Non-invasive Brain Computing Interface

Technology Discernment
Sagri Sharma, Krishnanjan Bhattacharjee, Sailandra Harsha, Pavan Kurariya, Anuradha Tomar, Priyanka
Pawar, Dipali Shinde, Lina Rani Nath, Atish Bundhe, Pooja Rai, Priyanka Jain, Karimullah Shaikh,
Anuradha Lele, Ajai Kumar, Hemant Darbari
[email protected]

Abstract - Brain Computer Interfacing (BCI) system has the

potential to enable people to use their brain generated
thoughts to control and communicate with computer and
other communication devices (e.g. Smart Phones,
NeuroPhone etc.) or other medical devices through
commands for disabled. Therefore, it can be called a way to
control external world with unspoken communication and
control.
BCI system can be used as an assistive technology that caters
as a communication channel capturing thoughts, especially
for people suffering from severe motor inhibitory diseases,
through control of voluntary variations in brain waves as
their brain is still intact. Apart from that normal people can
also use this thought-to-device and communication approach
for various application from affective computing to military
usage, smart device control etc.
The challenges to design such system involve low signal-tonoise ratio (SNR) due to artifacts present in
Electroencephalography (EEG) and low accuracies of
classifiers, to name a few. Also since brain wave patterns
vary person to person, it poses challenge for proper use of
algorithm usage and implementation as well as training to
normalize to form definite thought-to-signal simulation for
further application in computer domain including Natural
Language Processing (NLP)
This report describes the principle of brain-computer
interface (BCI) using independent, non-invasive and
asynchronous technique, based on recognition of subjectspecific EEG patterns.
The brain waves captured by EEG device is processed and
interpreted to communicate and control external devices as
per the intention of user. BCI provide unspoken and muscle
independent communication and control in form of
commands; character-by-character basis and characters
transformed.
Keywords- Brain Computing Interface, EEG, ECoG,
Augmented Communication, non-invasive BCI
Corresponding Author: Sagri Sharma, AAI Group Email ID:
[email protected]

1.

INTRODUCTION

BCI systems has a broad objective to create control

level and communication level brain generated
thoughts-controlled application that is activated by
signals measured directly from an individuals brain. It
is assumed that such brain-signal interpretation system
will allow an individual with a severe disability to have
an effective control of devices such as assistive
appliances, computers, and neural appendages in natural
environments as one of the applications. This type of
direct-brain thought processing interface would increase
an individuals independence to communicate/control to
external world where active intervention or sensory
organ usage is not viable, leading to a radically
improved quality of life and reduce costs.
This trailblazing technology is a powerful
communication and control option in the interaction
between users and systems through thoughts. The
principle of BCI will be used in various areas of medical
science like neuroscience, neurology, neuro-engineering
as they can monitor the acting brain in real time and in
the real world; therefore, scientists understand the role
of functional networks during behavioral tasks. The uses
for this technology are almost limitless, once the basic
mechanism of converting thoughts to computerized or
robotic action is achieved.
The outputs/commands of the system can be mounted
on mobile controlled machines, computer mouse
control, lie-detection, alertness monitoring, neurofeedback,
eye
tracking,
neuromarketing,
communication systems, and machine learning
procedures. The capability of providing substantial
independence due to universal, effective and efficient
interface has greatly improved quality of life for
severely disabled persons can be achieved in
communicating and monitoring. Some of the nonmedical area where BCI may be integrated seem

frivolous, such as the ability to control a smart phone

and video game by thought. It will enrich the gaming
and entertainment experience in games tailored to the
users affective state immersion, flow, frustration,
surprise, and so on. The prospects are bright for rapid
growth of BCI in non-medical areas. This is especially
true of gaming, with its large economic impact. The line
between medical and non-medical usage of BCI is likely
to be minimal, but a non-medical application might be
the optimized presentation for learning content.
Future interfaces will be able to understand and
anticipate the users state and intentions. One
observation is that the coordination of medical and nonmedical BCI research efforts is vital. Such alignment
could produce a shared roadmap and research agenda
that would benefit both areas. In pursuit of such an ideal
interface, the efforts of the current research are to have
the feasibility of utilizing electrical brain potentials to
directly communicate to devices such as a personal
computer system. To enable understanding of the
system, a scenario based approach is adapted to show
the capabilities of BCI.
BCI enabled system control computer cursor via mental
commands for a handicapped person while he thinks
about various movements of the cursor, as one of its
applications. Users can gain enough control over a
cursor to draw a circle, access computer programs with
enough practice. It could theoretically be expanded to
allow users to communicate through mind-typing
without keyword expressing their thoughts. If one thinks
of controlling a device with remote control is
convenient, system will have the potential for the user
changing TV channels for instance, with thoughts.
2.

TECHNOLOGICAL/RESEARCH TRENDS
(INTERNATIONAL/NATIONAL)

BCI. BCIs have a wide range of applications that can be

beneficial for users with and without medical indication.
In India, Centre For Neuroscience, Indian Institute of
Science, Bangalore and National Brain Research Center
(NBRC) are working in this niche area.
3.

A.

