A Wide-band and User-friendly EEG Recording System for

Wearable Applications

Lorenzo Bisoni, Enzo Mastinu and Massimo Barbaro

Department of Electrical and Electronic Engineering, University of Cagliari, Piazza dArmi, Cagliari 09123, Italy

Keywords:

EEG Recorder, Wearable EEG, Wide-band EEG, Wireless, Bluetooth EEG.

Abstract:

A wireless, wearable and non-invasive EEG recording system is proposed. The system includes a low-power

8-channel acquisition module and a Bluetooth (BT) transceiver to transmit acquired data to a remote platform.

It was designed with the aim of creating a cheap and user-friendly system that can be easily interfaced with

the nowadays widely spread smartphones or tablets by means of a mobile-based application. The presented

system, validated through in-vivo experiments, allows EEG signals recording at different sample rates and

with a maximum bandwidth of 524Hz. It was realized on a 19cm

custom PCB with a maximum power

consumption of 270mW.

1 INTRODUCTION

The electroencephalogram (EEG) is a common tech-

nique for detecting symptoms of neurological dis-

eases such as epilepsy, sleep disorders, anxiety and

learning disabilities. These pathologies have a great

impact on people common life and are quite common.

For example anxiety disorders affects approximately

13.6% of the European population (Alonso J., 2004),

and, in 2010, its overall cost in Europe was e74.4 bil-

lion (J. Olesen, 2012). Most of the mentioned mental

disorders require long-term EEG monitoring tofollow

the course of the disease and sometimes to prevent

further degradations of the patient condition such as

epileptic discharges. In these cases the longer is the

EEG measurements period the higher is the probabil-

ity of a successful event detection. Moreover EEG

acquisition during daily life activities is highly rec-

ommended to better reveal some pathologies.

Traditional ambulatory EEG systems do not sat-

isfy these requirements. In fact patients can be con-

tinuously observed for only a few hours because of

the costs and resource overheads. Moreover there are

some inconveniences such as forcing people to take

time off work and moving them from their natural en-

vironment. As a consequence, patients often feel un-

comfortable and, depending on their pathology, this

can affect the EEG acquisition, introducing undesired

artifacts. As a result of recent technology innovations,

new outpatient EEG systems were introduced. Such

mobile solutions overcome some of these limitations

reducing the overall patient monitoring costs and in-

creasing the effectiveness of the measurements (Wa-

terhouse, 2003). Despite their beneﬁts such systems

are still cumbersome and/or too complex to be used

outside hospitals and require expert assistance.

Wearable EEG is aimed to overcome these issues,

allowing the recording of a longer temporal window

that includes all stages of sleep and wakefulness and

increasing the likelihood of recording typical symp-

toms. Many efforts have been already put on the real-

ization of wearable EEG systems. Some examples are

presented in (Brown L., 2010) and (Carmo J.P., 2007)

in which respectively a semi-custom and a completely

custom CMOS EEG recorder was realized, whereas

a 4-channel BCI-cap based on off-the-shelf compo-

nents is described in (Lin et al., 2008). Other exam-

ples are the Epoc (Emotiv, 2013), the Imec‘s headset

(Patki et al., 2012) and the Quasar‘s (Quasar, 2013)

DSI 10/20. They all use proprietary radio link for data

transmission resulting in a reduced power consump-

tion though they require speciﬁc hardware to inter-

connect a remote back-end. Only a few systems have

been developed using a standard communication link

such as the ThinkGear (NeuroSky, 2009) and the Star-

fast (A. Riera, 2008). Furthermore currently available

EEG systems mainly operate with a bandwidth under

100Hz that may be enough to cover the most com-

mon diagnostic purposes, but a wider bandwidth, up

to 600Hz, is required to investigate some pathologies

(Mihajlovic et al., 2014). As fully discussed in (Mi-

hajlovic et al., 2014), many improvements are still re-

Bisoni L., Mastinu E. and Barbaro M..

A Wide-band and User-friendly EEG Recording System for Wearable Applications.

