A Multimodal Low-cost Platform for Acquisition of Electrophysiological

Signals Interfacing with Portable Devices

A. Santos Ribeiro

1,2

, D. Salvado

2,3

, G. Evans

, J. Soares Augusto

and H. A. Ferreira

Centre for Neuropsychopharmacology, Division of Brain Sciences, Department of Medicine,

Imperial College London, London, U.K.

Institute of Biophysics and Biomedical Engineering, Faculty of Sciences, University of Lisbon, Lisbon, Portugal

Institute of Nuclear Medicine, University College London, London, U.K.

Centro de F

ısica da Mat

eria Condensada, Faculty of Sciences, University of Lisbon, Lisbon, Portugal

Instituto de Engenharia de Sistemas e Computadores: Investigac¸

ao e Desenvolvimento, Lisbon, Portugal

Keywords:

Multimodal Acquisition, Physiological Signals, Arduino, Mobile, Low-Cost.

Abstract:

Advances in low-voltage integrated circuits have enabled the development of low-cost, low-power, and down-

sized portable instrumentation. In the biomedical ﬁeld, mobile sensing platforms provide an efﬁcient way to

monitor the physical condition of a subject. Moreover, these platforms provide an input for human-computer

interaction. We developed a low-cost platform that can be adapted to acquire different electrophysiological

signals, and interface with portable devices for storing, processing, and displaying of data. The developed

platform was used to acquire electrocardiography (ECG), electromyography (EMG), electroencephalography

(EEG), and electrooculography (EOG) signals, and the results were compared with signals obtained with the

benchmark BIOPAC system. For the same frequency bands, results show that our portable platform was able

to acquire electrophysiological signals with similar accuracy as those acquired with the BIOPAC system. Due

to its simplicity, low-cost design, and easy implementation, the developed platform suits researchers, devel-

opers, and hobbyists, in the ﬁelds of physiological monitoring, human-computer interaction, and perceptual

computing.

1 INTRODUCTION

Recent advances in the miniaturization and availabil-

ity of portable biomedical devices have shown to im-

prove healthcare quality (West, 2012). The appli-

cation of mobile health monitoring systems in am-

bulatory, emergency, home, and point-of-care set-

tings provide a greater access to physiological data,

leading to improved therapeutic decision-support, and

decision-making. Similarly, rehabilitation procedures

(Bin Ambar, 2012)(Roy, 2009), physiology-driven

robotics (Yin et al., 2012), and human-computer in-

teraction (Zheng, 2009)(Kim, 2004)(Andreoni et al.,

2007) should beneﬁt from the use of portable biomed-

ical devices. The wide range of applications promotes

further research in the area of system design and con-

trol for increased reliability, multimodality integra-

tion, and easy implementation and dissemination.

Most of commercially available biomedical

portable devices are dedicated systems and usually

focus on only one kind of electrophysiological sig-

nal (e.g. (Emotiv, 2013)(Alive Technologies, 2013)).

Although beneﬁcial, the use of such devices is con-

ﬁned to speciﬁc tasks and applications. On the other

hand, multimodal acquisition systems on the mar-

ket (e.g. (PLUX wireless biosignals, 2013)(Shimmer,

2013)) are designed based on two components: (1)

a main unit for data storage, transmission, and pro-

cessing, and (2) dedicated sensors for the acquisi-

tion of speciﬁc electrophysiological signals. Either

single or multimodal approaches lead to unnecessary

costs, as in the ﬁrst case a completely different plat-

form is needed for each electrophysiological signal

to be acquired, and in the second case, different sen-

sors are required. However most electrophysiological

signals require the same acquisition steps: differen-

tial ampliﬁcation, ﬁltering, and additional ampliﬁca-

tion, depending on the target signal (Webster, 2010).

Therefore, a single customizable platform for the ac-

quisition of the different electrophysiological signals

would be desirable and more cost-effective.

This work illustrates the development and imple-

Santos Ribeiro A., Salvado D., Evans G., Soares Augusto J. and Ferreira H..

A Multimodal Low-cost Platform for Acquisition of Electrophysiological Signals Interfacing with Portable Devices.

DOI: 10.5220/0004885000630070

In Proceedings of the International Conference on Physiological Computing Systems (PhyCS-2014), pages 63-70

ISBN: 978-989-758-006-2

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

mentation steps of a low-cost multimodal acquisition

platform, which interfaces with portable devices, such

as laptops, tablets or smartphones to collect, record,

process, and display electrophysiological data.