Device control

One of the driving forces behind the development of

BCIs was the desire to give users who lack full control
of their limbs access to devices and communication
systems. Under these circumstances, users can be
benefitted from a device even if it has limited speed,
accuracy and efficiency. The generation of control
signals and the extraction of these from the brain signals
may cause latency in the control loop of several
hundreds of milliseconds or more that makes smooth
closed-loop control impossible. The introduction of a
form of shared control between the device and the user
in which the latter gives more high level, open-loop
commands may overcome the latency issue. However, it
is possible, that healthy users could for limited
application scenarios also benefit from either
additional control channels or hands-free control.
Examples include drivers, divers, and astronauts who
need to keep their hands on the steering wheel, to swim,
or to operate equipment. Brain-based control paradigms
are developed for these applications in addition to, for
instance, voice control. This field has a strong body of
research work on single task situations and for patient
users but lacks data on control in multi-task
environments. The direct societal impact will be limited,
although a spin-off to medical applications may be
anticipated.
B.

University of California Los Angeles (UCLA) started

with the research on BCIs in the 1970s under a grant
from the National Science Foundation. The papers
published as a result of the researches defined brain
computer interface. The field of BCI research and
development has focused on neuro-based applications
that aim at restoring damaged hearing, sight and
movement. The biggest impediment to BCI technology
at present is the lack of a sensor modality that provides
safe, accurate and robust access to brain signals. It is
conceivable or even likely; however, that such a sensor
will be developed within the few years. The use of such
a sensor should greatly expand the range of
communication functions that can be provided using a

SALIENT AREAS OF BCI APPLICATIONS IN GLOBAL SCENARIO

Gaming and entertainment

The entertainment industry with the introduction of 3D

movies at home and gesture-based game controllers is
the front runner in introducing BCI concepts and
prototypes. Over the past few years, new games have
been developed that are exclusively for use with a EEG
headset by companies like Neurosky, Emotiv, Uncle
Milton, MindGames, and Mattel. Additional, there is a
general conviction that experiences of existing game
and entertainment systems are enriched, intensified
and/or extended by using BCIs , for example by
tailoring games to the affective state of the user. For
instance, several game designers linked mental states to
the popular games so that the appearance of an

incarnation is no longer controlled through the

keyboard, but reflects the mental state of the gamer. The
first (mass) application of non-medical BCIs may
actually be in the field of gaming/entertainment where
the first stand-alone examples came to the market in
2009 and a broadening to console games may follow
soon. We expect that successful applications will be
based on state monitoring and not on the user directly
controlling the game. Although BCIs for gaming were
suggested a decade ago, the research basis is still small
but growing rapidly in a mainly application-driven
manner, rather than theoretical. The societal impact is
judged low. The price sensitivity is very high, with
important requirements with regard to ease of use,
amongst others.
C.

Cognitive improvement

Development of cognitive performance is now a reality.

Enhancement by the use of prescription drugs like
Modafinil and Ritalin has rapidly increasing among
students and professional workers. BCIs are also
considered a means to improve cognitive functioning of
healthy users. Neuro-feedback training (brain activity
alteration through operant conditioning), for instance to
improve attention, working memory, and executive
functions is relatively common among healthy users.
Another potential application area is the optimized
presentation of learning content. Although there is
currently lack of good experimental data on its effects,
the effect size is probably small and limited to specific
cognitive tasks. Generally, there may be a thin line
between medical and non-medical use of neurofeedback.
D.

Evaluation

Evaluation applications, either online (constant

monitoring) or/and offline, such as neuro-marketing and
neuro-ergonomics are two evaluation examples. Neuroergonomics has been affected by BCI: it evaluates how
well a technology matches human capabilities and
limitations. An illustrative example is the recent
research on cell phone use during driving: brainimaging results show that even hands-free or voice
activated use of a mobile phone is as dangerous as being
under the influence of alcohol during driving. The body
of evidence in this area is mainly based on fundamental
neuro-scientific studies and would benefit from a
transition to more applied studies. The societal impact is
rated low as it is merely a tool that adds to current
evaluation tools but has no direct contribution to solving
societal issues. The economic viability is potentially

high; design and evaluation are relevant for many

services and products (e.g., electronics, food, buildings).
The value added by these methods is still to be proven
over the existing evaluation means, for instance
questionnaires.
E.

Training and education

Most aspects of training are related to the brain and its

manipulability and its measurement can help to improve
training methods. Indicators for these training methods,
such as learning state and rate of progress, are useful for
automated training systems and virtual instructors, to
name a few. Currently, this application area is in a
conceptual phase with limited experimental evidence.
However, it shall have a have a good societal as well as
economic impact, for proliferation of knowledge-based
economy.
F.

Safety and security

EEG alone or combined EEG and eye movement data of

expert observers could support the detection of deviant
behavior and suspicious objects. In an envisioned
scenario an observer or multiple observers are watching
surveillance recordings to detect suspicious behavior.
EEG and eye movements might be helpful to identify
potential targets that may otherwise not be noticed.
Other applications in this area include using EEG in lie
detection and person identification, but are debatable.
This application area can be considered a niche and the
field is still in a concept development phase. Successful
applications can contribute to societal issues, for
instance regarding the safety of main transport hubs.
G.

