IEETA BRAIN COMPUTER INTERFACE

Towards a Rapid Prototyping and Multi-Application System

Virgílio F. Bento, Filipe M. Silva

and João P. Silva Cunha

Institute of Electronic Engineering and Telematics of Aveiro (IEETA), University of Aveiro, 3810-193 Aveiro, Portugal

Dep. Electronics, Telecommunications and Informatics (DETI), University of Aveiro, 3810-193 Aveiro, Portugal

Keywords: Brain-machine interfaces, Electroencephalography (EEG), Mu rhythms, Motor imagery.

Abstract: Recent advances in computer hardware and signal processing assert that controlling certain functions by

thoughts may represent a landmark in the way we interact with many output devices. This paper exploits the

possibility of achieving a communication channel between the brain and a mobile robot through the

modulation of the electroencephalogram (EEG) signal during motor imagery tasks. A major concern was

directed towards designing a generalized and multi-purpose framework that supports rapid prototyping of

various experimental strategies and operating modes. Preliminary results of brain-state estimation using

EEG signals recorded during a self-paced left/right hand movement task are also presented. The user

successfully learned to operate the system and how to better perform the motor-related tasks based on

outcomes produced by its mental focus.

1 INTRODUCTION

In recent years, the appealing idea of a direct

interface between the human brain and an artificial

system – called Brain Computer Interface (BCI) –

has motivated a growing community of researchers

(McFarland, 2006). The conceptual approach is to

model the brain activity variations and map them

into some kind of actuation or command over a

target output (e.g., a computer interface or a robotic

system). Continuing advances in a number of fields

have supported the thesis that the concept is viable,

although a significant research and development

effort has to be conducted before these technologies

enter routine use. Nowadays, the principal reason for

the BCI research is the potential benefits to those

with severe motor disabilities (e.g., amyotrophic

lateral sclerosis, brainstem stroke or severe cerebral

palsy) (Kubler et al., 2005).

The combination of these reasons led the authors

to gradually start a project aiming to initiate a long-

term multidisciplinary research by combining

developments in relevant fields, such as cognitive

neuroscience, brain imaging, pattern recognition,

electronics and computing. The ultimate goal is to

promote the involvement of under and post-graduate

students in international level stages such as

competitions of the grade of BCI2000 (Schalk et al.,

2004) and similar. In the middle-term, the main

scope has been the design and development of a BCI

system to exploit the benefits of a closer interplay

between neurosciences and robotics. A hypothesis is

that brain-actuated control of a robotic device will

improve human-robot interfaces and facilitates robot

programming.

Bearing this in mind, this paper presents the first

steps towards the development of an EEG-based

BCI system that analyzes the brain activity of a

subject, tries to find out its intentions and generates

output commands controlling an appropriate output

device. The relevant feature of this implementation

includes the movement imagery, based on the mu

rhythms, as control strategy to command a Khepera

mobile robot (Pineda et al., 2000). A major concern

was directed towards designing a generalized

framework that supports rapid prototyping of

various experimental strategies and operating

modes. From the current stage of development,

based on Matlab and Simulink, it stands out the high

versatility of implementation that allows the

comparison of different spatial filters, spectral

analysis algorithms and signal processing methods.

Although some issues are yet to be addressed, our

BCI is already mature for practical experiments and

to obtain the first conclusions on the potential of the

proposed solutions.

337

F. Bento V., M. Silva F. and P. Silva Cunha J. (2009).

IEETA BRAIN COMPUTER INTERFACE - Towards a Rapid Prototyping and Multi-Application System .

In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing, pages 337-344

DOI: 10.5220/0001550003370344

 SciTePress

The remainder of the paper is organised as

follows. Section 2 reports previous studies that

produced valuable insights, focusing on strategies

employed and the potential of mu rhythms in actual

BCIs. Section 3 describes the design options, the

developed tools and the applications of the IEETA

BCI. Section 4 presents preliminary results of brain-

state estimation using EEG signals recorded during a

self-paced left/right hand movement tasks, while

controlling the Khepera mobile robot. Section 5

concludes the paper and outlines the perspectives of

future research.