DOI: 10.5220/0005200500290036

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2015), pages 29-36

ISBN: 978-989-758-071-0

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

quired in order to get a promising solution, easy to

use by non-expert users in a completely uncontrolled

environment. Some critical aspects that refer to the

electrode-skin adherence, the battery life time and the

quality of the acquired EEG signals have to be solved.

According to the idea of spreading the use of

wearable EEG recorders, the system requires a low

impact on people daily life. In other words the device

should not imply the use of additional special equip-

ments and it should be very practical. These require-

ments and the increasing spread of smartphones and

tablets among a wide variety of users, from teen to

elders (Smith, 2013), suggest these new generation

mobiles as the best solution to control the wearable

EEG recorders. Furthermore this choice is supported

by recent researches on moving the telemedicine to-

ward mobile platforms (Sapal Tachakra and Song,

2003), (Chan S.R. and P., 2014), (Lupu and Cosmin-

Constantin, 2013), (El Khaddar et al., 2012).

The EEG system proposed in this paper is based

on a custom PCB with off-the-shelf components

(COTS) and uses a standard BT link to transmit the

acquired signals. As a consequence it exhibits a

higher power consumption compared to those solu-

tions, previously presented, that use custom radio link

but it has the advantage of being easily interfaced with

any BT-based terminal and integrated with such new

healthcare systems. In addition, todays technology al-

lows mobile devices with high computational power,

huge storage memory and fully programmable. In

this way they can store and elaborate the EEG sig-

nals allowing a wide-range of applications. Anyway

our EEG recorder can also be connected to a tra-

ditional desktop PCs for which a simple Microsoft

Windows-based application for testing purposes was

developed. Being a wearable device, special efforts

were made in reducing its power consumption and in

device miniaturization. As a result, only the essential

components were included in the project: an ampli-

fying/ﬁltering block, an analog to digital converter, a

micro-controller, a BT transceiver and a power man-

agement module. In the following Sec.2 the sys-

tem architecture details will be described, Sec.3 con-

tains a brief description of a possible remote interface

whereas the validation test results will be presented in

Sec.4. Finally systems performance and future devel-

opment will be shortly summarized in Sec.5.

2 SYSTEM ARCHITECTURE

The designed system, named BlueThought by joining

the implemented transmission link with the nature of

the acquired signals, is based on a differential 8 chan-

nel recording unit. The EEG signals detected with a

standard EEG cap are ﬁrst ampliﬁed and then con-

verted into digital signals by an ADS1299 component

from Texas Instrument. Once acquired, digital sig-

nals are transmitted to a remote back-end by means

of a Microchip Bluetooth RN-42 module. Moreover a

USB connection was introduced to charge the EEG

recorder battery and as additional channel for data

transfer. A Microchip PIC18F46J50 coordinates data

exchange between ADC and BT or USB external con-

troller. The system architecture is depicted in Fig.1.

Figure 1: BlueThought: System Architecture.

In addition a power management unit generates

all digital and analog voltage supplies for the ADC,

the microprocessor and the BT transceiver from a

3.7V − 950mAh LiPo battery. Even the battery charg-

ing circuit was implemented on the board. The EEG

recorder was realized on the 5.5cmX3.5cm double

face board depicted in Fig.2. In the following further

details on the main modules of the EEG recorder will

be described.

2.1 Signal Conditioning and Digital

Conversion

Before being converted into digital format, the input

signals are ﬁltered and ampliﬁed. To reduce power

consumption and PCB area we decided not to insert

an additional signal conditioning block but to use the

one provided by the A/D converter (ADS1299). This

device contains eight independent differential chan-

nels allowing simultaneous acquisition. An internal

multiplexer allows to select the P and N input sig-

nals among various sources and depending on the

selected signals different recording modes are possi-

ble: normal recording, test and impedance monitor-

ing mode. The normal recording mode is the default

working set-up in which EEG signals are acquired in

both single-ended and differential conﬁguration. In

BIODEVICES2015-InternationalConferenceonBiomedicalElectronicsandDevices

Figure 2: EEG Interface prototype: power management

circuit on bottom side(a); ADS1299 (ADC), PIC18F46J50

(microprocessor), USB and RN-42 (BT transceiver) on top

face(b) and a 3.7V − 950mAh LiPo battery(c).

single-ended measurements the N signals are inter-

nally shorted to the external reference (typically mid-

supply voltage) or to the bias signal generated by the

internal bias module on the base of a desired input

signal combination. Whereas in differential measure-

ments both P and N signals come from the EEG cap.