User state monitoring

The ability to understand and anticipate users state and

users intentions is the future! For instance, automobiles
will react to driver drowsiness and virtual humans could
convince users to follow their diet. These future
implementations of so-called user system- symbiosis or
affective computing require systems to gather and
interpret information on mental states such as emotions,
attention, workload, fatigue, stress, and mistakes.
Another application field is the use of BCIs as a
research tool in neuro-scientific research. Compared to
slower, lab-based functional imaging methods it can
monitor the active brain in real time and in the real
world and thus help to understand the role of functional
networks during behavioral tasks. This is a fast growing
field in BCI applications and the extension to
physiological measures like heart rate variability and

skin conduction is in the pipeline. Better user interfaces

providing access to electronic systems and services for
all can have an important contribution to society.
4.

that use biofeedback. Yet at its beginning, adds a new

dimension in research by offering the user an additional
and independent communication channel based on brain
activity only.

BRIEF OUTLINE OF THE SCENARIO WITH SPECIFIC TECHNOLOGY

FALL-OUTS

BCI systems are divided into two mechanisms on the

basis of signal reception: Invasive and Non-Invasive.
Invasive BCI systems interact with brain directly by
penetrating electrodes into the brain cortex. In NonInvasive BCI systems, the electrodes are placed on the
scalp and the brain activity is recorded by using EEG.
Neurons communicate by passing electrical signals
along connecting fibers, and this neural activity can be
picked up on the surface of the head by the EEG
sensors. EEG signals are recorded from sensory-motor
areas during mental imagination of specific movements
and are classified on-line to produce output that can be
used to control computers or communication devices.
The scenario is depicting the use of BCI non-invasively
based the on recognition of subject-specific
electroencephalogram (EEG) patterns. These oscillatory
EEG signals are recorded during specific mental activity
called brain rhythms. Depending upon the changes in
brain rhythms there are many potentials and to acquire
the signals there are many techniques as Sensor motor
activity, Slow cortical potentials (SCPs), P300 Evoked
potential, Visual evoked potentials (VEPs), Response to
mental tasks, Activity of neural cells (ANC), Multiple
neuro-mechanisms (MNs) depending on user intention;
as mentioned in upcoming sections of this paper. The
acquired signal will undergo various phases as signal
analysis, signal pre-processing, signal enhancement,
feature extraction and signal classification. This
classified signal will be translated in the form of
commands/output.
In
the
current
set-up,
commands/output is of two types: character level and
control level.
Character level commands are based on typical BCI
event-related potential, providing muscle independent
communication on a character-by-character basis.
Control Level Commands can be integrated into
commercially available devices so that user can do text
entry through thought without physical intervention,
emailing and internet surfing options through mind. The
obtained output signals will be evaluated for different
purposes, such as cursor control, selection of letters or
words, or control of prosthesis.
The scenario proposes the setups where user is provided
with intuitive control strategies in plausible applications

5.

ARTIFACTS OF THE SCENARIO UNDER CONSIDERATION

The current BCI system fall into seven main categories,

based on the neuro-mechanisms and recording
technology as control signals are used to generate
signals.
A.

Sensor motor activity

This BCI designs use sensorimotor activity as the neural

source of Control can be divided into three subcategories:
Changes in brain rhythms (, , and ): All the brain
waves fall in different range, rhythms falls in range
of 8-12 Hz and rhythms in the range of 13-30 Hz.
Both originate in the sensorimotor cortex and are
displayed when a person is not engaged in
processing sensorimotor inputs or in producing
motor outputs. They are mostly prominent in frontal
and parietal locations. This phenomenon, called
event-related synchronization (ERS), is dominant
over the contralateral sensorimotor area and reaches
a maximum around 600 ms after movement offset.
Movement-related potentials (MRPs): MRPs are
low-frequency potentials that start about 1-1.5 s
before a movement having bilateral distribution and
present amplitude maximum at the vertex.
Other sensorimotor activities: These activities are
not restricted to a particular frequency band or scalp
location and cover different frequency ranges
usually.
B.

Slow cortical potentials (SCPs)

Slow cortical potentials (SCPs) are slow voltage shifts

which occur in the EEG in the frequency range 1-2 Hz.
Feedback training subjects can learn to voluntarily
control their SCPs.
C.

P300 Evoked potential

Infrequent or particularly significant auditory, visual, or

somatosensory stimuli, when interspersed with frequent
or routine stimuli, typically evoke in the EEG over the
parietal cortex a positive peak at about 300 ms after the
stimulus is received. This peak is called P300.

D.

Visual evoked potentials (VEPs)

VEPs are small changes in the ongoing brain signal.

They are generated in response to a visual stimulus such
as flashing lights and their properties depend on the type
of the visual stimulus.
E.

Response to mental tasks

BCI systems based on non-movement mental tasks

assume that different mental tasks (e.g., solving a
multiplication problem, imagining a 3D object, and
mental counting) lead to distinct, task-specific
distributions of EEG frequency patterns over the scalp.
F.

Activity of neural cells (ANC)

It has been shown that the firing rates of neurons in the

motor cortex are increased when movements are
executed in the preferred direction of neurons. Once the
movements are away from the preferred direction of
neurons, the firing rate is decreased.
G.

Multiple neuro-mechanisms (MNs)

BCI systems based on various neuro-mechanisms use a

combination of two or more of the above mentioned
neuro-mechanisms.
6.