2 MATERIAL AND METHODS

The developed platform is comprised of three main

blocks: an analogue circuit, an Arduino MEGA ADK

board, and a mobile platform.

The analogue circuit was simulated in a general-

purpose circuit simulation program (B

Spice, Beige

Bag Software), and implemented in a printed circuit

board (PCB) using CadSoft EAGLE PCB design soft-

ware. This platform was then connected to the Ar-

duino MEGA ADK board, which was programmed to

collect the signal, and connect to the mobile platform

for data processing.

In this section, we will discuss in detail the stages

of circuit development and implementation, and the

interface with the Arduino and mobile platforms.

2.1 Analogue Circuit

The main analogue circuit is comprised of 8 sub-

circuits, Figure 1: electrical protection circuit; differ-

ential (1st stage) ampliﬁcation with a Driven Right

Leg circuit (DRL); low-pass ﬁlter; 2nd stage ampliﬁ-

cation; notch ﬁlter; high-pass ﬁlter; 3rd stage ampliﬁ-

cation; and voltage offset circuit.

Protective

stage

Differential

amplification

with DRL

Drift removal

and low

pass filter

2nd stage

amplification

3rd stage

amplification and

voltage offset

High pass

filter

Notch filter

ADC and

USB

interface

Data collection,

visualisation,

and processing

TCP/IP

connection

with MATLAB

Figure 1: Pipeline of the developed multimodal acquisition

platform. Dark blue stages correspond to the analogue cir-

cuit, medium blue stages to the Arduino Mega ADK board,

and light blue stages to the mobile platform.

2.1.1 Circuit Development

The ﬁrst sub-circuit (Figure 2) attenuates high fre-

quencies present in the acquired signal, using a pas-

sive low-pass ﬁlter. Such circuit aims to reduce the

radio-frequency (RF) noise induced in the cables, and

its propagation from the electrodes to the analogue

circuit. Additionally, this sub-circuit protects the user

against electrical shock (high voltage) by means of

two inverted diodes. Diodes are analogue components

that only allow the passage of signal if the signal goes

in the diode forward direction and if there is a differ-

ential potential over 0.7V at the diode terminals. Us-

ing this conﬁguration, signals over +0.7V and under

-0.7V ﬂow directly to ground, while signals within

these limits are not affected. Note that this is a re-

dundant protective stage, and that the user should not

be at any time connected directly to the main power

supply. For example, if the computer is connected to

the main power supply, and the acquisition platform is

connected to the computer, the user is also connected

to the main power supply. Thus there would be a high

risk of electrical shock in the presence of an electrical

discharge followed by failure of the protective circuit.

In this case, the main protective stage is therefore the

use of batteries in the acquisition device.

Figure 2: Protective sub-circuit and RF removal.

The sub-circuit following the protective stage cir-

cuit is responsible for the acquisition and ampliﬁca-

tion of the low amplitude electrophysiological sig-

nals, which range from microvolts to millivolts (Fig-

ure 3). The typical design approach is to use an in-

strumentation ampliﬁer (IA) that multiplies the dif-

ference between the two inputs, typically between

10 and 100, reducing the common-mode noise. For

this type of application, a low-noise, low-drift and

low-power consumption IA is required, such as the

INA114, which is widely applied in medical instru-

mentation (Burr-Brown Corporation, 1998). To im-

prove the common-mode noise rejection a driven-

right-leg (DRL) circuit may be used. In this system,

the ground electrode is connected through a feedback

loop to the IA, instead of directly to the reference in-

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

put of the IA. This circuit feeds the user with a small

current that is the inverse of the common-mode noise

(obtained from the IA), therefore reducing the overall

noise.

Figure 3: Differential ampliﬁer sub-circuit with DRL feed-

back.

Although the chosen IA has low drift voltage

(50µV maximum for the INA114), this offset may

still lead to the circuit malfunction after the ampliﬁca-

tion blocks. For example, in a typical EEG board, an

ampliﬁcation between 1000 and 100000 is required,

therefore the offset is ampliﬁed from 50µV to 5V .

To account for this effect, a passive high-pass ﬁlter

(HPF) with a very low cut-off frequency can be used,

blocking continuous current while allowing variable

current to pass. Next to this circuit, a low-pass ﬁl-

ter (LPF) transparent to the frequencies below the fre-

quencies of interest is used (Figure 4).

Figure 4: Offset removal and LPF with variable cut-off fre-

quency sub-circuits.

