentication using brain waves. In 2016 Inter-
national Conference on Inventive Computation Tech-
nologies (ICICT), volume 2, pages 1–6. IEEE.
Sterman, M. B. (1996). Physiological origins and functional
correlates of eeg rhythmic activities: implications
for self-regulation. Biofeedback and Self-regulation,
21(1):3–33.
Thakor, N. V. and Sherman, D. L. (2013). Eeg signal pro-
cessing: Theory and applications. In Neural Engineer-
ing, pages 259–303. Springer.
Thomas, K. P. and Vinod, A. P. (2018). Eeg-based bio-
metric authentication using gamma band power dur-
ing rest state. Circuits, Systems, and Signal Process-
ing, 37(1):277–289.
Walsh, P., Kane, N., and Butler, S. (2005). The clinical
role of evoked potentials. Journal of neurology, neu-
rosurgery & psychiatry, 76(suppl 2):ii16–ii22.
Xavier, G., Ting, A. S., and Fauzan, N. (2021). P-eg002. an
exploratory study of brain waves and corresponding
brain regions of fatigue post-call doctors using quan-
titative electroencephalogram. Clinical Neurophysiol-
ogy, 132(8):e78.
Yang, S. and Deravi, F. (2017). On the usability of
electroencephalographic signals for biometric recog-
nition: A survey. IEEE Transactions on Human-
Machine Systems, 47(6):958–969.
Zhao, H., Chen, Y., Pei, W., Chen, H., and Wang, Y. (2021).
Towards online applications of eeg biometrics using
visual evoked potentials. Expert Systems with Appli-
cations, 177:114961.
Brain Waves and Evoked Potentials as Biometric User Identification Strategy: An Affordable Low-cost Approach
619