Content uploaded by Manuel Reis Carneiro
Author content
All content in this area was uploaded by Manuel Reis Carneiro on Mar 07, 2022
Content may be subject to copyright.
IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020 15107
Wearable and Comfortable e-Textile Headband
for Long-Term Acquisition of Forehead
EEG Signals
Manuel Reis Carneiro , Aníbal T. de Almeida,
Life Senior Member, IEEE
,
and Mahmoud Tavakoli ,
Member, IEEE
Abstract
—Electroencephalography (EEG) has a wide range
of applications in medical diagnosis, and novel form of Human
Machine Interfaces (HMI) for controlling prosthetic implants,
wheelchairs, and home appliances in various forms of paraly-
sis. However, the current EEG setups are composed of many
wires hanging down from the system, and individual elec-
trodes that must be set manually, which is time-consuming.
As a result, the overall system is neither comfortable, nor
aesthetically appealing. Here, we introduce for the first time,
a comfortable textile-basedEEG headband system that is soft,
conformal to the skin, and comfortable. We present materials
and methods for fabrication of multi-layer stretchablee-textile ,
that interfaces the human epidermis from one side through
printed electrodes, and interfaces a rigid PCB island on the second layer. We as well demonstrate a method that allows
creation of VIAs (vertical interconnect access) between the layers, using a CO2 laser. All Electrodes are integrated into
the headband, and thus there is no need for individual electrode placement, and individual wiring. By screen printing
a home-made conductive stretchable ink, patient-specific EEG headbands can be tailor made considering the optimal
positioning of the electrodes for each patient. We show that these printed electrodesbenefit from a very low skin-electrode
impedance, comparable to gold standard Ag/AgCl, or gold cup electrodes, thanks to the high surface area silver flakes
used in this work. The e-textile headband interfaceswith anEEG acquisitiondevice that captures, amplifies, and transmits
the data to an external mobile phone or a PC. Furthermore, the integrated amplification in the textile and the use of an
EMF rejection layer on top of the electrodes were shown to reduce the unwanted EM noise that is picked up by the
system. We as well show application of the developed headband for usage in Human Machine Interfaces and Sleep Data
Acquisition. Altogether, this device is step toward wider use of EEG acquisition devices for daily-use applications.
Index Terms
—Conformable Electronics; E-textile, wearable sensor, biomonitoring, electroencephalography, epidermal
electronics.
I. INTRODUCTION
IN SPITE of the growing availability of neuroimaging
techniques (magnetic resonance imaging - MRI, com-
Manuscript received June 10, 2020; revised July 13, 2020; accepted
July 13, 2020. Date of publication July 16, 2020; date of current
version November 18, 2020. This work was supported in part by
the Foundation of Science and Technology (FCT) of Portugal through
the Carnegie Mellon University (CMU)-Portugal project Wireless biO-
monitoring stickers and smart bed architecture: toWards unthered
patients (WoW) under Grant 45913, in part by the Dermotronics under
Grant PTDC/EEIROB/31784/2017, in part by the European Union (EU)
Structural and Investment Funds (FEEI) through operational program
of the center region, in part by the Materiais e Tecnologias Indus-
triais Sustentáveis (MATIS) under Grant CENTRO-01-0145-FEDER-
000014, in part by CENTRO2020, in part by Add.Additive under
Grant POCI-01-0247-FEDER-024533, and in part by the Regional
Development Funds (FEDER), through Programa Operacional Com-
petitividade e Internacionalização (POCI). The associate editor coor-
dinating the review of this article and approving it for publication was
Dr. Edward Sazonov.
(Corresponding author: Mahmoud Tavakoli.)
The authors are with the Soft and Printed Microelectronics Lab, Institute
of Systems and Robotics, University of Coimbra, 3004-531 Coimbra,
Portugal (e-mail: mahmoud@isr.uc.pt).
This article has supplementary downloadable material available at
https://ieeexplore.ieee.org, provided by the authors.
Digital Object Identifier 10.1109/JSEN.2020.3009629
puted tomography - CT, single-photon emission tomogra-
phy - SPECT and positron emission tomography - PET),
electroencephalography (EEG) is still the most commonly
used sensing method for brain signal acquisition, which can
be used for detection of epileptic activity [1], [2] and non-
convulsive status epilepticus (NCSE) [3]. It also has a key
role in the investigation and research on encephalitis [4],
head trauma [5], coma [6], stroke [7] and brain death [8]
and other neurological conditions [9]–[11]. In addition to the
medical diagnosis, EEG signals have received an increasing
attention as an HMI for control of implanted prosthetics, and
wheelchairs [12]–[16], [17] and for study of emotions and
sensorial perception [18]–[20].
Despite the advances in the signal processing and clas-
sification techniques that allow real-time EEG based con-
trol of prosthetics and wheelchairs, one limiting factor for
wider application of the EEG-based HMIs is the current data
acquisition setups. Today’s EEG data acquisition systems are
bulky, have many wires, and require time-consuming prepa-
ration. They are not comfortable, nor aesthetically appealing.
In contrast to biological organs that are made of soft tissues,
1558-1748 © 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
15108 IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020
current electrophysiological monitoring devices are made of
rigid materials. The mechanical mismatch between the current
rigid electronic systems, and human tissues, is one of the
main driving forces for the field of stretchable electronics.
The main objective in this field is to develop soft and
stretchable e-skins that are able to adapt to the dynamic
morphology of the human skin [21], [22], without losing their
functionality. In the last decade, materials and methods for
the fabrication of stretchable circuits have been developed and
improved, including mask deposition [23], [24], laser pattern-
ing [25]–[27], and printing techniques [28]–[36]. Some efforts
were made in order to develop soft and stretchable e-skin
devices for conformal and comfortable electrophysiological
monitoring [29], [37]–[47]. For most applications in bio-
electronics, such circuits need to be epidermis-conformable,
i.e. they must keep their functionality while being bent, folded,
twisted and when subject to strains up to 30%, the normal
stretchability of the human skin [29], [30]. Among them, few
works presented soft electronic skins for electrophysiological
monitoring and EEG sensing [48], [49]. While promissing
as a disposable system, these ultrathin polymer-based archi-
tectures are generally fragile for long-term utilization. As a
consequence, the state of the art EEG technology can’t be
widely used for long-term research and daily applications [50]
and its utilization is mostly limited to experiments inside
laboratories. On the other hand, textile, as the common and
widely accepted “wearable” for human being, has an out-
standing potential to host electronics circuits for wearable
bio-electronics [51], [37], [52], [53]. In contrast to polymeric
e-skins, textile can be cut and easily tailor-made to the desired
dimensions.