As the frequencies of interest vary with the tar-

get electrophysiological signal (Table 1), a four-

input one-output switch was used for the purpose of

achieving multimodality. Hence either electrocardio-

graphy (ECG), electromyography (EMG), electroen-

cephalography (EEG) or electrooculography (EOG)

signals can be recorded adjusting the values of the re-

sistors, in order to change the value of the LPF cut-off

frequencies.

As the ﬁltered signal still has a very low ampli-

tude, a second ampliﬁcation stage is needed (Figure

5). For the same reasoning used for the LPF, a four-

input one-output switch was used to increase differ-

ently the amplitude of the signal according to the elec-

trophysiological signal of interest. For example, the

EEG signal has a very low amplitude and will need

higher ampliﬁcation when compared to an ECG sig-

nal, which can have an amplitude 1000 times higher

(Table 1).

Figure 5: Variable second stage ampliﬁcation sub-circuit.

The next circuit module is an active notch-ﬁlter

(NF), designed to attenuate the electric grid noise at

50/60 Hz

. For this purpose, a band-stop ﬁlter centred

at 50Hz was implemented according to the schematic

in Figure 6. As the main hub noise has frequencies

in the range of those of electrophysiological signals, a

switch was implemented to turn on and off the notch

ﬁlter, depending on the application.

Figure 6: Main hub (50Hz) notch ﬁlter sub-circuit.

At this stage, the acquired signal has already been

ﬁltered for low-frequencies and main hub noise, as

well as ampliﬁed. Yet, an HPF is required to reduce

the noise which has frequencies above the frequen-

cies of interest. For the purpose of generalization and

customization, the HPF was also adapted for different

inputs (Figure 7).

Although the analogue signal could already be

passed to a microprocessor for analog-to-digital con-

version (ADC), and passed to a mobile device or lap-

top via USB or Bluetooth connection, programming a

microprocessor is not straightforward. An easy way

to do it is to use an Arduino platform, which is our

purpose. Therefore two ﬁnal analogue steps are re-

quired to use the Arduino board (Figure 8). Firstly,

EU and US mains frequency, respectively.

AMultimodalLow-costPlatformforAcquisitionofElectrophysiologicalSignalsInterfacingwithPortableDevices

Table 1: Frequency and amplitude ranges for ECG, EMG, EEG, and EOG(National Instruments, 2013).

Signal Frequency range (Hz) Amplitude range (mV)

ECG 0.01 − 300 0.05 − 3

EMG 50 − 3000 0.001 − 100

EEG 0.1 − 100 0.001 − 1

EOG 0.1 − 10 0.001 − 0.3

Figure 7: High-pass ﬁlter with variable cut-off frequency

sub-circuit.

a third ampliﬁcation stage is required to increase the

signal amplitude of the acquired signals to the 0-5V

interval, specially for the EEG signal. Secondly, a

voltage offset circuit is required to translate the signal

to positive voltages only.

Figure 8: Third stage ampliﬁer and voltage offset sub-

circuits.

2.1.2 PCB Implementation

After having the circuit design completed, tested, and

simulated, it was implemented in a printed circuit

board (PCB). The main issue in this stage is to oc-

cupy the least possible space and avoid acute angles

- to reduce induced noise -, whilst trying to have few

crossing connections. Therefore the components are

placed in the board and organised in a way such that

connections between components have the shortest

length possible. Ground connections were performed

with a polygon instead of lines, such as to reduce the

overall noise (by increasing the ground area). The ﬁ-

nal manufactured board has 52mm height and 92mm

width.

In the developed board (Figure 9) a two-layer PCB

was implemented. This conﬁguration was chosen be-

cause it is the standard used by many PCB manufac-

tures, and therefore reduces costs of production.

Figure 9: Multimodal electrophysiological signal acquisi-

tion PCB board (with components). The board has 52mm

height and 92mm width, and two layers.

2.2 Arduino Platform

The analogue board was connected to an Arduino

MEGA ADK platform in order to use its ADC. The

Arduino MEGA ADK board has 54 digital input/out-

put pins, of which 15 can be used as Pulse Width

Modulation (PWM) outputs, 16 analogue inputs, 4

UARTs (hardware serial ports), a 16MHz crystal os-

cillator, and a USB connection. It has a 10-bit resolu-

tion, and is able to acquire analogue signals at a maxi-

mum sampling frequency of 10kHz, therefore suitable

for handling biomedical data.

Due to the high number of analogue inputs, the

Arduino platform can be easily set up for additional

acquisition boards, i.e. channels; this improves the

customization of the device, and facilitates the inte-

gration of different platforms.