To reduce the number of interconnectionsbetween the

EEG recorder and the bonnet all N input lines are

shorted together getting only one common reference

electrode conveniently placed on the patient body. In

test mode, different internally-generated test signals

can be selected as input allowing the signal acquisi-

tion chain to be tested out. Another important feature

provided by the ADS1299 is the lead-off detection. It

consists in a continuous patient electrode impedance

monitoring to verify if a suitable connection is present

or not.

The ﬁrst stage of each acquisition channel is a

differential low-noise programmable gain ampliﬁer

(PGA). It offers seven gain settings (1,2,4,6,8,12, and

24) that can be set-up by writing the channel-setting

registers (one per channel) of the ADS1299. As men-

tioned in Sec.1, our EEG recorder can acquire sig-

nals with a bandwidth wider then standard EEG mon-

itor. In fact, as reported in Tab.1, the system supports

different sample rates from 250SPS up to 2000SPS

resulting in a maximum bandwidth of 524Hz. This

makes our device suitable for a wide range of ap-

plications even those requiring the analysis of sig-

nals out of standard EEG frequency. After being

ampliﬁed, the signal is digitalized by a 24-bit sigma

delta converter. The ADC operates in two different

modes: continuous mode (default) and single-shot

Table 1: EEG Recorder Output Data Rate and -3dB BW.

OUTPUT DATA RATE -3dB BANDWIDTH

(SPS) (Hz)

250 65

500 131

1000 262

2000 524

mode. In the ﬁrst modality, when a start command

is sent, it continuously converts the input signal. The

conversion ends when a stop command is received.

Whereas, if the device is in single-shot mode it gen-

erates only one sample per received start command.

This means that to begin a new conversion, a new

start command has to be sent. Regardless of the op-

erating mode, as a single sample conversion ends, a

data-ready signal (DRDY) is pulled down to notify

the microprocessor that a new sample is ready. Af-

ter being converted, the eight samples (each per chan-

nel) are packed and sent to the micro-controller over

a 3MHz SPI connection. Each data packet contains

24-bit of header and 216-bit of sampled data. In the

following the control unit is described. It forwards the

samples received from the ADC to the BT transceiver

or to the USB controller depending if a wireless or a

wired connection is being used.

2.2 Control Unit

The Microchip PIC18F46J50 is used as control unit

to serve two main tasks: system set-up and data ex-

change. The PIC is powered at 3.3V with a CPU clock

frequency of 48MHz generated by an on-chip oscilla-

tor. At system power-up, the PIC is used to setup the

EEG interface deﬁning both recordingand connection

parameters such as acquisition gain and bandwidth,

ADC SPI clock frequency, BT data rate and commu-

nication protocol parameters. All values are tuned to

ﬁnd a good compromise between the acquired EEG

signal quality and the power consumption. Once that

the system started to acquire the EEG signals, the con-

trol unit coordinates data exchange among the con-

verter and BT or USB remote back-end.

The microprocessor provides several internal pe-

ripherals that, if not used, can be disabled to save

power. In particular we are interested in using the

USB and the UART in/out ports to respectively con-

nect the PIC to a remote USB controller or to the BT

transceiver.

Although the system communication mode can be

on-line modiﬁed by the user, if, on power-up, any

device is connected to the USB port, the PIC auto-

matically enables the USB controller otherwise the

AWide-bandandUser-friendlyEEGRecordingSystemforWearableApplications

BT transceiver is turned on. Once deﬁned the con-

nection mode, the microprocessor starts a polling cy-

cle waiting for data coming from the remote con-

troller. The received commands are decoded and ex-

ecuted. Such commands, generally are aimed at con-

trolling the ADC or the BT transceiver but they can

also be addressed to the same microprocessor for ex-

ample to setup the USB controller. The main steps

of the ﬁrmware are described in the ﬂow chart of

Fig.3. Regardless of the back-end connection mode

Initialize PIC & ADC

with default settings

Is any USB device

connected ?