In this article, we propose a plug-and-play wearable e-textile
(Fig. 1), that integrates multiple flexible electrodes and all
the necessary electrical interconnects, which can easily and
inexpensively be adapted to the specific needs of each user
in terms of electrodes quantity and positioning. By stencil
printing of a novel stretchable conductive ink developed by the
authors, and laser patterning of electrically conductive vias,
we implemented a two layer e-textile architecture, in which
electrodes and interconnects are on two opposite sides of the
textile. The conductive stretchable ink was synthesized by
mixing a high surface area Ag flake s(Technic inc), and a
Styrene-isoprene block copolymers (SIS) (Aldrich Chemistry).
In order to verify the suitability of the printed electrodes
for recording of electrophysiological data, we analyzed the
impedance of the skin-electrode interface over a frequency
range of 0-100kHz, and compared it with the gold-standard
Ag/AgCl and Goldcup electrodes. Results showed that the
printed electrodes using the mixture of high surface area
Ag flake sand (SIS) have an impedance in the same order
of gold standard electrodes, and extremely lower than the
composite made by a blend from the same polymer and
Carbon particles. Moreover, a modular electronic hardware
architecture was designed and implemented. This architecture
allows acquisition, amplification, processing and real-time
transmission of 8 EEG channels. Different setups of the
system in terms of the distance between the amplification
module and the electrodes were compared and the effect of a
Fig. 1. Textile EEG acquisition headband proposed in this work. Top
detail: Connection between rigid PCB and textile printed conductive lines
by means of a ink drop joint. Right detail: Contact between the printed
electrodes and the skin in the forehead.
EMF rejection layer on the acquired signal was also studied.
We show that integration of the amplification circuit near
the electrodes, as well as the implemented Faraday Cage, lead
to a decrease in the noise picked up by the system. Finally,
as a case study, examples of application of the developed
e-textile architecture are demonstrated as an HMI, and as
well for a comfortable data recording system during the
sleep.
II. MATERIALS AND METHODS
The wearable EEG system is composed of two main compo-
nents. The skin interfacing e-textile and the electronics system,
required for signal acquisition, amplification and communica-
tion. It is capable of recording electrical brain activity and
streaming it wirelessly for further processing.
A. E-Textile Headband
A conductive paste was developed by mixing 0.62wt% Ag
flakes (Technic inc), with a 0,17wt% SIS (Sigma-Aldrich)
solution. SIS is first dissolved in Toluene in a magnetic
mixer and heater (Agimatic-N). Ag flakes are then added to
the solution and mixed using a planetary mixer (THINKY
ARE-250) (3 minutes, 2000rpm). In contrast to common
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
CARNEIRO
et al.
: WEARABLE AND COMFORTABLE e-TEXTILE HEADBAND FOR LONG-TERM ACQUISITION OF FOREHEAD EEG SIGNALS 15109
Fig. 2. Schematic of the patch fabrication process. A:Thelatex
impregnated fabric is cured on top of a glass in order to cure uniformly.
B: The vias on the fabric are cut using a Co2 laser. C: A previously cut
stencil is placed on the fabric and the ink is spread using a sharp blade or
spatula. D: The stencil is carefully removed, and the printed electrodes
are let do dry. E: The fabric is flipped over and the process is repeated
to create the circuit interconnects.
Fig. 3. Schematic of the textile patch and respective dimensions, with the
position of each printed electrode in terms of the international 10-20 EEG
electrode positioning system.
Ag based conductive inks and pastes that demand for a
post-baking (thermal sintering) procedure, the resulting com-
posite is surfactant-free and does not require a baking step. The
whole printing process is performed at the room temperature,
which is especially relevant and important for textile-based
electronic, as commercial Ag based pastes for screen printing,
require a thermal sintering step (100-250◦C), and therefore are
not compatible with heat sensitive substrates, including most
textiles. Fig. 2 summarizes the process for fabrication of the
EEG headband.
First, a lycra-mesh fabric is impregnated in latex (PVS
Elastica, FORMX). Then using a CO2 laser machine (Uni-
versal Laser Systems VLS 3.50), electrical vias are cut on the
fabric (Fig. 2B). Then a stencil (LineafixAdhesive, Lineafix
Hogar) is placed over the fabric and is patterned by the same
CO2 laser to desired pattern (size, geometry and position) of
the electrodes. The conductive ink is screen printed over the
fabric (Fig. 2C). While printing the first layer of the circuit,
the previously patterned VIAs are also filled. Then the stencil
is removed, and the fabric is turned over (Fig. 2D). Finally,
the process is repeated on the other side of fabric, this time
to print the electrical interconnect.
In Fig. 3 the shape of the textile patch, printed lines,
vias and printed electrodes is shown. Furthermore, each of
the 8 recording electrodes is identified by its chosen location
in terms of the 10-20 system (the internaltionaly accepted EEG
electrodes positioning system), alongside eith the reference
and ground electrodes.
In order to attach the rigid amplification PCB (described
later) to the printed lines, a simple joint is created by a drop
of conductive ink that bonds both to the line and the contacts
of the rigid PCB, as seen in Fig. 1 (top insertion).
B. Signal Amplification and Transmission
Fig. 4A illustrates the high level system block diagram
describing the architecture of the electronics system. The
electrodes patch is worn by the subject’s on his/her forehead
and the raw data is then amplified, digitized and transmited
through WiFi to a device connected to the same network for
processing.
The electronics system is composed of two main cir-
cuits. Amplification circuit, and processing/transmission cir-
cuit. Amplification circuits are placed directly on the e-textile,
to be near to the acquisition electrodes. Placing the amplifier
near the signal source is generally a good practice, since it
eliminates amplification of the noise acquired by cables in
a tethered system. As can be seen in Fig. 4B, the system
architecture is designed in order to allow integration of several
amplifiers that can be all wired to the main processing and
transmission board. In this way, one can customize the number
and location of amplifiers on the headband.
Brain electrical signals acquired by surface electrodes are of
very low amplitude (between 10 μV and 100 μV) [54]–[56]
and therefore, require high gain amplification prior to being
sent to the Analog-to-Digital converter. The ADS1299 module
(ADS1299, Texas Instruments) was selected as the signal
amplifier and ADC for the final design. This 8-channel
amplifier with an integrated 24-bit sigma-delta ADC enables
prolonged biopotential monitoring with sufficient precision.
The amplifier sits directly on the electrodes’ textile patch.