SIGNAL ANALYSIS

The goal of signal analysis in a BCI system is to

maximize the signal-to-noise ratio (SNR) of the EEG or
single-unit features that carry the users messages and
commands. To achieve this goal, consideration of the
major sources of noise is essential. Noise can be from
other non-cognitive activities like eye movements, EEG
features other than those used for communication.
Numerous options are available for BCI signal
processing. Ultimately, they need to be compared in online experiments that measure speed and accuracy. The
new Graz BCI system, based on Matlab and Simulink,
supports rapid prototyping of various methods.
Different spatial filters and spectral analysis methods
can be implemented in Matlab and compared in regard
to their online performance. Signal processing methods
are important in BCI design, but they cannot solve every
problem. Enhancement can be done in signal-to-noise
ratio; they cannot directly address the impact of changes
in the signal itself. Factors such as motivation, intention,
frustration, fatigue, and learning affect the input features

that the user provides. Thus, BCI development depends

on appropriate management of the adaptive interactions
between system and user, as well as on selection of
appropriate signal processing methods.
A.

Signal pre-processing

To extract features, it is necessary to pre-process first

the data. Three steps are to be considered necessarily to
achieve this goal; Initially, Referencing should be done,
temporal filtering is later and then signal enhancement.
In the case of EEG recordings from the cortex or from
scalp, these recordings are gained using generally,
different electrodes at different positions. The brain
activity voltage which is measured by a given electrode
is a relative measure; the measurement obtained can be
compared to some another reference brain voltage
which is situated on some other site. This concludes in a
combination of brain activity at that given electrode,
brain activity at the reference site and also noise. Due to
this, the reference site should be considered such that
the brain activity at that particular site is almost zero.
Generally, the mastoids, nose and earlobes are used.
There are three referencing methods:
Common reference: The common reference
technique is widely used in BCIs. Here this method
uses one Common reference for all electrodes. The
site of this considered reference is situated at very
large distance from all electrodes. The activity
taking place at the reference site influences the
measurements at equal level, but still the differences
between electrode measurements contain all the
needed information.
Average reference: The average reference subtracts
the average of the activity at all electrodes from the
Measurements. This method is based on the
principle that the activity sums up to zero at every
moment at the whole head. Thus, the average of all
activity gives a representation of an estimate of the
activity that has taken place at the reference site.
This produces a dereferenced solution on subtracting
this average from principle. The relatively low
density of the considered electrodes with the fact
that the beneath part of the head is not taken into
consideration. It should bring some Practical
problems along with.
Current source density (CSD): The current source
density (CSD) is used in many BCIs. It is the rate of
change of current flowing into and through the scalp.
This can be depicted from EEG data, and it can be
interpreted as the difference between weighted
average of their surrounding electrodes and an
electrode. The CSD can be given by computing it
with the laplacian. The laplacian later computes the

sum of the differences between an electrode and its

neighbor electrode. This estimation is valid only
when the electrodes are in a two dimensional plane
and equally distant.
The brain signals are naturally contaminated by many
internal and external noises. They can be removed using
simple filters. The relevant information in BCIs is found
in the frequencies below 30Hz. Therefore, all noise with
higher frequencies (e.g. noise from the electrical net has
a fixed frequency of 50Hz or 60 Hz) can be removed
using FIR low pass filter. Specific frequency bands may
also be selected using FIR band pass filters.
The choice of a suitable enhancement technique is
dependent on several factors such as the recording
technology, number of electrodes, and neuromechanism of the BCI.
Spatial filters Referencing methods: The proper
selection of a spatial filter for any BCI is determined
by the location and extent of the selected brain
control signal and of the various sources of EEG or
non-EEG noise.
Surface Laplacian (SL): The SL is defined as the
2nd order spatial derivative of the surface potential.
Due to its intrinsic spatial high-pass filtering
characteristics, the SL can reduce the volume
conduction effect by enhancing the high-frequency
spatial components; therefore can achieve higher
spatial resolution than surface potentials.
Principal component analysis (PCA): The PCA is a
linear mapping that transforms a number of possibly
correlated variables into a smaller number of
uncorrelated variables called principal components.
Independent component analysis (ICA): The more
important artifacts in BCIs are generated by muscles
and eyes blink. Classical automatic methods for
removing such artifacts can be classified into
rejection methods and subtraction methods.
Common spatial patterns (CSP): The CSP is a
technique used to find the common projection matrix
that decomposes the different classes of single trial
EEG datasets, and more specifically to find spatial
structures of event-related (de)synchronization
(ERD/ERS) in a EEG context.
Common spatial subspace decomposition (CSSD):
The CSSD can extract signal components specific to
one condition from multiple MEG/EEG data sets of
multiple task conditions
Figure 1 shows, the statistical results of the study
realized concerning pre-processing methods in BCI
designs. The statistics show that 96 BCI designs that
employ signal enhancement techniques before

extracting the features from the signal, 32% use surface

Laplacian (SL), 22% use either principal component
analysis (PCA) or independent component analysis
(ICA), 14% use common spatial patterns (CSP) and
11% use common average referencing (CAR)
techniques.