Initialize USB controller

with default settings

Initialize Bluetooth

with default settings

Yes

Has been any command

received ?

Decode command

Execute command

Yes

Figure 3: Main steps of the control unit ﬁrmware.

(via USB or BT) and only in single-shot recording,

the same polling cycle is used by the PIC to send

the sampled data to the remote controller. Other-

wise, in continuous recording, sampled data transmis-

sion is handled by an interrupt service routine. The

ADS1299 data-ready signal (DRDY) is connected to

an interrupt sensitive pin of the PIC acting as an ex-

ternal interrupt. When DRDY is pulled down (i.e.

new samples are available) an exception is raised and

the interrupt service routing is executed. The new

samples are transferred from the ADC to the micro-

processor that forwards data to the remote controller.

Some details about the USB and the BT connection

are given in the following paragraph.

2.3 Data Transmission

In normal operation mode, the BT transceiver allows

wireless data transmission between the EEG recorder

and the remote back-end, whereas the USB controller

is used for battery recharging. However, during test

mode, the wired connection can be used for both

data transfer and system powering. Enabling high

frequency EEG recording, the ADC needs to oper-

ate at sample frequencies up to 2000SPS (see Tab.1)

resulting in a maximum outputs data rate (ODR) of

54Kbyte/sec (Eq.1). This is a critical parameter to

deﬁne the data-exchange channel speciﬁcations.

Single data packet: 216bit

ODR(@2000SPS):

2000∗ 216

= 54Kbyte/sec

(1)

The USB port is directly handled by an on-chip USB

controller and can operate in two different modali-

ties: CDC and HID mode. To make the USB suit-

able for our application, a standard HID protocol was

implemented. Working at full speed (48MHz) with

64-byte data packet size, the data transfer speed is

limited to 64Kbytes/sec. In addition, to make the

transmission more efﬁcient, two sampled data pack-

ets (2x216bit = 432bit) are grouped in the same USB

frame. As a result, in worst conditions (2000SPS), the

required data transfer rate amounts to 27Kbyte/sec

that is below the USB transfer rate limit.

In contrast to the USB HID protocol, the BT

transceiver does not require ﬁxed size packets, but

their length is adapted to the amount of transferred

data. The RN-42 is a small form factor, low power,

class 2 BT radio with on-chip antenna. It delivers up

to a 3Mbps data rate for distances up to 20 meters.

It uses an UART port to communicate with the con-

trol unit and operates in two modes: data mode (de-

fault) and command mode. In data mode, the module

works as a data pipe. When the module receives data,

it strips the BT headers and forwards the data to the

UART port. When data is written to the UART port,

the module constructs the BT packet and sends it out

over the BT wireless connection. Thus, the entire pro-

cess of sending/receiving data to the host is transpar-

ent to the PIC. The command mode is used to deﬁn-

ing the BT operating mode, the UART baud rate and

others control ﬂow parameters. Moreover the RN-42

operates in slave mode so that other BT devices (PC,

tablet or smartphone) can discover and connect to the

module.

BIODEVICES2015-InternationalConferenceonBiomedicalElectronicsandDevices

Figure 4: Visual C++ EEG interface: system setup window.

3 REMOTE INTERFACE

The designed system is a general purpose EEG

recorder and depending on the treated pathology a

speciﬁc software can be developed. At this ﬁrst stage

of the project a Visual C++ application was written,

implementing only the essential features for the hard-

ware debugging. The ADC module can be completely

conﬁgured in terms of PGA gain and sample rate and

both continuous and single-shot modes are selectable.