An ESP8266 microcontroller (Expressif Systems) was cho-
sen, due to its low power consumption and WiFi capabil-
ities. The communication between the ADS1299 and the
ESP8266 is via an SPI interface, which allows for a modular
design, where up to 3 amplification boards can be connected
to the same microcontroller, allowing for simultaneous reading
of up to 24 EEG signals
The system is powered by 2 li-Po cells, totalizing 7.4V and
1600mAh, allowing for 24 hours operation time.
III. RESULTS AND DISCUSSION
A. EEG Data Acquisition
The data was acquired by a working prototype (Fig. 4C)
of the proposed device. It was transferred to a computer and
processed in Matlab.
Processing includes removal of the DC frequency and
removal of any frequency above 100Hz, as well as removal of
power line interference at 50Hz.
B. Electrode-Skin Impedance Measurements
In order to analyse the suitability of the printed electrodes
for acquisition of electrophysiological signals, we analysed the
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
15110 IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020
Fig. 4. A: Block diagram illustrating the main components of the device, comprising data collection and data transmission blocks, power circuit
and data processing unit. B: The modularity of the system allows for several amplifiers to be connected simultaneously to the same processing and
transmission board. C: Working prototype of the EEG acquisition system proposed in this work.
impedance of the electrode-skin interface over a wide range
of frequencies and compared the results with gold standard
electrodes. Generally, it is accepted in the literature that a
lower skin-electrode impedance is desired, since it leads to a
higher signal to noise ratio [29].
A bioimpedance measurement system (PalmSens4, Fig. 5A)
connected to PSTrace Software was used to measure the
electrode-skin impedance in different frequencies (between
0.1Hz and 10000Hz), by using an alternating sinusoidal exci-
tation electrical current. Relying on the equivalent model
proposed by Albulbul [57], we are able to estimate each of the
electrode-skin impedance parameters. For the measurements,
3 electrodes [58] are applied in the ventral side of the subject’s
forearm and the excitation signal is applied between 2 of the
3 electrodes placed in the subject.
The ventral side of the forearm was selected to perform
the electrode-skin impedance measurements due to accessi-
bility for long-term measurement. A careful skin preparation
procedure was performed before every measurement. This
consists of cleaning the skin with alcohol swabs in order
to remove dead skin cells, sweat and oil. We then let the
skin to air-dry for 5 minutes and placed the electrodes over
the forearm. We then wait for 5 minutes before taking the
measurement. This period is considered to allow stabilization
of the temperature, and humidity. All tests were taken in
ambient humidity ranging from 52-70%. This preparation
procedure allowed to decrease the influence of external factors
(other than the electrode materials) in the taken measurements.
All the electrodes had the same displacement of 3cm
between them. The Ag/AgCl electrodes were fixed to the skin
by their adhesion, while the Goldcup and AgSIS ink electrodes
were fix by means of a textile (as seen in Fig. 5E) secured
to the forearm, maintaining the pressure constant in both
cases.
Different combinations of materials were tested and
compared– Ag/AgCl (Fig. 5C); Goldcup (with Ten20 con-
ductive paste) (Fig. 5D) Ten20; AgSIS ink (Fig. 5E)and
CSIS [59].
Using the Levenberg-Marquadt [60] fitting method, imple-
mented in PSTrace5 software, we are able to estimate the val-
ues of each component of the electro-skin interface equivalent
model in Fig. 5F, according to equation (1):
Ze=Rs+Rd
1+j2πfCdRd
(1)
where Rdand Cdrepresent the impedance associated with the
electrode-skin interface and the polarization at this interface,
while Rsis the series resistance associated with the type of
material of the electrodes.
Regarding the Electrode-Skin equivalent model, our team
has previously shown that the Rsvalue is not a relevant
parameter for signal to noise ratio measurement [29]. On the
other hand, lower Rdand higher Cddirectly contribute to a
better signal quality, since they lead directly to a lower value
of Ze, as shown by the previous equation. Parameters such
as the electrode “softness”, and conformable skin-electrode
interface were parameters found to be directly contributing to
a better electrode-skin interface [29].
From Fig. 5G, we see that the AgSIS printed electrodes
have a Rd value similar to the gold standard electrodes (Gold-
cup (with conductive paste) and Ag/AgCl). This is a very desir-
able result, since lower resistance paths to the ionic current
biological signals lead to higher quality measurements [29].
As a comparison, we can see that the Rd Value for CSIS
printed electrodes is much higher than the previously referred
electrodes. This might be due to the high aspect ratio of silver
flakes, that allow a better percolating network compared to the
spherical type conductors.
It is known, by previous works, that higher Cd values
are related to the ability of the electrodes to accumulate
charges at theelectrode-skin interface and are thus expected in
polarizable electrodes, which are likely to behave capacitively
at the electrode-skin interface. The highest values of this
parameter are observed in the Ag/AgCl and AgSIS electrodes,
followed by Goldcup (with Ten20 conductive paste) and CSIS
electrodes.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
CARNEIRO
et al.
: WEARABLE AND COMFORTABLE e-TEXTILE HEADBAND FOR LONG-TERM ACQUISITION OF FOREHEAD EEG SIGNALS 15111
Fig. 5. Electrode-skin impedance analysis and electrode characterization. A: experimental setup - two electrodes placed over the subject’s forearm
are connected to the impedance analyzer and a frequency sweep is performed. B: single electrode/skin interface simplification circuit for the system’s
equivalent model. C: Gold-standard commercially available pre-gelled Ag/AgCl and D: Goldcup electrodes. E: Novel conductive ink electrodes.
F: Bode plots of the electrode-skin impedance of each of the tested electrodes (Ag/AgCl, Goldcup – interfaced to the skin via a layer of Ten20
conductive paste, CSIS and AgSIS). G: Components of the equivalent electrode-skin model (Rd, Cd and Rs) for each of the tested electrode
materials.
We are able to infer, from the previously presented results,
that this AgSIS printed electrodes benefit from a low electrode-
skin impedance, comparable to the Gold standard electrode
types. This means that this conductive ink, contrary to what
is observed in other printable conductive inks (Carbon-SIS for
example), can be reliably used for biomonitoring applications,
as we demonstrate later in this work for the case of EEG
recordings.
Regarding the Electrode-Skin Impedance shown in the
Bode plots from Fig. 8A, we observe, as expected, that the
electrode-skin impedance of AgSIS electrodes are similar to
the values of the Goldcup (with Ten20 conductive paste) and
Ag/AgCl electrodes and much lower than the CSIS electrodes.