2 RELATED WORK

Over the past decade, several working BCI systems

have been described in the literature (Jos et al.,

2004, Pfurtscheller et al., 2006, Wolpaw et al.,

2003). These systems use a variety of signal

acquisition methods, experimental paradigms,

pattern recognition approaches and output interfaces,

requiring different types of cognitive activity. Most

solutions rely on brain electrical activity measured

through electroencephalogram (EEG). Despite their

poor spatial resolution, this non-invasive technique

has proven to be a useful and practical tool in

experimental research, mainly due to fast recording,

easy subject preparation and reduced equipment

required. Furthermore, the relationship between

EEGs and brain function is well documented in the

literature.

One type of BCI that has been extensively

studied derives information either from the user’s

movements or the imagination of movement. These

movement-based BCIs recognize changes in the

human mu rhythm, which is an EEG oscillation

recorded in the 8-13 Hz range from the central

region of the scalp overlying the sensoriomotor

cortex (Kuhlman, 1978, Pfurtscheller and Lopes da

Silva, 1999). This activity is most pronounced when

subjects are at rest, but not planning to initiate

voluntary movement. At least a second before

subjects initiate voluntary movement, the mu rhythm

over the hemisphere contralateral to the region

moved shows a decrease in amplitude and thus

power. This attenuation becomes more symmetric

over both hemispheres as subjects actually initiate

the movement and remains until shortly after the

movement is initiated. Mu activity returns to

baseline levels within a second after movement is

initiated and may briefly increase above baseline

(Fatourechi et al., 2007, Pineda et al., 2000). These

activity dependent changes in mu activity have also

been called Event Related Desynchronization (ERD)

and Event Related Synchronization (ERS) by

Pfurtscheller and his co-workers (Pfurtscheller and

Lopes da Silva, 1999).

The mu rhythm thus has potential for BCIs for

many reasons. It is present in nearly all adults,

including many individuals with motor disabilities.

Since it’s easy to train in subjects while they are

awake with eyes open (Kuhlman, 1978, Pfurtscheller

and Lopes da Silva, 1999) and can be affected by

visual and imagined input (Muthukumaraswamy et

al., 2004, Pineda, 2005, Hoshi and Tanji, 2006), it

may be possible for users to learn to use a mu

rhythm based BCI system by means of a multiplicity

of stimuli and cognitive strategies. Therefore, the

pattern recognition may be simple: detecting power

changes can be fruitful in BCI design. Finally, the

mu rhythm can be modulated in either or both

hemispheres (Pfurtscheller and Lopes da Silva,

1999, Pineda et al., 2000).

These observations led us to utilise motor

imagery as control strategy to achieve asymmetrical

electrocortical responses and to use left-right

differences in the sensoriomotor EEG to provide the

required control options of a two dimensional

environment.

3 IEETA BCI: DESIGN AND

OPERATION

The IEETA BCI system was conceived having in

mind the application to which it will be applied:

control of a robotic device. Moreover, the BCI tools

are optimized for each individual user by providing

him with a training period in the presence of

feedback. Indeed, present BCI systems depend on

user control of brain electric activity, such as

amplitude in a specific frequency band (e.g., mu

rhythms) in EEG recorded over a specific cortical

area (e.g., sensoriomotor cortex).

This section describes the relevant development

steps since they will reveal much about the

problems, challenges and tradeoffs of the complete

BCI prototype, as well as guide the selection of

alternative designs.

Matlab

and Simulink

were the platform chosen to

develop the BCI system. This choice is justified by

the fact that it’s the widely tool used in signal

processing and classification. These two areas

represent the main base of any BCI system.

Due to the flexibility of Matlab programming, all

the algorithms are written in Matlab code whereas

BIOSIGNALS 2009 - International Conference on Bio-inspired Systems and Signal Processing

338

the driver of the acquisition hardware was created

employing C++, using a wrapper to integrate it.

Another important aspect of developing any

system in Matlab

is that, this can be done in a high

abstraction level, letting the developer focus on the

problems of the system and not on the tools that

support it. Each new module can be chosen from

Simulink library. Simulink library provides many

Signal Processing modules for direct

implementation. When a new module is needed with

a specific task, it can be implemented using an S-

Function.

S-Functions uses a callback method to perform

each task that operate flawlessly within the main

system, this way we guarantee a temporal

performance, essential in online analysis.

3.1 EEG Recording

EEG signals were recorded from eight scalp

electrodes placed over central (C3, C4), frontal (F3,

F4) and parietal (P7, P8, P3 and P4) locations

according to the 10-20 international system and

referred to a linked-ear reference (Jasper, 1952).

Using this spatial location we assume that a generic

motor imagery task can relate to different subsets of

cortical areas activation, resulting in the excitability

of different regions such as the Premotor Cortex, the

supplementary motor area, the primary motor cortex

and the sensoriomotor cortex (Porro et al., 1996,

Lacourse et al., 2005, Lotze et al., 1999).

Figure 1: Spatial Location of the EEG electrodes over the

frontal, central and parietal areas.

The BCI user utilizes a portable EEG acquisition

system (Figure 2) with a sampling rate of 256 Hz.

This EEG system imposes a maximum number of 8

acquisition electrodes. This difficulty becomes

secondary by the advantage of using a portable

system that represents minor power consumption,

essential to the implementation of a future

ambulatory system. Another advantage resulting

from using a small number of electrodes is the

smooth online performance of the BCI system, as

trivial to assume that a higher number of EEG

electrodes results in a higher number of signals to

process that if not equipped with fast, and expensive,

hardware system could inflict slowdowns in the real-

time processing.

Figure 2: TrackIt system: ambulatory acquisition of EEG

signals (from LifeLines Ltd).

3.2 Control Paradigm

According to (Arroyo et al., 1993) mu waves are

almost constantly present when the subject is relaxed

and are heavily suppressed when the subject

performs a motor (imagery or real) task exciting the

contralateral side, i.e. mu waves disappear over the

left brain hemisphere when the right hand is moved

and vice versa. In addition, humans can learn to

modify the amplitude of the mu rhythm after

prolonged training (on the order of weeks or

months) with the help of mental activities alone.

This is the starting point of the systems described in

(Fabiani et al., 2004, Guger et al., 2001, Wolpaw et

al., 2000). Their idea is to take that amplitude –

measured only by one pair of electrodes – and

translate it into (one-dimensional) cursor movement.

Using a simple computation, it estimates the FFT of

the ongoing EEG (“online”), taking the square root

of the power associated with the mu rhythm

frequency range, and comparing the resulting value

with adaptable voltage ranges. This leads to a trivial

quantification (or classification) encoding the mu

rhythm amplitude, and is directly translated into the

movement of a cursor on a feedback video screen,

where low amplitudes move the cursor down, while

high amplitudes move it up (the magnitude of the

upward or downward movement being part of the

quantification, too).

Although the accuracy that can be achieved with

this system is relatively high (in up to 95% of all

cases, the system really does what the user wants it

to do), it cannot serve as the basis for a practical

IEETA BRAIN COMPUTER INTERFACE - Towards a Rapid Prototyping and Multi-Application System

339

device, since it is very slow. To cope with this

problem, the solution relies in the multi-dimensional

control provided by distinct EEG features. The idea

was to record the EEG at two different sites on the

scalp, hoping that subjects would be able to learn to

intentionally vary the two mu rhythm amplitudes

simultaneously and independently. The projected

system was much more practicable, since it

possessed the potential to “emulate” a computer

mouse to a certain extent. However, despite the

correctness of the “independence hypothesis”, the

outcome was not much more than a laboratory

phenomenon, because the achieved accuracy did not

exceed 70%. Kostov and Polak, (2000) have also

shown that

subjects can move a virtual object up and

down on a computer screen by issuing various

mental activities during a time window delimited by

the pressing of two manual switches.

3.3 Signal Processing and Classifier

Each EEG raw signal was filtered in the 8-13 Hz

band by a 20-order Band-Pass Butterworth Filter.

After filtering, the signal was segmented in

customized blocks of 128 samples (0.5s). Each

signal block was transformed by a surface Laplacian

in F3, C3, P3 and P7 for the left Hemisphere and F4,

C4, P4 and P8 for the Right Hemisphere (see Figure

1).

The power spectrum Estimation was performed

using the Yule-Walker Method (Kay, 1998). Each

vector (8 channels with 128 frequency components)

is then analysed by the ERD (Event Related

Desynchronization) block, which verify for a

specific frequency band if the ERD is confirmed.

There are two ERD modules, one for each

hemispheric signal.

The Classifier (

Figure 3

) has two inputs, one for

each ERD block. It was implemented by mean of a

decision tree, so if only the Right Hemisphere signal

verifies the ERD, the classifier output is “LEFT”. If

only the Left Hemisphere signal verifies the ERD,

then the classifier output is “RIGHT”. If both the

signals verify the ERD then the output is

“FORWARD”. If neither one of the signals verifies

the ERD, then the output is “STOP”. In this way,

with only 2 mental tasks, we obtain 4 possible

control orders.

The majority of the BCI systems implemented

devote a great part of the system resources and

development time in the classifier. Nonetheless, a

very important part of any BCI system, the authors

of this BCI system realize that the importance of the

classifier can easily be minored if we get better and

enhanced features from the mental tasks carried out.

This can be achieved selecting the best electrodes

set-up, signal processing methods and new EEG

processing techniques such as ERD, ERS, P300 and

others yet to be found.

Figure 3: Binary Tree Classifier perceiving four possible

states based on the two ERD features.

3.4 BCI Output

The mobile robot used as the control application was

a small Khepera. The Khepera robot (5.7 cm

diameter), is a two-wheeled vehicle with 8 infrared

sensors representing the ideal analogy for a

wheelchair.

Two other graphical applications where

developed, BioFeedback I and BioFeedback II, both

mainly used in user training.

BioFeedback I provides the feedback in a

graphical way, displaying one of four images

according to the classifier output - “RIGHT”,

”LEFT”, ”FORWARD”, ”STOP”.

Figure 4: Training Module where each image correlates

with a classification output (BioFeedback I).

Feedback is provided by means of coloured

arrows, one for each mental task, for easily

recognition of the system output.

The other graphical application (BioFeedback II)

is also used in the online test of the system, and it’s a

classical application of the BCI systems. The goal of

this application is to place the cursor on one of three

possible areas (Figure 5). The cursor is controlled by

BIOSIGNALS 2009 - International Conference on Bio-inspired Systems and Signal Processing

340

means of the classifier outputs – “RIGTH”, “LEFT”

and “FORWARD”.

Figure 5: Second Graphical Application proposed for user

training (Biofeedback II).

The final application is the control application

depicted before. This output module, depending on

what the classifier output is, controls the velocity on

each wheel. The Khepera robot provides 8 infrared

sensors, this enables, in integration with the system,

to detect when the robot reaches a wall, allowing in

this case only two possible movements – forward

and the opposite wall direction movement.

The control communication with the Khepera

Robot is performed through RS232 protocol and

using the communication toolbox available from

Matlab

in real-time performance.

Figure 6: Control output.

4 EXPERIMENTAL RESULTS

4.1 Users’ Protocol

During a session, the subject sat in front of a screen,

and was asked to remain still (Figure 7a). Scalp

electrodes (see montage in Figure 1) acquired 8 EEG

channels, each one of them referenced to an

electrode in the right ear lobe.

The Experimental Procedure was designed in 3

steps. First we acquire the Baselines through the

BCI_GetBase sub-system that is essentially the BCI

system without the output module, sharing all the

main modules and its configuration.

The Baseline is the core of any ERD based

system. We acquire 3 Baselines, for 3 mental tasks:

imagery finger tapping; imagery open/close hand;

imagery free hand movement. We choose three

different baselines, due to the fact that the definition

of baseline is related with the definition of No

Control (NC) state, that is when there is no

intentional control, e.g., during periods of thinking

or monitoring that do not correspond to the cortical

excitation achieved by the two motor imagery tasks

that are asked to the user. NC control support is

necessary for most types of machine or device

interactions where frequent actions are spaced by

periods of inaction. In this way, and using the

contralateral propriety of cortical activation, the

baseline that shares all the common underlying brain

activity of the motor imagery tasks asked to the user,

is its opposite task, e.g., if we are analysing the

presence of a right motor imagery task we compare

the EEG signal with the recorded baseline for the

left motor imagery task performed before, detecting

in this way all the difference in signal amplitudes

related to the modulation of the mu rhythms.

After the record of the baselines the subject was

given three possible conditions to control the

Khepera robot.

1) Rest: the subject, sat in a comfortable chair,

was asked to relax as much as possible and think of

nothing in particular.

2) Self-generated movements: subject is asked

to move each finger;

(3) Imagination: subjects were instructed to

image performing the self-generated movement

without actually doing it.

Only the results achieved using the motor

imagery tasks were explored in validating the

preliminary system results.

4.2 Data Analysis

The degree of mu suppression occurring during the

imagery of movement can be expressed as a relation

with the peak power value at rest and typically

shows an average decrease that depends on the level

and “quality” of attention.

After we got all the 3 Baselines for each

ipsilateral hemisphere (Figure 8 and 9), the system

automatically chooses what is the Baseline that

offers the best results. The notion of a good Baseline

IEETA BRAIN COMPUTER INTERFACE - Towards a Rapid Prototyping and Multi-Application System

341

Figure 7: Experimental Environment.

is associated with the absence of involuntary

desynchronization or artefacts.

With the baselines for each mental task, the next

step is to train the user in the system. To promote a

fast learning on how to better control the cerebral

rhythms it’s proposed to the user the BioFeedback I

and II usage for a typical period of 30 minutes that

can be (desirably) increased depending on the user

available time

Figure 8: Baselines for: imagery finger tapping, imagery

close/open Hand and imagery free hand movement

respectively (Ipsilateral Movement - Right Hand, Right

Hemisphere).

Finally, after the acquisition of the Baseline and

the (ideal) extensive training of the user, we analyse

the user performance in controlling the Khepera

Robot. In this step, all the BCI system configuration

is set up specifically for the user, this is easily

achieved due to the rapid-prototyping characteristic

we included in the system.

These distinctive features of any BCI system

relate to the importance of the user training that after

a few hours of train could certainly boost the system

performance.

The control of the Khepera robot was done in a

free environment. The user had to move it to two

possible areas as shown on Figure 7b using the

motor imagery tasks depicted before.

Figure 9: Baselines for: imagery finger tapping, imagery

close/open Hand and imagery free hand movement

respectively (Ipsilateral Movement - Left Hand, Left

Hemisphere).

Examples of the ERD achieved for both the

“Right” and “Left” areas in the contralateral spatial

filtered electrodes (C3 and C4) are shown below.

Figure 10: ERD for the spatial filtered signal at C3 (Right

Area target).

Figure 11: ERD for the spatial filtered signal at C4 (Left

Area target).

The user achieved a 70% correct classification

rate for each direction in a total of 7 trials. The

BIOSIGNALS 2009 - International Conference on Bio-inspired Systems and Signal Processing

342

classification rate has evolved during the sequential

trials. Through time, the user developed own ways

to better control the robot. This fact implies that an

extensive training is essential to obtain very good

results.

Nonetheless the lack of extent of the online

results, as referred earlier, these results are

preliminary and mainly used to validate the system

as a promising BCI structure.

5 CONCLUSIONS

We have shown the development of a multi-

application BCI system from the source to the

output. Using rapid-prototyping tools we ensured an

efficient time-progress window of development.

This also represents a proficient ability to perform

several optimizations quickly and in highly

integration with the structural hierarchy of the BCI

system implemented.

An important aspect about this BCI system is its

modular structure that allows it to perform a

different function just by creating a new output

module. This modular structure also improves the

time-progress window due to its parallel

development and optimization suited for each

module individually.

This system represents a new BCI platform

developed using efficient and widely used signal

processing tools ensuring in this way a maximum

focus on the project itself and not on the

development tools that support it.

In spite of being in an inborn stage this system

provided encouraging results in the preliminary

online test made. The user demonstrated satisfaction

in using the system and confirmed its controllability.

More and extended online tests are needed to

perform increasable optimizations, nonetheless, this

process is already on course in two different BCI

areas (Control and Bio-Encryption), that due to the

system modularity interchange results and possible

optimization between them in order to achieve the

best possible results.

ACKNOWLEDGEMENTS

The authors would like to thank Luis Paula for the

voluntary testing of the system and its valorous

commentaries.

Partly supported by "EpilBI - Epileptogenic

focus localization in a 3D multimodal Brain Imaging

system." (POSC/EEA-CPS/60977/2004 – FCT)

project.

REFERENCES

Arroyo, S., Lesser, R. P., Gordon, B., Uematsu, S.,

Jackson, D. & Webber, R. (1993) Functional

significance of the mu rhythm of human cortex: an

electrophysiologic study with subdural electrodes.

Electroencephalogr Clin Neurophysiol, 87, 76-87.

Fabiani, G. E., Fabiani, G. E., Mcfarland, D. J., Wolpaw,

J. R. & Pfurtscheller, G. A. P. G. (2004) Conversion of

EEG activity into cursor movement by a brain-

computer interface (BCI).Neural Systems and

Rehabilitation Engineering, IEEE Transactions on

[see also IEEE Trans. on Rehabilitation Engineering],

12, 331-338.

Fatourechi, M., Birch, G. E. & Ward, R. K. (2007) A self-

paced brain interface system that uses movement

related potentials and changes in the power of brain

rhythms. J Comput Neurosci, 23, 21-37.

Guger, C., Guger, C., Schlogl, A., Neuper, C.,

Walterspacher, D. A. W. D., Strein, T. A. S. T. &

Pfurtscheller, G. A. P. G. (2001) Rapid prototyping of

an EEG-based brain-computer interface (BCI)Neural

Systems and Rehabilitation Engineering, IEEE

Transactions on [see also IEEE Trans. on

Rehabilitation Engineering], 9, 49-58.

Hoshi, E. & Tanji, J. (2006) Differential involvement of

neurons in the dorsal and ventral premotor cortex

during processing of visual signals for action planning.

J Neurophysiol, 95, 3596-616.

Jasper, H. H. (1952) Electroencephalography. Pediatrics,

9, 786-7.

Jos, Del, R. M., Fr, D, Ric, R., Josep, M. & Wulfram, G.

(2004) Brain-actuated interaction. Elsevier Science

Publishers Ltd.

Kay, S. M. (1998) Modern Spectral Estimation: Theory

and Application., Englewood Cliffs, Prentice-Hall.

Kostov, A. & Polak, M. (2000) Parallel man-machine

training in development of EEG-based cursor control.

IEEE Trans Rehabil Eng, 8, 203-5.

Kubler, A., Nijboer, F., Mellinger, J., Vaughan, T. M.,

Pawelzik, H., Schalk, G., Mcfarland, D. J., Birbaumer,

N. & Wolpaw, J. R. (2005) Patients with ALS can use

sensorimotor rhythms to operate a brain-computer

interface.

Kuhlman, W. N. (1978) Functional topography of the

human mu rhythm. Electroencephalogr Clin

Neurophysiol, 44, 83-93.

Lacourse, M. G., Orr, E. L., Cramer, S. C. & Cohen, M. J.

(2005) Brain activation during execution and motor

imagery of novel and skilled sequential hand

movements. Neuroimage, 27, 505-19.

Lotze, M., Montoya, P., Erb, M., Hulsmann, E., Flor, H.,

Klose, U., Birbaumer, N. & Grodd, W. (1999)

Activation of cortical and cerebellar motor areas

IEETA BRAIN COMPUTER INTERFACE - Towards a Rapid Prototyping and Multi-Application System

343

during executed and imagined hand movements: an

fMRI study. J Cogn Neurosci, 11, 491-501.

Mcfarland, D. (2006) BCI Communication Devices.

WTEC Workshop on Brain Computer Interface (BCI)

Research.

Muthukumaraswamy, S. D., Johnson, B. W. & Mcnair, N.

A. (2004) Mu rhythm modulation during observation

of an object-directed grasp. Brain Res Cogn Brain Res,

19, 195-201.

Pfurtscheller, G. & Lopes da Silva, F. H. (1999) Event-

related EEG/MEG synchronization and

desynchronization: basic principles. Clin

Neurophysiol, 110, 1842-57.

Pfurtscheller, G., Pfurtscheller, G., Muller-Putz, G. R.,

Schlogl, A., Graimann, B. A. G. B., Scherer, R. A. S.

R., Leeb, R. A. L. R., Brunner, C. A. B. C., Keinrath,

C. A. K. C., Lee, F. A. L. F., Townsend, G. A. T. G.,

Vidaurre, C. A. V. C. & Neuper, C. A. N. C. (2006) 15

years of BCI research at graz university of technology:

current projects.Neural Systems and Rehabilitation

Engineering, IEEE Transactions on [see also IEEE

Trans. on Rehabilitation Engineering], 14, 205-210.

Pineda, J. A. (2005) The functional significance of mu

rhythms: translating "seeing" and "hearing" into

"doing". Brain Res Brain Res Rev, 50, 57-68.

Pineda, J. A., Allison, B. Z. & Vankov, A. (2000) The

effects of self-movement, observation, and

imagination on mu rhythms and readiness potentials

(RP's): toward a brain-computer interface (BCI). IEEE

Trans Rehabil Eng, 8, 219-22.

Porro, C. A., Francescato, M. P., Cettolo, V., Diamond, M.

E., Baraldi, P., Zuiani, C., Bazzocchi, M. & Di

prampero, P. E. (1996) Primary motor and sensory

cortex activation during motor performance and motor

imagery: a functional magnetic resonance imaging

study. J Neurosci, 16, 7688-98.

Schalk, G., Schalk, G., Mcfarland, D. J., Hinterberger, T.,

Birbaumer, N. A. B. N. & Wolpaw, J. R. A. W. J. R.

(2004) BCI2000: a general-purpose brain-computer

interface (BCI) system

BCI2000: a general-purpose brain-computer interface

(BCI) system. Biomedical Engineering, IEEE

Transactions on, 51, 1034-1043.

Wolpaw, J. R., Wolpaw, J. R., Mcfarland, D. J. &

Vaughan, T. M. (2000) Brain-computer interface

research at the Wadsworth Center.Rehabilitation

Engineering, IEEE Transactions on [see also IEEE

Trans. on Neural Systems and Rehabilitation], 8, 222-

226.

Wolpaw, J. R., Wolpaw, J. R., Mcfarland, D. J., Vaughan,

T. M. & Schalk, G. A. S. G. (2003) The Wadsworth

Center brain-computer interface (BCI) research and

development program.Neural Systems and

Rehabilitation Engineering, IEEE Transactions on

[see also IEEE Trans. on Rehabilitation Engineering],

11, 1-4.

BIOSIGNALS 2009 - International Conference on Bio-inspired Systems and Signal Processing

344