Fig. 1 Comparative Study of Pre-processing methods in BCI

B. Feature extraction methods in BCI designs

This section describes briefly the common BCI features
extraction methods. Concerning the design of a BCI
system, some critical properties of these features must
be considered:
Noise and outliers: the brain signals (e.g. EEGs)
have a poor signal-to-noise ratio.
High dimensionality: in BCI systems, feature vectors
are often of high dimensionality.
Several features are generally extracted from several
channels and from several time segments before
being concatenated into a single feature vector;
Time information: BCI features should contain time
information as brain activity patterns are generally
related to specific time variations of EEG;
The brain signals are non-stationary in nature;
The brain signals are non-linear in nature;
Non-sufficient training sets: training process is time
consuming and demanding for the subjects.
There are many methods used in BCI, depending of the
type of the BCI systems. We describe some main and
specific methods like Band powers (BP), Crosscorrelation between EEG band powers, frequency
representation (FR), time-frequency representation
(TFR), Hjorth parameters, parametric modeling, inverse
model and specific techniques used for P300 and VEP
such as Peak picking (PP) and slow cortical potentials
calculation (SCPs).
Band powers (BP) features may be extracted from
the EEG signals by estimating the power distribution
of the EEG in predefined frequency bands. These
frequency bands varied between 9 and 14 Hz and
between 18 and 26 Hz. In the case of EEG
measurements the cross-correlation coefficients

between the EEG activities may be calculated to

obtain some information from comparing different
locations and different frequency bands.
Frequency representation (FR) features have been
widely used in signal processing because of their
ease of application, computational speed and direct
interpretation of the results.
Time-frequency representation (TFR) due to the
non-linearity and non-stationary nature of the EEG
signal, the classical methods based on Fourier
transform (FT) are, in general, not efficient for
feature extraction because they obtained features do
not provide any time domain information, i.e. these
features do not analyze the time-varying spectral
content of the signals. .
Inverse model have shown to be promising feature
extraction algorithms. Such models are able to
compute the activity in the whole brain volume, only
using EEG and a head model that generally
represents the brain as a set of volume elements.
Peak picking (PP) method detects a specific pattern
based on its peak value in a region associated with a
specific cognitive component. It is used specifically
for the evoked potential P300 (or P3)-based BCI
system.

C. Feature classification in BCI Design

Feature classification of the acquired signal getting in
BCI, is accomplished using following approaches:
SVM (support vector machines) are supervised
learning models with associated learning algorithms
that analyze data and recognize patterns, used for
classification and regression analysis. A Support
Vector Machine (SVM) is a discriminative classifier
formally defined by a separating hyperplane, as
shown in Fig. 4. In other words, given labeled
training data (supervised learning), the algorithm
outputs an optimal hyperplane which categorizes
new examples.

Fig. 4 Support Vector Machine

Fig. 2 P300 evoked potential

Slow cortical potentials (SCPs) calculation methods

amplitudes are extracted on-line from the regular
electroencephalogram, filtered, corrected for eye
movement artifacts and fed back to the patient with
visual, auditory or tactile feedback.

LDA (Linear discriminant analysis) is used in

statistics, pattern recognition and machine learning
to find a linear combination of features which
characterizes or separates two or more classes of
objects or events. The resulting combination may be
used as a linear classifier, or, more commonly, for
dimensionality reduction before later classification.
Bayesian classifier is a simple probabilistic classifier
based on applying Bayes' theorem (from Bayesian
statistics) with strong (naive) independence
assumptions. A more descriptive term for the
underlying probability model would be "independent
feature model".
D.

Fig. 3 Design patterns in BCI

BCI TRANSLATION ALGORITHMS

Translation algorithm is a series of computations that

transforms the BCI input features derived by the signal
processing stage into actual device control commands.
Stated in a different way, a translation algorithm takes
abstract feature vectors that reflect specific aspects of
the current state of the users EEG or single-unit activity
(i.e., aspects that encode the message that the user wants
to communicate) and transforms those vectors into
application-dependent device commands. Different BCI
systems use different translation algorithms.

Each algorithm can be classified in terms of three key

features: transfer function, adaptive capacity, and
output. The transfer function can be linear (e.g., linear
discriminant analysis, linear equations) or nonlinear
(e.g., neural networks). The algorithm can be adaptive
or non-adaptive. Adaptive algorithms can use simple
handcrafted rules or more sophisticated machinelearning algorithms. The output of the algorithm may be
discrete (e.g., letter selection) or continuous (e.g., cursor
movement). The diversity in translation algorithms
among research groups is due in part to diversity in their
intended real-world applications. Nevertheless, in all
cases the goal is to maximize performance and
practicability for the chosen application.
7.

A.

TECHNICAL DESCRIPTION OF THE SCENARIO

Experimental Set-Up

The idea is to focus on controlling external devices and

thought depiction in written format and further
application through communication based on that. This
can be conceived as potentially resulting from a natural
sequence of motor intention, followed by preparation
and completing by the execution. Accordingly, the
neuro-physiological approach aims of current system
are to capture EEG indices of preparation for an
immediately upcoming motor action. As mentioned
above, the interpretation of signals is of two types
Control and Character level commands.
B.

rhythm. It is also known as the posterior dominant

rhythm. At the time of wakefulness, rhythm is present
over posterior regions of head, maximal with the subject
relaxed and closed-eyes. It diminishes with eye opening.
The frequency ranges from 8 to 13 Hz in adults. It is
typically sinusoidal which is shown in Fig 5.
Some of the normal individuals do not have rhythm
during wakefulness. By itself, this is not abnormal, e.g.
EEG in encephalopathy
Seizures and status epilepticus
Periodic discharges along with other controversial
EEG patterns
EEG in cerebrovascular disease.
Artifacts which mimic seizures and other
physiologic patterns.
Prolonged EEG monitoring , quantitative EEG
techniques for detecting seizures and ischemia
Event-related and evoked potentials in the ICU

Output (Control Level / Character Level)

Fig. 5 International 1020 system

Application running on a separate computer receives

combined results (commands) of the classification via
an asynchronous interface and acquires them in a
temporal queue. It examines the queue repeatedly for
stationary signals persisting for a certain time length.
On primary level signals are captured via EEG and in
the analysis of the EEG the following need to be
evaluated:

Frequency
Voltage
Location
Morphology
Polarity
State
Reactivity
Symmetry
Artifact.