C. Signal-to-Noise Ratio Measurements
SNR is a good indicator of the quality of the acquired
data and thus a good indicator of the quality of the overall
system. Different setups of the system were implemented,
tested and compared, in an effort to reduce the Signal-to-noise
ratio (SNR) in similar EEG readings acquired by the proposed
system. The SNR value is calculated by computing the ratio of
the filtered signal (x) summed squared magnitude to that of the
noise (y), using the snr(x,y) matlab function. First, the ampli-
fier was connected to the patch through wires, like usually
seen in similar devices used in a clinical setting or research
Fig. 6. A: Connection through long wires between the amplification
board and the electrodes’ patch. B: Amplification connected directly in
the electrodes’ patch. C: Tin foil faraday cage covering the system.
labs (Fig. 6A). In another setup, the amplification board was
fixed directly in the fabric patch, near the electrodes (Fig. 6B).
In an effort to reduce further the noise levels in the system,
a Faraday cage was implemented, by covering the fabric patch
with tin foil, as it can be seen in Fig. 6C.
In Table I, we can see that the usage of a faraday cage
covering the exposed printed lines is effective in reducing
noise levels acquired by the system, increasing the SNR from
−39.8139dB to −36.0465dB.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
15112 IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020
TABLE I
RELATION BETWEEN THE SNR IN THE SYSTEM WHEN THE CIRCUIT
LINES ARE EXPOSED TO EM NOISE AND WHEN THE CIRCUIT
INTERCONNECTS ARE PROTECTED BY A TIN FOIL
FARADAY CAGE
TABLE II
COMPARISON OF THE SNR VALUES IN THE SYSTEM WHEN USING
LONG WIRES AND WHEN THE AMPLIFICATION BOARD I S DIRECTLY
CONNECTED ON THE TEXTI LE PATC H ,ON TOP OF THE
PRINTED ELECTRODES
Furthermore, as seen in Ta b le I I, there is an improvement
in the SNR value (from −35.5489dB to −30.1996dB), when
the amplification is integrated near the electrodes, compared
to the case of tethered headband.A parallel experiment was
conducted to compare the SNR between EEG data acquired
by goldstandard electrodes (Ag/AgCl and goldcup) and the
custom printed electrodes presented in this work. After careful
skin preparation, all 3 electrodes were attached around the FP1
location (10-20 electrode positioning system) so they would
acquire the same EEG signal and connected to the developed
EEG acquisition board through 1-meter wires, and EEG was
recorded simultaneously from the 3 electrodes.
8 independent EEG datasets were acquired, and the results
are presented in Table III. We can observe that, although
Ag/AgCl electrodes (SNR=−32.1874dB), followed by gold-
cup electrodes (SNR=−32.5452dB), perform slightly bet-
ter than AgSIS printed electrodes (SNR=−32.7113dB),
the results are comparable.
D. Mechanical Characterization
A latex-impregnated lycra dogbone (width 10mm, length
50mm, thickness 0.55m) with an AgSIS printed track
TABLE III
SNR COMPARISON BETWEEN EEG DATA ACQUIRED WITH
AG/AGCLELECTRODES,GOL DCUP ELECTRODES AND
AGSIS PRINTED ELECTRODES
Fig. 7. Variation of electrical resistance of an AgSIS track with tensile
strain.
(with 1.5mm, length 50mm) was stretched while measuring
the track’s electrical resistance. This test was performed in a
Instron 5943 and the results are presented in Fig. 7. We can
observe that between 0 and 20% tensile strain, the resistivity
doubles, which is not a relevant variation since the initial
resistance was 11.93and a normal everyday usage of the
developed headband would never subject the printed tracks
to strains bigger that 20%. We can also observe that at 32%
strain there is a ramp up in electrical resistance leading us to
conclude that that is the tensile strain limit for the developed
ink.
E. Comparison With Commercially Available Devices
Table IV (adapted from [61]), summarizes the characteris-
tics of current EEG acquisition headsets that are both available
in the market and presented in state-of-the-art literature.
The system developed in this work is low weight- only 115g
regarding the 24 channels version (including textile electrodes,
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
CARNEIRO
et al.
: WEARABLE AND COMFORTABLE e-TEXTILE HEADBAND FOR LONG-TERM ACQUISITION OF FOREHEAD EEG SIGNALS 15113
TABLE IV
COMPARISON BETWEEN FULL-HEAD (A,B,C,D,E,F,G,H,I,J)AND FOREHEAD-ONLY (K,L,M,N,O)EEG
HEADSETS AND THE PROPOSED EEG ACQUISITION HEADBAND IN TERMS OF DIFFERENT PARAMETERS
electronic circuits and batteries), and low cost - only
$225.33 BOM. The maximum operation time was found to be
24 hours and 42 minutes, more than any of the other presented
devices.
Regarding the used electrodes, in contrast with most of
the presented commercially available devices (which rely on
disposable, wet contact electrodes), the headband proposed in
this work relies on dry ink-based stretchable electrodes that are
comfortable and can be used for long periods. Furthermore,
this headband can accommodate more electrodes than any of
the other forehead EEG headbands presented and it is fully
customizable in terms of electrode positioning.
F. Case Studies
1) Human-Machine Interface:
Since the eyes act as electrical
dipoles, with the cornea being positive and the retina negative,
when the eyes move, changes in electrical potentials can be
observed [71]–[73]. Using the 2 outermost electrodes of the
fabric patch, and observing one of them relative to the other,
we can see a potential that changes as follows: When the eyes
move to the right, the potential gets positive and in contrast,
gets negative when the eyes move to the left. The results
from this simple application can be observed in Fig. 8 and
Multimedia extension 1.
2) Sleep Data Acquisition:
The system developed in this
work was used to acquire and record sleep data over a night
of sleep (Fig. 9D). Although the data was recorded but not
analyzed in depth, a simple visual inspection of the EEG data
plots allows us to see clear differences in brain activity along
Fig. 8. Human-Machine Interface by eye position detection. A: subject
looking left (blue box activated). B: subject looking right (red box acti-
vated). In each figure we can also see the measured potential, in the
white text box, in uV.
the night, as seen in Fig. 9. During deep sleep (Fig. 9 A,B,E)
slow waves are predominant and high frequency activity is
almost non-existing. For instance, in Fig. 9 B, one can easily
observe predominant 7Hz brain activity usually related to
stage 1 non-REM sleep in healthy subjects [74]Furthermore,
in Fig 9 E, it can be observed that there is no relevant
brain activity above 40 Hz and low frequency waves are
predominant (below 15Hz, with peaks around 2Hz and 9Hz),
which correspond to either sleep stage 3 (predominant Delta
waves in the 1-4Hz range) or a drowsy-like non-awake state
(predominant alpha waves in the 9,10Hz range)[74].