The eight recorded signals are plotted for a real-time

view and stored in a text ﬁle for off-line data com-

puting. Fig.4 and Fig.5 show the two main window

of the developed interface. The ﬁrst refers to the sys-

tem settings and the second to the plotting of the eight

recorded signals. For all channels is possible to setup

the amplitude scale (µV, mV or V) and the temporal

window size.

4 RESULTS

All system features were ﬁrst characterized and than

compared with a standard laboratory equipment. To

start with its static electrical characterization, Tab.2

collects the EEG interface power consumption in dif-

ferent working conditions. In idle state (only the mi-

croprocessor is on) it has a minimum power consump-

tion of about 119mW whereas in worst conditions (i.e.

all devices are on, sample rate of 2000SPS and ac-

tive BT data transmission) it absorbs a maximum of

Figure 5: Visual C++ EEG interface: on-line plotting win-

dow in a µV amplitude scale and with a 6s temporal window.

Table 2: EEG Recorder Power Consumption in different

working conditions.

Power A/D A/D

Consumption Converter Converter

(mW) (OFF) (ON)

Bluetooth (OFF) 119 158

Bluetooth (ON) 230 270

270mW. Under this conditions and with the chosen

battery (3.7V − 950mAh LiPo) the system can contin-

uously work for about 13 hours. Further experiments

were performed to study the dynamic behaviour of the

EEG acquisition channel. Its gain programmability

from 1V/V up to 24V/V was conﬁrmed acquiring a

12mV − 30Hz sine as depicted in the above plot of

Fig.6. The device showed a 63.5Hz −3dB bandwidth

at sample rate of 250SPS and the magnitude bode di-

agram of the recording channel transfer function is

depicted in Fig.6 (below plot). Moreover both wired

(USB) and wireless (BT) connections were tested.

Once the system main functions have been proved,

some in-vivo EEG measurements, on one human sub-

ject, were performed. To evaluate the signal quality of

the designed EEG recorder, the system was compared

with a commercial device (Brain QUICK,(Micromed,

2014)) depicted in Fig.9 where the huge difference

in terms of dimensions between the two devices is

also highlighted. Moreover, the experimental setup,

depicted in Fig.10, includes a commercial EEG cap

(KIT-CAP-SPEXT61 from Micromed) with 61 elec-

trodes used to acquire the neural signals.

To better compare the two devices, they were

AWide-bandandUser-friendlyEEGRecordingSystemforWearableApplications

0 5 10 15 20 25 30

−200

−150

−100

−50

100

150

200

time [ms]

Voltage [mV]

PGA Gain test (input signal: sine, f=30Hz, Vmax=12mV)

G=1, Vmax =12.5mV

G=2, Vmax=24mV

G=4, Vmax=46mV

G=6, Vmax=68mV

G=8, Vmax=87mV

G=12, Vmax=120mV

G=24, Vmax=180mV

−1

−40

−20

EEG Acquisition Channel −3dB Bandwidth at 250SPS

Frequency [Hz]

Magnitude [dB]

Figure 6: The EEG interface gain programmability vali-

dated recording a 12mV − 30Hz sine at 250SPS (on the top)

and the system -3dB bandwidth (63.5Hz) at 250SPS (on the

bottom).

connected to adjacent electrodes and simultaneous

recordings were performed in different patient con-

ditions. During the ﬁrst test, the human subject was

in resting state with closed eyes to avoid any kind of

artifact. Fig.7 and Fig.8 show the EEG signals re-

spectively acquired by the Brain Quick and by our

EEG recorder. The signals recorded at 250SPS are

quite similar in both time and frequency domains. In

Fig.11 it is possible to appreciate the differences be-

tween an open-eyes (on the left) and a closed-eyes (on

the right) EEG signal perfectly recorded by our device

at 250SPS. Finally, to further validate our system,

some typical EEG artifacts such as the teeth-grinding

signal (Fig.12) and the eyes-blinking effect (Fig.13)

were recorded. They respect the typical shapes and

amplitudes of such signals.

5 CONCLUSIONS AND

DISCUSSION

An off-the-shelf based EEG interface was presented.