An important feature of the EEG is the frequency of

Different patterns of signals captured via EEG with

respective activity is depicted in the Fig 6.
Activity

Alpha rhythm and

eye blinks

Muscle artifact

Signals Image via EEG

Lateral
eye movement

Alpha rhythm

Fig. 6 EEG signals

C.

Eye closure

The architecture overview

The block architecture of the scenario is shown in Fig.

6.

Eye blinks

Mu rhythm

Fig. 7 BCI Architecture

D.

Leftward rapid
eye movement

Stepwise depiction of architecture

rightward
rapid
eye movement
Fig. 8 Capture Signal

Beta activity
Fig. 9 Signal processing

E.
Lambda waves

Phase by Phase Architecture Nuances

BCI scenario System is divided into two phases

Training Module:
The guiding motto of BCI is: Let the machines learn
through thought interpretations!, thus the user should
require only a minimum of training for operating it. The
training procedure described here serves for teaching

the machine and adjusting its model parameters to

better match the user and his brain signals properties.
During the training procedure EEG has to be acquired
from the user while performing a certain task, e.g.,
execution or imagination of left- vs right-hand
movement of the index fingers. The user should be
instructed to sit comfortably and should omit any
muscular artifacts, like biting, gulping, yawning,
moving the head, arms, legs or the whole body as these
may induce electromyography (EMG) noise activity that
interferes with EEG signals, such that the SNR
converges to zero. Eye movements are to be minimized
for the same reason.

In training phase system will capture sensory signals

response and verify the same for training data and
generalized commands.

The fig. 10 below depicts the possible latency of signal

to processing:

Fig. 10 Latency of Signal Processing

Fig. 12 Training Phase of Handshake Protocol

The image shown above is illustrated assuming that the

tasks that are performed for a period of 6 min
repetitively with an interlude of 0.52 s. All the training
sessions may be performed in two experimental kinds:
(1) imagined, (2) executed. In the executed task
experiment response markers may be acquired via
keyboard, while the user determines himself which task
to perform next. For the imagined task experiment a
visual cue indicates the task, which has to be executed
on the next auditory beat produced by a digital
metronome. Both stimuli place corresponding markers
into the data, stored with a timestamp. A time series of
EEG activity acquired within a certain time region
before the marker in order to train the learning machine
and adjust its parameters.

In training phase sensory signals response shall be

generalized via classification.

Fig. 11 Training Architecture

Fig. 13 Signal Response Classification

As depicted in Fig 14, there are following phases:

Signal Acquisition: The electric signals generated by the
neurons are acquired and processed by the signal
acquisition and processing techniques and devices. NonInvasive acquisition are used to capture the signals or
electrophysiological signals from the scalp by using the
technologies like electroencephalogram (EEG),
functional magnetic resonance imaging (fMRI),
magneto-encephalogram (MEG), P-300 based BCI etc.
Among non-invasive Brain Computer Interfaces (BCIs),
electroencephalogram (EEG) has been the most
commonly used for them because EEG is advantageous
in terms of its simplicity and ease of use, which meets
BCI specifications when considering practical use.

Functional Phase

Fig. 14 Functional Architecture

Signal Pre-Processing: The unwanted signal like noise

is inevitable. EEG recordings contain electrical signals
from the brain along with several unwanted signals like
interference from electronic equipment. Those
unwanted components may bias the analysis of the EEG
leading to wrong conclusions. So the signal must be preprocessed to remove the noise.

Fig. 15 Functional Phase of Handshake Protocol

The step by step execution process of functional module

of Handshake Protocol is must to understand. In
Functional phase system will identify and cross verify
signals coming from user (object).

Signal Classification: Simultaneously capturing of the

signals in large amounts is not possible as the brain
signals are continuously captured through electrodes.
The waves are generally classified on frequency and on
their shape. There are six important signals Beta, Alpha,
Delta, Gama, Mu and Theta.
Computer Interaction: Once the signals are classified,
they will be used by an appropriate algorithm for the
development of a certain application. Some limitations
in implementing the BCI system are variability in the
acquired EEG signals. Different types of signals are
acquired from the same person in different sessions,
different signals are acquired when many people are
performing the same mental task and EEG signals are
affected by the persons eye blinks, muscular
movements, suddenly hearing sound and interference
from electronic devices etc. Based on these different
types of applications like, character recognition, phrase
recognition, object movement etc., it is observed that it
is also possible to reconstruct the images and videos on
to the screen from brain signals.

Fig. 16 Sequence Diagram of function phase

8.

BENEFITS OF BCI

The system is capable of giving a better way to live

instead of adulthood disabilities. It can be used to
control devices which are helpful for disabled
peoples like those on wheel chair, according to the
willingness of users.