In contrast, we can observe in Fig. 9 C a greater range of
frequencies coexisting in the plot, as well as some artifacts
related to eye-blinks (−2s and −3.5s), which indicate that the
subject is awake [74]. The same happens in Fig. 9 F where
an even distribution of amplitudes between all frequencies can
be observed.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
15114 IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020
Fig. 9. Sleep data acquisition. A: subject wearing the proposed EEG acquisition headband during sleep. B,C: EEG activity during sleep. Low
frequency waves can be observed D: Normal awake EEG activity. E: Predominant frequencies in EEG during sleep. High frequencies in EEG during
sleep. High frequency activity is almost non-existing during sleep F: Existing frequencies in a normal awake EEG.
The portability and comfort of the system bring important
advantages for long-term recordings. It must be noted that the
tested subject reported no discomfort during sleep and showed
no signs of skin-damage or irritation after waking up.
IV. CONCLUSIONS
A comfortable and reusable e-textile with integrated flexible
and stretchable electrodes and amplification circuitry for EEG
biosignals’ acquisition was successfully developed in this
work. The main goal of creating this flexible, fabric-based
headband was to overcome the existing problems in state-of-
the-art EEG setups: The bulkiness of the systems, the dis-
comfort caused by rigid electrodes and the time consuming,
laborious task of placing individual electrodes in a subject,
using conductive pastes.
The developed textile patch with integrated electrodes is
easy to place in the subject, allowing the immediate placement
of multiple electrodes in the forehead in a matter of seconds,
without the need for time-consuming individual electrode
placement, skin abrasion, or complex skin preparation. Due to
its comfort, it can be used on a daily basis for Human Machine
Interfaces, continuous brain signal monitoring, or sleep-data
acquisition.
Regarding the screen printed electrodes, the Electrode-Skin
impedance was measured over a wide range of frequencies
(0.1Hz - 10000Hz) and it was found that these electrodes have
a behavior comparable to state of the art electrodes used in
medical environment (Ag/AgCl and Goldcup (with Ten20 con-
ductive 8paste) electrodes). Also, in terms of acquired noise,
the printed electrodes are comparable to Ag/AgCl and Gold-
cup electrodes. The fabrication technique is simple, fast and
is performed entirely at ambient conditions. Therefore, this
headband can be easily adapted to each patient in terms of
quantity and positioning of electrodes in the system.
The proposed architecture allows approximation of the
amplification circuit to the electrodes, which is an efficient
measure in reducing the noise, as longer wires in traditional
systems are a known cause of induced noise. The current
electronic circuit developed in this work, is a preliminary
proof-of-concept version and should be still analyzed and
improved in terms of internal contribution to the noise, which
the subject of future works.
It was also observed that the use of an electromagnetic
radiation rejection layer (Faraday cage) covering the patch
interconnects proved to be an effective way of further reducing
noise levels in the system.
The wireless capabilities of the headband make the proposed
EEG acquisition system ideal for outpatient EEG studies and
use in remote sites, since it is self-contained and there are no
wires holding the patient under study. Also, the operating time
of more than 24 hours is an outstanding advantage allowing
for long term brain activity recordings, alongside with the
advantages of the device being comfortable and low weight
(115g).
The before presented characteristics of the proposed EEG
acquisition system, is a major development in the field of
EEG acquisition hardware, that can have strong impact in the
way brain-activity studies and medical exams are performed,
allowing for long-term patient monitoring. Furthermore, this
work is a major step toward application of wearable EEGs in
HMI for control of appliances.
CONFLICT OF INTERESTS
The authors declare no conflict of interest.
REFERENCES
[1] M. Sundaram, R. M. Sadler, G. B. Young, and N. Pillay, “EEG in
epilepsy: Current perspectives,” Can. J. Neurol. Sci., vol. 26, no. 4,
pp. 255–262, Nov. 1999.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
CARNEIRO
et al.
: WEARABLE AND COMFORTABLE e-TEXTILE HEADBAND FOR LONG-TERM ACQUISITION OF FOREHEAD EEG SIGNALS 15115
[2] S. Noachtar and J. Rémi, “The role of EEG in epilepsy: A critical
review,” Epilepsy Behav., vol. 15, no. 1, pp. 22–33, May 2009, doi:
10.1016/j.yebeh.2009.02.035.
[3] B. J. Kolls and A. M. Husain, “Assessment of hairline EEG as a
screening tool for nonconvulsive status epilepticus,” Epilepsia, vol. 48,
no. 5, pp. 959–965, May 2007, doi: 10.1111/j.1528-1167.2007.01078.x.
[4] R. Sutter et al., “Electroencephalography for diagnosis and prognosis of
acute encephalitis,” Clin. Neurophysiol., vol. 126, no. 8, pp. 1524–1531,
Aug. 2015, doi: 10.1016/j.clinph.2014.11.006.
[5] J. N. Ianof and R. Anghinah, “Lesão cerebral traumática: Um ponto de
vista EEG,” Dementia Neuropsychologia, vol. 11, no. 1, pp. 3–5, 2017,
doi: 10.1590/1980-57642016dn11-010002.
[6] B. Young, “The EEG in coma,” J. Clin. Neurophysiol., vol. 17, no. 5,
pp. 473–485, 2000.
[7] K. G. Jordan, “Emergency EEG and continuous EEG monitoring in acute
ischemic stroke,” J. Clin. Neurophysiol., vol. 21, no. 5, pp. 341–352,
2004, doi: 10.1097/01.WNP.0000145005.59766.D2.
[8] Z. Chen et al., “An empirical EEG analysis in brain death diagnosis for
adults,” Cognit. Neurodyn., vol. 2, no. 3, pp. 257–271, Sep. 2008, doi:
10.1007/s11571-008-9047-z.
[9] S. F. Khan, R. Ashalatha, S. V. Thomas, and P. S. Sarma, “Emer-
gent EEG is helpful in neurology critical care practice,” Clin.
Neurophysiol., vol. 116, no. 10, pp. 2454–2459, Oct. 2005, doi:
10.1016/j.clinph.2005.06.024.
[10] J. Praline et al., “Emergent EEG in clinical practice,” Clin.
Neurophysiol., vol. 118, no. 10, pp. 2149–2155, Oct. 2007, doi:
10.1016/j.clinph.2007.07.003.