The USB host interface given by MAX3421E IC

allows the connection, and subsequent interaction, of

the Arduino board with any type of device that has

a USB port. For example, it may be used to interact

with many types of phones, control Canon cameras,

or interface with keyboards, mouses and games con-

trollers as the Wii remote and PS3 controller.

The Arduino programming language is an imple-

mentation of Wiring, a similar physical computing

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

platform, which is based on the Processing multi-

media programming environment. Arduino programs

are written in C/C++. The Arduino IDE comes with

the software library Wiring from the original Wiring

project (Wiring, 2013), which makes many common

input/output operations easier. To make a runnable

cyclic executive program developers only need de-

ﬁne two functions: setup() - a function run once at

the start of a program that can initialize settings; and

loop() - a function called repeatedly until the board

powers off.

2.3 Mobile Platform

To communicate with the Arduino board, a desktop

computer or mobile platform can be used. The in-

terface is programmed such as to receive the digital

signal from the Arduino via microUSB connection,

and to store the data either in a desktop hard disk,

or in a mobile internal disk or SDcard. In a typical

laptop-Arduino connection, high level programming

languages (e.g. MATLAB) can be used to decrease

programming hassle. This becomes even more impor-

tant as more complex applications are needed, such as

real-time processing or visualization.

The ability to interface with mobile platforms

brings full portability to the acquisition device, yet

usually at the expense of more complex program-

ming software. On the other hand, high level pro-

gramming languages, e.g. Octave and MATLAB, and

typical desktop operating systems (OS), e.g. Linux,

are also being implemented in mobile platforms due

to their increased processing capabilities. These ad-

vances open new possibilities of mobile communica-

tion for users with low programming skills, and for

the development of general applications, opposed to

OS-speciﬁc applications.

Based on these ideas, a mobile communication

script was developed to store the data acquired with

the Arduino platform in the mobile platform, to be

further processed and visualized with the GNU Oc-

tave application, or with MATLAB via computer with

a TCP/IP connection. It is important to note that cur-

rently there is no direct communication from the Ar-

duino platform to the mobile Octave application, and

the mobile MATLAB software only works through

cloud computing. An example of communication

from the Arduino platform to MATLAB is presented

below.

1 uns ign ed l on g t i m e ;

2 i n t s e n s o r V a l u e ;

3 i n t po s ;

5 vo i d e s t a b l i s h C o n t a c t ( ) {

6 w h i l e ( S e r i a l . a v a i l a b l e ( ) <= 0) {

7 S e r i a l . p r i n t l n ( ’A’ ) ;

8 d e l a y ( 3 0 0 ) ;

9 }

10 }

12 vo i d s e t u p ( ) {

13 S e r i a l . b e g i n ( 1 1 5 2 0 0 ) ;

14 e s t a b l i s h C o n t a c t ( ) ;

15 }

17 vo i d l o o p ( ) {

18 t i m e = m i l l i s ( ) ;

19 S e r i a l . p r i n t ( t i m e ) ;

20 f o r ( pos =0 ; pos <10; p os ++) {

21 S e r i a l . p r i n t ( ’ , ’ ) ;

22 s e n s o r V a l u e = a n al o g R e ad ( A0 ) ;

23 S e r i a l . p r i n t ( s e n s o r V a l u e ) ;

24 }

25 S e r i a l . p r i n t l n ( ’ , ’ ) ;

26 }

arduino code.c

1 s 1 = s e r i a l ( ’COM# ’ ) ;

2 s 1 . B audR ate = 1 15 20 0;

3 s e t ( s1 , ’ t e r m i n a t o r ’ , ’LF ’ ) ;

4 f op en ( s1 ) ;

6 w= ’B ’ ;

7 w h i l e w˜= ’A’

8 w= f s c a n f ( s1 , ’%s ’ ) ;

9 f p r i n t f ( s1 , ’%s \n ’ , ’A ’ ) ;

10 end

12 a c q t i m e = z e r o s ( 1 , 1 0 0 0 0 0 0 ) ;

13 a c q d a t a = z e r o s ( 1 , 1 0 0 0 0 0 0 ) ;

14 p os =1 ;

15 w h i l e 1

16 s i z = 1 0 ;

17 r a w d a t a = f s c a n f ( s1 ) ;

18 d e l i m i t e r = ’ , ’ ;

19 r a w d a t a = t e x t s c a n ( r a w d a t a , ’%d

’ , ’ d e l i m i t e r ’ , d e l i m i t e r ) ;