It is a wearable system that, thanks to its small di-

mensions (height: 5.5cm x width: 3.5cm x depth:

1.0cm), can be easily placed on the patient head and

integrated with the electrodes framework. The devel-

oped device has 8 independent acquisition channels

and was designed with the aim of being user-friendly

and suitable for all applications in which a contin-

Figure 7: Closed-eyes EEG signal, in time and frequency

domain, recorded at 250SPS by Micromed Brain QUICK

(Micromed, 2014).

Figure 8: Closed-eyes EEG signal, in time and frequency

domain, recorded by our device at 250SPS.

uous EEG monitoring is required. In fully working

condition (i.e. when acquiring and transmitting data)

the system exhibits an overall power consumption of

270mW. Even-though it is higher than of other sys-

tems (Tab.3), the device allows 13 hours of continu-

ous signal recording that is in line with other wearable

devices. The higher power consumption is mainly

due to the choice of using a COTS solution and a BT

link to connect a remote controller. However it gives

BIODEVICES2015-InternationalConferenceonBiomedicalElectronicsandDevices

Table 3: Comparison between some state-of-the-art EEG Recorders.

Our device Quasar Imec Emotiv Epoc NeuroSky Brown L. Enobio

CMRR > 110dB > 120dB 115dB

Input Impedance 1GΩ 47GΩ

Bandwidth 0.01− 524Hz 0.02− 120Hz 0.3− 100Hz 0.2− 45Hz 3− 100Hz 0.5− 375Hz 0− 250(500)Hz

Channel number 8 12 12 14 1 8 8− 20(32)

Noise < 2µVpp 3µVpp 4µVpp 1µVpp < 1µVrms

Bit number 24 16 12 16 11 24

Wireless protocol BT Proprietary Nordic RF Proprietary BT Proprietary BT

Power consumption 270mW 42mW 130mW 12mW

Run time 13h 24h 12h 10h 30h 16h

Technology COTS ASIC / COTS ASIC / COTS

Figure 9: Comparison between our wearable EEG interface

(red circle) and a cumbersome commercial device (green

box).

the device the great advantage to easily connect any

BT-based end-terminal in contrast to other systems

that, using a proprietary wireless link, require speciﬁc

hardware. Moreover, compared to others state-of-the-

art equipments, our EEG recorder has a wider band-

width, up to 524Hz, allowing high-frequency EEG

monitoring. This can be very useful to deeper under-

stand and investigate a certain number of pathologies.

In addition, a Windows-based Visual C++ software

was written for the EEG recorder testing purpose. The

system was completed validated by in-vivo measure-

ments on human patient and compered with a com-

mercial laboratory equipment.

Moving towards a device that can easily become

part of everyday life for all people improving their

Figure 10: Experimental setup for in-vivo EEG measure-

ments.

Figure 11: EEG recorded signal with opened-eyes (on the

left) and closed-eyes (on the right).

living conditions from both health and entertainment

points of view and with the least economical and

daily-activity impact is our main goal. Therefore, be-

ing a wearable device, next developments are the re-

duction of the power consumption and the develop-

ing of smartphone-based application to respectively

increase the battery life and to make the system com-

pletely portable. In particular, some future improve-

ments include the use of a new generation BT called

Bluetooth 4.0 Low Energy (BTLE) that drastically re-

duce the power transmission and a review of the con-

trol unit strategy turning off, time by time, all on-

AWide-bandandUser-friendlyEEGRecordingSystemforWearableApplications

board unused devices. Moreover the possibility to op-

tionally expand the number of input channels by plug-

ging in an additional acquisition module and the intro-

ducing of on-board data storage capabilities might be

considered. Finally, a custom chip solution for sig-

nal conditioning and converting will might be investi-

gated to further reduce both power consumption and

system dimensions.

Figure 12: EEG recorded signal with teeth-grinding arti-

facts.

Figure 13: EEG recorded signal with eyes-blinking arti-

facts.

ACKNOWLEDGEMENTS

The authors would like to thank Dr. Matteo Fraschini

and Matteo Demuru from the University of Cagliari

for their support on EEG recording in-vivo experi-

ments. L. Bisoni gratefully acknowledges Sardinia

Regional Government for the ﬁnancial support of his

PhD scholarship (P.O.R. Sardegna F.S.E. Operational

Programme of the Autonomous Region of Sardinia,

European Social Fund 2007-2013 - Axis IV Human

Resources, Objective l.3, Line of Activity l.3.1.).

REFERENCES

A. Riera, S. Dunne, I. C. G. R. (2008). Starfast: a wireless

wearable eeg/ecg biometric system based on the eno-

bio sensor. Proceedings of the International Workshop

on Wearable Micro and Nanosystems for Personalised

Health (pHealth08).

Alonso J., Angermeyer MC., B. S. e. a. (2004). Prevalence

of mental disorders in europe: results from the euro-

pean study of the epidemiology of mental disorders

(esemed) project. Acta Psychiatrica Scandinavica,

109:21–27.

Brown L., Van de Molengraft J., Y. R. T. T. e. a. (2010). A

low-power, wireless, 8-channel eeg monitoring head-

set. Engineering in Medicine and Biology Society

(EMBC), 32

d Annual International Conference of

the IEEE EMBS:4197–4200.

Carmo J.P., Dias N.S., S.H.M.P. e. a. (2007). A 2.4-ghz low-

power/low-voltage wireless plug-and-play module for

eeg applications. Sensors Journal,7:1524–1531.

Chan S.R., Torous J., H. L. and P., Y. (2014). Mobile tele-

mental health: Increasing applications and a move to

hybrid models of care. Healthcare, 2(2):220–233.

El Khaddar, M., Harroud, H., Boulmalf, M., ElKoutbi, M.,

and Habbani, A. (2012). Emerging wireless technolo-

gies in e-health trends, challenges, and framework

design issues. Multimedia Computing and Systems

(ICMCS), 2012 International Conference on, pages

440–445.

Emotiv (2013). http://emotiv.com.

J. Olesen, A. Gustavsson, M. S. H.-U. W. e. a. (2012). The

economic cost of brain disorders in europe. European

Journal of Neurology, 19:155–162.

Lin, C.-T., Ko, L.-W., Chiou, J.-C., Duann, J.-R., Huang,

R.-S., Liang, S.-F., Chiu, T.-W., and Jung, T.-P.

(2008). Noninvasive neural prostheses using mo-

bile and wireless eeg. Proceedings of the IEEE,

96(7):1167–1183.

Lupu, C. and Cosmin-Constantin, M. (2013). Actual

portable devices as base for telemedicine and e-health:

Research and case study application. E-Health and

Bioengineering Conference (EHB), 2013, pages 1–4.

Micromed (2014). Brain quick. http://www.micromed.eu.

Mihajlovic, V., Grundlehner, B., Vullers, R., and Penders, J.

(2014). Wearable, wireless eeg solutions in daily life

applications: What are we missing? Biomedical and

Health Informatics, IEEE Journal of, PP(99):1–1.

NeuroSky (2009). http://neurosky.com.

Patki, S., Grundlehner, B., Verwegen, A., Mitra, S., Xu,

J., Matsumoto, A., Yazicioglu, R., and Penders, J.

(2012). Wireless eeg system with real time impedance

monitoring and active electrodes. Biomedical Circuits

and Systems Conference (BioCAS), 2012 IEEE, pages

108–111.

Quasar (2013). http://www.quasarusa.com.

Sapal Tachakra, X.H. Wang, R. S. I. and Song, Y.

(2003). Mobile e-health: The unwired evolution of

telemedicine. Telemedicine Journal and e-Health,

9(3):247–257.

Smith, A. (2013). Smartphone ownership 2013.

Pew Research Center, http://www.pewinternet.org

/2013/06/05/smartphone-ownership-2013/.

Waterhouse, E. (2003). New horizons in ambulatory elec-

troencephalography. Engineering in Medicine and Bi-

ology Magazine, 22:74–80.

BIODEVICES2015-InternationalConferenceonBiomedicalElectronicsandDevices