[11] P. N. Varelas. (2003). Emergent EEG Indications and Diagnostic Yield
Hours 2 or Any EEG Done to Exclude NCSE With 24-Hours-a-Day,
7-Days-a-Week. [Online]. Available: https://www.neurology.org
[12] A. Ferreira, W. C. Celeste, F. A. Cheein, T. F. Bastos-Filho, and
M. Sarcinelli-Filho, “Human-machine interfaces based on EMG and
EEG applied to robotic systems,” in Proc. 2nd Int. Workshop Biosignal
Process. Classification (BPC), 2006, pp. 116–125, doi: 10.1186/1743-
0003-5-10.
[13] J. Zhang, B. Wang, C. Zhang, Y. Xiao, and M. Y. Wang,
“An EEG/EMG/EOG-based multimodal human-machine interface to
real-time control of a soft robot hand,” Frontiers Neurorobotics, vol. 13,
p. 7, Mar. 2019, doi: 10.3389/fnbot.2019.00007.
[14] M. Tariq, P. M. Trivailo, and M. Simic, “EEG-based BCI control
schemes for lower-limb assistive-robots,” Frontiers Hum. Neurosci.,
vol. 12, p. 312, Aug. 2018, doi: 10.3389/fnhum.2018.00312.
[15] Z. T. Al-qaysi, B. B. Zaidan, A. A. Zaidan, and M. S. Suzani,
“A review of disability EEG based wheelchair control system: Coher-
ent taxonomy, open challenges and recommendations,” Comput. Meth-
ods Programs Biomed., vol. 164, pp. 221–237, Oct. 2018, doi:
10.1016/j.cmpb.2018.06.012.
[16] S. K. Swee, K. Desmond, T. Kiang, and L. Z. You, “EEG controlled
wheelchair,” in Proc. MATEC Web Conf., Apr. 3016, Art. no. 02011,
doi: 10.1051/matecconf/20165102011.
[17] F. Babiloni et al., “On the use of brain–computer interfaces outside sci-
entific laboratories: Toward an application in domotic environments,” Int.
Rev. Neurobiol., vol. 86, pp. 133–146, Jan. 2009, doi: 10.1016/S0074-
7742(09)86010-8.
[18] G. Cartocci et al., “Gender and age related effects while watching TV
advertisements: An EEG study,” Comput. Intell. Neurosci., vol. 2016,
pp. 1–10, Oct. 2016, doi: 10.1155/2016/3795325.
[19] A. G. Maglione et al., “A neuroelectrical brain imaging study on the
perception of figurative paintings against only their color or shape
contents,” Frontiers Hum. Neurosci., vol. 11, p. 378, Jul. 2017, doi:
10.3389/fnhum.2017.00378.
[20] G. Cartocci et al., “Electroencephalographic, heart rate, and galvanic
skin response assessment for an advertising perception study: Applica-
tion to antismoking public service announcements,” J. Vis. E xp. , no. 126,
Aug. 2017, Art. no. e55872, doi: 10.3791/55872.
[21] A. Nag, S. C. Mukhopadhyay, and J. Kosel, “Wearable flexible sensors:
Areview,”IEEE Sensors J., vol. 17, no. 13, pp. 3949–3960, Jul. 2017,
doi: 10.1109/JSEN.2017.2705700.
[22] Y. Shao, H. Hu, and Y. Visell, “A wearable tactile sensor array for
large area remote vibration sensing in the hand,” IEEE Sensors J.,
vol. 20, no. 12, pp. 6612–6623, Jun. 2020, doi: 10.1109/JSEN.2020.
2972521.
[23] R. K. Kramer, C. Majidi, and R. J. Wood, “Masked deposition
of gallium-indium alloys for liquid-embedded elastomer conductors,”
Adv. Funct. Mater., vol. 23, no. 42, pp. 5292–5296, Nov. 2013, doi:
10.1002/adfm.201203589.
[24] Y.-L. Park, B.-R. Chen, and R. J. Wood, “Design and fabrication of soft
artificial skin using embedded microchannels and liquid conductors,”
IEEE Sensors J., vol. 12, no. 8, pp. 2711–2718, Aug. 2012, doi:
10.1109/JSEN.2012.2200790.
[25] I. D. Joshipura, H. R. Ayers, C. Majidi, and M. D. Dickey, “Methods to
pattern liquid metals,” J. Mater. Chem. C, vol. 3, no. 16, pp. 3834–3841,
2015, doi: 10.1039/c5tc00330j.
[26] T. Lu, L. Finkenauer, J. Wissman, and C. Majidi, “Rapid prototyp-
ing for soft-matter electronics,” Adv. Funct. Mater., vol. 24, no. 22,
pp. 3351–3356, Jun. 2014, doi: 10.1002/adfm.201303732.
[27] R. Rocha, P. Lopes, A. T. de Almeida, M. Tavakoli, and C. Majidi,
“Soft-matter sensor for proximity, tactile and pressure detection,” in
Proc. IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), Sep. 2017,
pp. 3734–3738, doi: 10.1109/IROS.2017.8206222.
[28] A. Eshkeiti et al., “Screen printing of multilayered hybrid printed
circuit boards on different substrates,” IEEE Trans. Compon., Packag.,
Manuf. Technol., vol. 5, no. 3, pp. 415–421, Mar. 2015, doi:
10.1109/TCPMT.2015.2391012.
[29] P. A. Lopes et al., “Soft bioelectronic stickers: Selection and evaluation
of skin-interfacing electrodes,” Adv. Healthcare Mater., vol. 8, no. 15,
Aug. 2019, Art. no. 1900234, doi: 10.1002/adhm.201900234.
[30] P. Filipe, A. Lopes, H. Paisana, A. T. de Almeida, C. Majidi, and
M. Tavakoli. (2018). Hydroprinted Electronics: Ultrathin Stretchable
Ag-In-Ga E-Skin for Bioelectronics and Human-Machine Interaction.
[Online]. Available: http://pubs.acs.org
[31] M. Tavakoli et al., “EGaIn-assisted room-temperature sintering of
silver nanoparticles for stretchable, inkjet-printed, thin-film electron-
ics,” Adv. Mater., vol. 30, no. 29, Jul. 2018, Art. no. 1801852, doi:
10.1002/adma.201801852.
[32] M. Tavakoli et al., “Carbon doped PDMS: Conductance stability
over time and implications for additive manufacturing of stretchable
electronics,” J. Micromech. Microeng., vol. 27, no. 3, Feb. 2017,
Art. no. 035010, doi: 10.1088/1361-6439/aa5ab1.