20 r a w d a t a = d o u b l e ( r a w d a t a { 1 }) ;

21 a c q d a t a ( ( pos −1)∗ s i z + 1 : po s ∗ s i z ) =

a l l d a t a ( 2 : s i z +1 ) ∗ 5 / 1 0 2 4 ;

22 a c q t i m e ( pos ) = a l l d a t a ( 1 ) ;

23 pos = p os +1 ;

24 end

25 f c l o s e ( s 1 ) ;

matlab code.m

AMultimodalLow-costPlatformforAcquisitionofElectrophysiologicalSignalsInterfacingwithPortableDevices

2.4 Set-up

The developed platform was used to acquire: a lead I

ECG; an EMG of the brachioradialis muscle contrac-

tion; an EEG of the occipital cortex; and an EOG of

the horizontal direction for a reading task. Recorded

signals were then compared to the corresponding sig-

nals obtained with the commercially available bench-

mark BIOPAC system to evaluate the performance of

the developed platform. All acquisitions were per-

formed with one channel - three electrodes: positive,

negative and reference. The platform was supplied

with a dual voltage source of 9V. In addition, the high-

pass and low-pass ﬁlters’ cut-off frequencies and am-

pliﬁcation gains were adjusted to those of interest for

acquiring ECG, EEG, EMG, and EOG signals (Web-

ster, 2010).

3 RESULTS

The multimodal platform is shown in Figure 10. The

communication of the analogue circuit with the Ar-

duino Mega ADK platform was performed through

jumper cables (not shown in Figure 10). Other inter-

face options could have been implemented, such as

adapting the analogue circuit into an Arduino shield.

Such approaches are easily implemented as both plat-

forms have roughly the same size. Further considera-

tions have been made to implement the proposed plat-

form through surface-mount technology to decrease

size or to provide additional channels within the same

size. Although such approach is highly recommended

in commercial devices it may impair the implementa-

tion for researchers and hobbyists with low practice

in soldering electronic circuits.

The communication of the Arduino with the mo-

bile platform is performed through a microUSB ca-

ble. This approach leads to simpler hardware and

software implementation. A wireless communication

between the two could be implemented without hav-

ing to re-design the analogue circuit: one could com-

bine a Bluetooth shield with the Arduino platform.

The acquired ECG, EMG, EEG, and EOG signals

obtained with the developed acquisition platform are

shown in Figure 11 (left) as well as the correspond-

ing signals obtained with the BIOPAC system (right).

The cut-off frequencies of the low-pass (fcLP) and

high-pass (fcHP) ﬁlters were adjusted, respectively,

to fcLP = 116, 248, 116, 48 Hz and fcHP = 0.04, 16,

0.16, 0.05 Hz. For the BIOPAC platform the cut-off

frequencies of the LPF and HPF were selected, re-

spectively, to fcLP = 100, 250, 100, 30 Hz and fcHP

= 0.05, 30, 0.5, 0.05 Hz. Results show that the differ-

Figure 10: Developed multimodal acquisition platform.

From left to right: Analogue board, Arduino MEGA ADK

and mobile platform (connections between the Analogue

board and the Arduino MEGA ADK, and battery power

supply not shown for clarity).

ent signals present similar traces for both acquisition

platforms. Moreover a higher signal-to-noise ratio is

observed for the ECG signal of the proposed system

when compared to the BIOPAC platform, while for

the EMG and EOG a lower signal-to-noise ratio is

veriﬁed.

Figure 11: Acquired signals. ECG (ﬁrst row), EMG (sec-

ond row), EEG (third row) and EOG (fourth row) signals

acquired by the developed platform (left) and the BIOPAC

system (right).

An estimation of the cost of the proposed proto-

type is presented in Table 2, for a 1- and 4-channel

acquisition platform. The predicted cost for both

the single, and multichannel acquisition platform is

≤ 250e, thus presenting a low-cost approach to mul-

timodal acquisition systems. Note that the price does

not scale linearly with the number of channels, even

for prototyping, because of the fact that one Arduino

Mega ADK is able to simultaneously acquire up to 16

channels, requiring no additional boards, and a higher

number of PCB boards is cheaper.

These results suggest that accurate and low-cost

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

Table 2: Approximated cost estimation for a 1- and 4-channel acquisition board in euros (e).