 BCI is capable to serve human criticality by giving

assistance in trauma. If user is suffering from trauma
and is in under observation in ICU of hospital, then
under these circumstances BCI system can be
helpful for monitoring the critical situation like, if
the patient is in panic then his changed brain signals
can be read and by interpreting the signals it can
send an alert to the observer and prompt the observer
for necessary action.
BCI will bring a revolution by giving utility to
control devices independent of sensory muscles. If
suppose user is paralyzed and need to operate
external devices then BCI gives the facility to the
user to operate the devices using brain signals which
are interpreted into character level or control level
commands.
BCI provides an easy way to control surroundings
using smart phones. The character and control level
commands can easily be used to instruct the smart
gadgets for ease of life. Thus, gives a new way to
control devices using thoughts.
BCI can be used for natural language processing.
The output of the system can be given to NLP
system for translation and synthesis. This translated
and synthesized output can be used for
communication for disabled peoples.
9.

EVALUATION (CONSTRAINTS/DEPENDENCIES)

Developments of BCI systems face mainly following

issues;
Directly using BCI for differently abled people or
critical people who are unable to use language for
commands or communicate. The challenge remains
using it without training phase with generalized
assumed commands.
One of the major problems in designing and
developing BCI System is to gather EEG signals
with variation in pattern. Variability is observed in
the signal of the same person acquired during
different sessions. Different person performing same
mental task produced different pattern in their EEG
signals. Due to this variability it is difficult to design
and developed a generic BCI system which works
for multiple subjects.
Another problem is that the EEG signal of a person
contains artifacts like Eye movements/blinks,
Muscular movement, interference from electronic
devices etc. The presence of such artifacts lowers the
signal-to-noise ratio (SNR) of the EEG signal. This
signal to noise ratio problem is solved in by the use
of artifact Processor.

The information transfer rates for experienced

subjects and well-tuned BCI systems are relatively
low.
A communication system for a person with severe
disabilities should be completely controlled by its
user. Unfortunately BCI systems require extensive
assistance from caretakers who need to apply
electrodes or signal-receiving devices before a user
can communicate this problem can be solved using
hybrid systems, combinations of different BCI
techniques, and other biometric interfaces to address
this problem.
Current EEG recording systems use standard wet
electrodes. These electrodes consist of electrolytic
gel which is required to reduce electrode skin
interface impedance. A use of electrolytic gel is very
inconvenient and uncomfortable and, as a large
number of electrodes is used to read the brain waves.
An electrode cap with a large numbers of electrodes
is uncomfortable for users to wear and thus
unsuitable for long-term recording. EEG recording
takes a long time, thus making BCI a time
consuming task. Moreover, recording hardware with
a large amount of channels is quite expensive, so that
it is difficult for common users to afford.
The cost of BCI equipment is relatively modest,
current systems require substantial and continuing
expert oversight, which is extremely expensive and
currently limited to a few research laboratories. As a
result, these systems are not readily available to most
potential users.
The need to select appropriate brain signals, to
record them accurately and reliably, to analyze them
appropriately in real time, to control devices that
provide functions of practical value to people with
severe disabilities, to manage the complex shortterm and long-term adaptive interactions between
user and system, and to integrate BCI applications
into the lives of their users, means that the expertise
and efforts of all these disciplines are critical for
success.
10. CONCLUSION

At present, the main motivation to BCI research and

development is the expectation that BCI technology will
be valuable for those whose suffer from neuromuscular
disabilities that prevents them from using conventional
enhancive communication methods.
BCI have inputs, outputs and translation algorithms like
other communication and control system that convert
the former to the latter. BCI operations depends on the
interaction of adaptive controllers, the human brain,
which produces input (i.e. the electrophysiological

activity measured by the BCI system) and the system [7] Donchin E, Smith DB (1970) The contingent negative
variation and the late positive wave of the average evoked
itself, which translates that activity into outputs (i.e.
potential. Electroencephalogr Clin Neurophysiol 29:201203
specific commands that act on the external world).
BCI inputs include slow cortical potentials, P300
evoked potentials, and rhythms from sensory-motor
cortex. Recording methodologies seek to maximize
signal to noise ratio. Noise consists of EMG, ECoG and
other activity from source outside the brain, as well as
brain activity different from specific rhythms or evoked
potentials that comprise the BCI input. A variety of
temporal and spatial filters can reduce such noise and
thereby increase the signal-to-noise ratio. BCI
translation algorithms include linear equations, neural
networks, and numerous other classification techniques.
The most difficult aspect of their design and
implementation is the need for continuing adaptation to
the characteristics of the input provided by the user.
BCI development depends on the close interdisciplinary
cooperation between neurological scientists, engineers,
psychologists, computer scientists, and therapy
specialists. Appropriate user populations must be
identified, and BCI applications must be configured to
meet their most important needs.
We have given a scenario based approach and presented
a concept for the research and development of a typical
BCI system which could drive applications in a new
way. We believe that BCI system is an important
development precisely because anyone could simply
think and potentially reconstruct the thoughts.
11. REFERENCES

[1] Berger H (1929) The Berlin Brain-Computer Interface (BBCI)

Archiv Psychiatrischer Nervenkrankheiten 87:527580
[2] Birbaumer N (1997) Slow cortical potentials: their origin
meaning and clinical use. In: Brain and behavior: past, present
and future. Tilburg University Press, Tilburg, pp 2539
[3] Birbaumer N, Ghanayim N, Hinterberger T, Iversen I,
Kotchoubey B, Kbler A, Perelmouter J, Taub E, Flor H
(1999) Spelling device for the paralyzed. Nature 398:297298
[4] Blankertz B, Curio G, Mller K-R (2002) Classifying single
trial EEG: towards brain computer interfacing. Adv Neural Inf
Process Syst 14:157164
[5] Blankertz B, Dornhege G, Schfer C, Krepki R, Kohlmorgen
J, Mller K-R, Kunzmann V, Losch F,Curio G (2003)
Boosting bit rates and error detection for the classification of
fast-pace motor commands based on single-trial EEG analysis.
IEEE Trans Neural Syst Rehabil Eng 11(2):127131
[6] Cui RQ, Huter D, Lang W Deecke L (1999) Neuroimage of
voluntary movement: topography of the Bereitschaftspotential,
a 64-channel DC current source density study. NeuroImage
9:124134

[8] Ebrahimi T, Vesin J-M, Garcia G (2003) Braincomputer

interface. A new frontier in multimedia communication. IEEE
Signal Processing SocietySignal Process Mag 20(1):1424
[9] Green JB, Sora E, Bialy Y, Ricamato A, Thatcher RW (1999)
Cortical motor reorganization after paraplegia: an EEG study.
Neurology 53(4):736743
[10] Hardwick A, Rush J, Furner S, Seton J (1996) Feeling it as
well as seeing it. Haptic displays within gestural HCI for
multimedia. In: Proceedings of the York Gesture Workshop,
University of York, York, Great Britain, March 96, pp 105
116
[11] Harel D, Carmel L, Lancet D (2003) Towards an odor
communication system. Computational Biology and
Chemistry 27:121133
[12] Lang W, Zilch O, Koska C, Lindinger G Deecke L (1989)
Negative cortical DC shifts preceding and accompanying
simple and complex sequential movements. Experiments in
Brain Research 74:99104
[13] MacIntyre B, Feiner S (1996) Future multimedia user
interfaces. Multimed Syst J 4(5):250268
[14] MacKenzie IS (1991) Fitts law as a performance model in
humancomputer interaction. Doctoral dissertation, University
of Toronto, Canada
[15] Mika S, Rtsch G, Mller K-R (2001) A mathematical
programming approach to the kernel fisher algorithm. Adv
Neural Inf Process Syst 13:591597
[16] Mller K-R, Mika S, Rtsch G, Tsuda K, Schlkopf B (2001)
An introduction to kernel-based learning algorithms. IEEE
Trans Neural Netw 12(2):181201
[17] Nicolelis MAL, Chapin JK (2002) Controlling robots with the
mind. Sci Am 287(4):4653 Multimed Tools Appl
[18] Pantic M, Rothkrantz LJM (2000) Automatic analysis of facial
expressions. IEEE Trans Pattern Anal Mach Intell
22(12):14241445
[19] Pentland A (1995) Machine understanding of human action.
In: Proceedings of the 7th international forum frontier of
telecommunication technology, Tokyo, Japan, November
1995, pp.757764
[20] Pfurtscheller G (1999) EEG event-related desynchronization
(ERD) and event-related synchronization (ERS). In:
Electroencephalography: basic principals, clinical applications
and related fields, 4th edn. Williams and Wilkins, Baltimore,
MD, pp 958967
[21] Pfurtscheller G, Flotzinger D, Kalcher J (1993) Brain
computer interface. A new communication device for
handicapped persons. J Microcomput Appl 16:293299
[22] Sharbrough F, Chatrian G-E, Lesser RP, Lders H, Nuwer M,
Picton TW (1991) American Electroencephalographic Society
guidelines for standard electrode position nomenclature. J Clin
Neurophysiol 8:200202
[23] Sutter EE (1992) The brain response interface: communication
through visually induced electrical brain responses. J
Microcomput Appl 15:3145
[24] Vapnik V (1995) The nature of statistical learning theory.
Springer, Berlin Heidelberg New York 25. Wolpaw JR,
[25] McFarland DJ, Neat GW, Forneris GW (1991) An EEG-based
braincomputer
interface
for
cursor
control.
Electroencephalogr Clin Neurophysiol 78:252259

[26] Wolpaw JR, Birbaumer N, McFarland DJ, Pfurtscheller G,

Vaughan TM (2002) Braincomputer interfaces for
communication and control. Clin Neurophysiol 113:767791
[27] Krepki R (2004) Brain-Computer Interfaces. Design and
implementation of an online BCI-System for the control in
gaming applications and virtual limbs. Doctoral dissertation,
Technical University of Berlin, Germany
[28] Gupta, S. & Singh, H. (1996) Preprocessing EEG signals for
direct human-system interface, Intelligence and Systems.
IEEE International Joint Symposia on, pp. 32-37
[29] Brain-Computer Interfaces: Beyond Medical Applications
Computer -IEEE Computer Society- 45, 4 (2012) 26-34" DOI
: 10.1109/MC.2012.107
[30] T. Kameswara Rao et al An Exploration on Brain Computer
Interface and Its Recent Trends www.ijarai.thesai.org Vol. 1,
No. 8, 2012
[31] Signal Processing and Classification Approaches for Braincomputer
Interface.
http://cdn.intechopen.com/pdfswm/6798.pdf
[32] Robert Horlings Man-Machine Interaction Group TU Delft
27 March 2008 (IJARAI) International Journal of Advanced
Research in Artificial Intelligence