[33] J. Krzeminski, A. Kanthamneni, D. Wagner, M. Detert, B. Schmidt, and
M. Jakubowska, “Microscale hybrid flexible circuit printed with aerosol
jet technique,” IEEE Trans. Nanotechnol., vol. 17, no. 5, pp. 979–984,
Sep. 2018, doi: 10.1109/TNANO.2018.2841639.
[34] C. Votzke, U. Daalkhaijav, Y. Mengüç, and M. L. Johnston,
“3D-printed liquid metal interconnects for stretchable electronics,”
IEEE Sensors J., vol. 19, no. 10, pp. 3832–3840, May 2019, doi:
10.1109/JSEN.2019.2894405.
[35] D. F. Fernandes, C. Majidi, and M. Tavakoli, “Digitally printed
stretchable electronics: A review,” J. Mater. Chem . C, vol. 7, no. 45,
pp. 14035–14068, Nov. 2019, doi: 10.1039/c9tc04246f.
[36] S. Khan, L. Lorenzelli, and R. S. Dahiya, “Technologies for print-
ing sensors and electronics over large flexible substrates: A review,”
IEEE Sensors J., vol. 15, no. 6, pp. 3164–3185, Jun. 2015, doi:
10.1109/JSEN.2014.2375203.
[37] M. A. Yokus and J. S. Jur, “Fabric-based wearable dry electrodes for
body surface biopotential recording,” IEEE Trans. Biomed. Eng., vol. 63,
no. 2, pp. 423–430, Feb. 2016, doi: 10.1109/TBME.2015.2462312.
[38] D.-H. Kim et al. Epidermal Electronics. Accessed: Oct. 26, 2019.
[Online]. Available: https://www.sciencemag.org
[39] M. Mehrali et al., “Blending electronics with the human body: A path-
way toward a cybernetic future,” Adv. Sci., vol. 5, no. 10, Oct. 2018,
Art. no. 1700931, doi: 10.1002/advs.201700931.
[40] Kenry, J. C. Yeo, and C. T. Lim, “Emerging flexible and wearable
physical sensing platforms for healthcare and biomedical applica-
tions,” Microsyst. Nanoeng., vol. 2, no. 1, pp. 1–19, Dec. 2016, doi:
10.1038/micronano.2016.43.
[41] D. G. Marques, P. A. Lopes, A. T. de Almeida, C. Majidi, and
M. Tavakoli, “Reliable interfaces for EGaIn multi-layer stretchable
circuits and microelectronics,” Lab Chip, vol. 19, no. 5, pp. 897–906,
Feb. 2019, doi: 10.1039/c8lc01093e.
[42] C. Leal, P. A. Lopes, A. Serra, J. F. J. Coelho, A. T. de Almeida,
and M. Tavakoli, “Untethered disposable health monitoring electronic
patches with an integrated Ag2O–Zn battery, a AgInGa current collector,
and hydrogel electrodes,” ACS Appl. Mater. Interfaces, vol. 12, no. 3,
pp. 3407–3414, Jan. 2020, doi: 10.1021/acsami.9b18462.
[43] J. Alberto et al., “Fully untethered battery-free biomonitoring electronic
tattoo with wireless energy harvesting,” Sci. Rep., vol. 10, no. 1,
Dec. 2020, Art. no. 5539, doi: 10.1038/s41598-020-62097-6.
[44] V. Goverdovsky, D. Looney, P. Kidmose, C. Papavassiliou, and
D. P. Mandic, “Co-located multimodal sensing: A next generation solu-
tion for wearable health,” IEEE Sensors J., vol. 15, no. 1, pp. 138–145,
Jan. 2015, doi: 10.1109/JSEN.2014.2338612.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
15116 IEEE SENSORS JOURNAL, VOL. 20, NO. 24, DECEMBER 15, 2020
[45] Y.-L. Zheng et al., “Unobtrusive sensing and wearable devices for health
informatics,” IEEE Trans. Biomed. Eng., vol. 61, no. 5, pp. 1538–1554,
May 2014, doi: 10.1109/TBME.2014.2309951.
[46] G. Wang et al., “Stretchable optical sensing patch system integrated
heart rate, pulse oxygen saturation, and sweat pH detection,” IEEE
Trans. Biomed. Eng., vol. 66, no. 4, pp. 1000–1005, Apr. 2019, doi:
10.1109/TBME.2018.2866151.
[47] G. S. Guvanasen et al., “A stretchable microneedle electrode array for
stimulating and measuring intramuscular electromyographic activity,”
IEEE Trans. Neural Syst. Rehabil. Eng., vol. 25, no. 9, pp. 1440–1452,
Sep. 2017, doi: 10.1109/TNSRE.2016.2629461.
[48] L. M. Ferrari et al., “Ultraconformable temporary tattoo electrodes for
electrophysiology,” Adv. Sci., vol. 5, no. 3, Mar. 2018, Art. no. 1700771 ,
doi: 10.1002/advs.201700771.
[49] P. Lepola et al., “Screen-printed EEG electrode set for emergency
use,” Sens. Actuators A, Phys., vol. 213, pp. 19–26, Jul. 2014, doi:
10.1016/j.sna.2014.03.029.
[50] X. Chen and Z. J. Wang, “Design and implementation of a wearable,
wireless EEG recording system,” in Proc. 5th Int. Conf. Bioinf. Biomed.
Eng., Wuhan, China, 2011, pp. 1–4, doi: 10.1109/icbbe.2011.5781501.
[51] D. Patron et al., “On the use of knitted antennas and inductively
coupled RFID tags for wearable applications,” IEEE Trans. Bio-
med. Circuits Syst., vol. 10, no. 6, pp. 1047–1057, Dec. 2016, doi:
10.1109/TBCAS.2016.2518871.
[52] W. Wu, “Stretchable electronics: Functional materials, fabrication strate-
gies and applications,” Sci. Technol. Adv. Mater., vol. 20, no. 1,
pp. 187–224, Dec. 2019, doi: 10.1080/14686996.2018.1549460.
[53] T.-H. Kang, C. R. Merritt, E. Grant, B. Pourdeyhimi, and H. T. Nagle,
“Nonwoven fabric active electrodes for biopotential measurement during
normal daily activity,” IEEE Trans. Biomed. Eng., vol. 55, no. 1,
pp. 188–195, Jan. 2008, doi: 10.1109/TBME.2007.910678.
[54] J. Malmivuo and R. Plonsey, Bioelectromagnetism,1sted.NewYork,
NY, USA: Oxford Univ. Press, 1995.
[55] W. O. Tatum and A. M. Husain, Handbook of EEG Interpretation,1sted.
New York, NY, USA: Demos Medical Publishing, 2007.
[56] M. C. Fields, J. J. Yoo, and L. V. Marcuse, Rowan’s Primer of EEG,
2nd ed. Amsterdam, The Netherlands: Elsevier, 2003.
[57] A. Albulbul, “Evaluating major electrode types for idle biological signal
measurements for modern medical technology,” Bioengineering,vol.3,
no. 3, p. 20, Aug. 2016, doi: 10.3390/bioengineering3030020.
[58] S. H. Nasrollaholhosseini, J. Mercier, G. F is cher, a nd W. G. B esio,
“Electrode–electrolyte interface modeling and impedance characterizing
of tripolar concentric ring electrode,” IEEE Trans. Biomed. Eng., vol. 66,
no. 10, pp. 2897–2905, Oct. 2019, doi: 10.1109/tbme.2019.2897935.
[59] R. Tchoudakov, O. Breuer, M. Narkis, and A. Siegmann, “Conduc-
tive polymer blends with low carbon black loading: High impact
polystyrene/thermoplastic elastomer (styrene-isoprene-styrene),” Polym .
Eng. Sci., vol. 37, no. 12, pp. 1928–1935, Dec. 1997.
[60] C. Kanzow, N. Yamashita, and M. Fukushima, “Levenberg–marquardt
methods with strong local convergence properties for solving nonlinear
equations with convex constraints,” J. Comput. Appl. Math., vol. 172,
no. 2, pp. 375–397, Dec. 2004, doi: 10.1016/j.cam.2004.02.013.
[61] B. Farnsworth. (2017). EEG Headset Comparison—A Technical
Overview of 10 Headsets. Accessed: Oct. 26, 2019. [Online]. Available:
https://imotions.com/blog/eeg-headset-comparison/
[62] J. Xu et al., “A wearable 8-channel active-electrode EEG/ETI acqui-
sition system for body area networks,” IEEE J. Solid-State Cir-
cuits, vol. 49, no. 9, pp. 2005–2016, Sep. 2014, doi: 10.1109/JSSC.
2014.2325557.
[63] BrainLink Specs. Accessed: Nov. 12, 2019. [Online]. Available:
https://www.mindtecstore.com/Macrotellect-BrainLink-Lite-EEG-
Headset-V20_1
[64] Dreem. Accessed: Nov. 12, 2019. [Online]. Available: https://shop.
dreem.com/europe/
[65] iBand Specs. Accessed: Nov. 12, 2019. [Online]. Available:
https://www.kickstarter.com/projects/arenar/iband-eeg-headband-that-
helps-you-sleep-and-dream/description
[66] BrainBit. Accessed: Nov. 12, 2019. [Online]. Available: https://brainbit.
com/
[67] MindWave Specs. Accessed: Nov. 12, 2019. [Online]. Available:
https://store.neurosky.com/pages/mindwave
[68] L. Inzelberg, M. D. Pur, S. Steinberg, D. Rand, M. Farah, and Y. Hanein,
“Wireless electronic-tattoo for long-term high fidelity facial muscle
recordings,” Proc. SPIE, vol. 10194, May 2017, Art. no. 101940U, doi:
10.1117/12.2263522.
[69] L. Inzelberg and Y. Hanein, “Electrophysiology meets printed elec-
tronics: The beginning of a beautiful friendship,” Frontiers Neurosci.,
vol. 12, p. 992, Jan. 2019, doi: 10.3389/fnins.2018.00992.
[70] A. Dabbaghian, T. Yousefi, S. Z. Fatmi, P. Shafia, and H. Kassiri,
“A 9.2-g fully-flexible wireless ambulatory EEG monitoring and diag-
nostics headband with analog motion artifact detection and compensa-
tion,” IEEE Trans. Biomed. Circuits Syst., vol. 13, no. 6, pp. 1141–1151,
Dec. 2019, doi: 10.1109/TBCAS.2019.2936327.
[71] R. Zhang et al., “An EOG-based human–machine interface to control a
smart home environment for patients with severe spinal cord injuries,”
IEEE Trans. Biomed. Eng., vol. 66, no. 1, pp. 89–100, Jan. 2019, doi:
10.1109/TBME.2018.2834555.
[72] Q. Huang et al., “An EOG-based human–machine interface for wheel-
chair control,” IEEE Trans. Biomed. Eng., vol. 65, no. 9, pp. 2023–2032,
Sep. 2018, doi: 10.1109/TBME.2017.2732479.
[73] G. M. Paul, F. Cao, R. Torah, K. Yang, S. Beeby, and J. Tudor,
“A smart textile based facial EMG and EOG computer interface,”
IEEE Sensors J., vol. 14, no. 2, pp. 393–400, Feb. 2014, doi:
10.1109/JSEN.2013.2283424.
[74] G. Torres, M. P. Cinelli, A. T. Hynes, I. S. Kaplan, and J. R. Leheste,
“Electroencephalogram mapping of brain states,” J. Neurosci. Neuroeng.,
vol. 3, no. 2, pp. 73–77, Dec. 2014, doi: 10.1166/jnsne.2014.1098.
Manuel Reis Carneiro received the M.Sc.
degree in electrical and computer engineer-
ing with a specialization in automation from
the University of Coimbra, Portugal, in 2019.
He is currently a Researcher with the Institute
of Systems and Robotics, University of Coim-
bra. His research interests include stretchable
electronics, printed electronics, biosignals acqui-
sition, human machine interfaces, and wearable
computing.
Aníbal T. De Almeida (Life Senior Member,
IEEE) received the degree in electrical engi-
neering from the Faculty of Engineering, Porto
University, in 1972, the Ph.D. degree in com-
puter control from London University in 1977,
and the Aggregation degree in electrical engi-
neering from the University of Coimbra in 1981.
He is currently a Full Professor with the Electrical
Engineering Department, University of Coimbra,
and the Director of the Institute for Systems and
Robotics, Coimbra.
Mahmoud Tavakoli (Member, IEEE) is currently
an Assistant Professor with the Department
of Electrical Engineering with the University of
Coimbra. He is the author of over 70 publications.
His current research interests are wearable com-
puting, soft and stretchable electronics, printed
electronics, and human-machine interfaces.
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 22,2022 at 14:54:24 UTC from IEEE Xplore. Restrictions apply.