Component 1-channel acquisition board 4-channel acquisition board

Instrumentation ampliﬁer (INA114AP) 10e 40e

Other through hole components 15e 60e

PCB Board 35e 100e

Arduino Mega ADK 50e 50e

Total 110e 250e

multimodal solutions can be developed for biomed-

ical signal acquisition, without requiring expertise

in both electronics and programming. As previ-

ously suggested, some enhancements of this plat-

form can be performed to increase robustness, reli-

ability, and portability, making this system useful for

advanced biomedical applications at the expense of

higher knowledge of electronics and programming.

4 CONCLUSIONS

A low-cost, simple and easy to implement portable

multimodal acquisition platform was developed using

an analogue circuit, an Arduino MEGA ADK and a

mobile platform. The developed platform was able to

acquire different electrophysiological signals, such as

ECG, EMG, EEG, and EOG, by changing the low-

pass and high pass ﬁlters’ cut-off frequencies and am-

pliﬁcation gain.

Two further developments to increase portability

and usability of the acquisition platform were fore-

seen. Firstly, the modiﬁcation of the design of the

analogue platform in order to use it as an Arduino

shield. This modiﬁcation allows the user to add ex-

tra analogue acquisition platforms, up to 15 additional

boards for the Arduino Mega ADK, such that different

electrophysiological signals can be acquired simulta-

neously (e.g. 16 EEG channels, or 8 EEG channels

+ 8 EMG channels). Secondly, the replacement of

manual switching to digital switching. Such conver-

sion allows the user to digitally control the acquisition

parameters without physical interaction, allowing for

the abstraction of the electronics, and enhancing us-

ability.

The developed platform is ideal for researchers,

developers and hobbyists, as it is portable, low-cost,

easily adaptable to acquire various physiological sig-

nals, and scalable/customizable in order to acquire a

larger number of channels. Due to its characteristics,

the developed platform is suitable for application de-

velopment in the ﬁelds of physiological monitoring,

human-computer interaction, and perceptual comput-

ing.

ACKNOWLEDGEMENTS

Research supported by Fundao para a Ciłncia e

Tecnologia (FCT) and Ministrio da Ciłncia e Edu-

cao (MCE) Portugal (PIDDAC) under grant PEst-

OE/SAU/UI0645/2011.

REFERENCES

Alive Technologies (2013). Alive website.

http://www.alivetec.com/.

Andreoni, G., Parini, S., Maggi, L., Piccini, L., Panﬁli, G.,

and Torricelli, A. (2007). Human machine interface

for healthcare and rehabilitation. In Vaidya, S., Jain,

L., and Yoshida, H., editors, Advanced Computational

Intelligence Paradigms in Healthcare-2, volume 65

of Studies in Computational Intelligence, pages 131–

150. Springer Berlin Heidelberg.

Bin Ambar, R. (2012). Multi-sensor arm rehabilitation

monitoring device. Biomedical Engineering, Interna-

tional Conference on, pages 424–429.

Burr-Brown Corporation (1998). BB INA114 Precision In-

strumentation Ampliﬁer Datasheet.

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

Kim, J. (2004). A new means of HCI: EMG-MOUSE, Sys-

tems, Man and Cybernetics. IEEE International Con-

ference on, 1:100–104.

National Instruments (2013). Biomedical engineer-

ing education portal. http://www.ni.com/white-

paper/5593/en/.

PLUX wireless biosignals (2013). Plux website.

http://www.plux.info/.

Roy, S. (2009). A Combined sEMG and Accelerometer

System for Monitoring Functional Activity in Stroke.

IEEE Transactions on Neural Systems and Rehabili-

tation Engineering, 17:585–594.

Shimmer (2013). Shimmer website. http://www.shimmer-

research.com/.

Webster, J., editor (2010). Medical Instrumentation Appli-

cation and Design. John Wiley & sons.

West, D. (2012). How Mobile Devices are Transforming

Healthcare. Issues in Technology Innovation, no.18:1–

14.

Wiring (2013). Wiring framework website.

http://wiring.org.co/.

AMultimodalLow-costPlatformforAcquisitionofElectrophysiologicalSignalsInterfacingwithPortableDevices

Yin, Y., Fan, Y., and Xu, L. (2012). Emg and epp-integrated

human-machine interface between the paralyzed and

rehabilitation exoskeleton. Information Technology in

Biomedicine, IEEE Transactions on, 16(4):542–549.

Zheng, X. (2009). A portable wireless eye movement-

controlled Human-Computer Interface for the dis-

abled. International Conference on Complex Medical

Engineering, pages 1–5.

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems