Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

Brain Computer Interface: Implementation and

Applications
1
Seemant Singh, 2Sharayu Barwe, 3Niraj Bhosale ,4Tanmay Shendkar, 5Shilpa Bhosale,
1,2,3,4,5
Student, Department of Computer engineering, NBNSSOE, Maharashtra, India

ABSTRACT
A brain computer interface, which is likewise known as an neural-manage interface, direct neural interface is an
instantaneous conversation pathway between an improved or stressed brain and an external tool. Mind studying and
remote communication have their precise fingerprint in numerous fields along with instructional, self-law,
production, advertising, security as well as games and enjoyment. It creates a mutual information between
customers and the encompassing structures. This paper indicates the implementation elements of the tool as well as
the applications. The main device i.e. EEG is also explained well in the paper to understand the BCI from very basic
level. The review paper also shows the algorithms that can be used in this BCI and machine learning platform
development. Different solutions that aim to limit and decrease their effects have also been represented.

Keyword BCI, EEG, non-invasive, classifications, electrodes, gaming and entertainment, mind control.

1. Introduction
A brain computer interface (BCI) which is also referred as a brain device interface (BMI), is a hardware and
software program communications device that facilitates people to engage with their environment, without the
involvement of peripheral nerves and muscular tissues, by using the usage of control signals generated from
electroencephalographic device. A BCI is an artificial intelligence device that can apprehend a positive set of styles
in brain signals following 5 consecutive degrees: signal acquisition, pre-processing or signal enhancement, feature
extraction, classification, and the manage interface the signal acquisition level captures the mind alerts and may also
perform noise reduction and artifact processing. The pre-processing stage prepares the indicators in a suitable form
for further processing. The function extraction stage identifies discriminative information inside the mind indicators
which have been recorded. The extraction of this thrilling records is a very tough challenge. Brain signals are mixed
with different signals coming from a finite set of mind sports that overlap in both time and area. The function vector
must also be of a low measurement, to reduce feature extraction degree complexity, without a relevant data loss.
Sooner or later the manipulate interface stage interprets the labeled alerts into meaningful commands for any related
tool, inclusive of computer.

8415 www.ijariie.com 318

Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

Fig-1BCI component and stages

2. Literature Survey

Sr. Year Title Author Proposed system Drawbacks

no

1 2015 Mind guided motion Antoni MalkP, -How EEG can be -The p300 speller
control of robot Chenguang Yang, Ning acquired through could be improved
manipulator using Wang!, and Zhijun Li2 headset. and adapted to get
EEG signals better and faster
-OpenViBE platform results.
to Process EEG
-More robot control
possibilities are not
better, and the
functionality can be
further enriched.

2 2015 Brain computer Rabie A. Ramadan, -Detailed phase -Comparisons and

interface: control Athanasios V. Vasilakos description classification of
signals reviews signals is not
-Up-to-date BCI effective.
algorithms and
techniques for all -problem of noise
phases reduction is not
solved well

8415 www.ijariie.com 319

Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

3 2015 2015: Brain- Narisa N.Y. Chu -brain signal -algorithms are not
Computer processing algorithmic more accurate
Interface approach in achieving
Technology and the level of -implantations of
Development: The interpretability of headsets is not
emergence of brain signals comfortable
imprecise
brainwave -standardize the brain
headsets in the signal databank,
commercial world. anticipating its reach
to big data

4 2012 Brain computer Luis Fernando Nicolas- -provide -invasive methods

interface: Review Alonso* and Jaime communication ways need more
Gomez-Gil to disabled people clarification

-Determination of -information bit rate

users intention is low

5 2015 Brain computing Sarah N.Abdulkader, -use of BCI in games -training challenges
interfacing: AymanAtia,Mostafa- and entertainment
applications and Sami M.Mostafa -non-linearity
challenges -creation of mutual
understanding -information
between user and transfer rate
surrounding system
-noise

6 2013 EEG-Based Brain- Han-Jeong Hwang a , -BCI paradigms -brain complexity is

Computer Soyoun Kim b , Soobeom not solved well
Interfaces: A Choi a & Chang-Hwan -target applications
Thorough Literature Im a -algorithms and
Survey -feature types feature
classifications
-classification should be more
algorithm precise

3. Proposed System
3.1 EEG Device
EEG is a non-invasive manner to inspect mind. EEG measures electric mind interest resulting from the float of
electric currents at some stage in synaptic excitations of the dendrites in the neurons and is extraordinarily sensitive
to the results of secondary currents. EEG alerts can be effortlessly recorded through electrodes located on the scalp.
The EEG recording machine consists of electrodes, amplifiers, A/D converter, and a recording tool.

The EEG recording machine consists of electrodes, amplifiers, A/D converter, and a recording tool. The electrodes
gather the signal from the scalp, the amplifiers system the analog signal to extend the amplitude of the EEG signals
in order that the A/D converter can digitalize the signal in a extra correct manner. In the end, the recording tool,
which can be a personal laptop or comparable, stores, and displays the facts. Electrodes are commonly fabricated
from silver chloride. Electrode-scalp touch impedance have to be among 1Ω and 10Ω to record an accurate signal.

8415 www.ijariie.com 320

Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

The electrode-tissue interface isn't simplest resistive however also capacitive and it consequently behaves as a low
skip filter out. EEG gel creates a conductive course among the skin and every electrode that reduces the impedance.
Use of the gel is cumbersome, however, as persisted maintenance is needed to guarantee a tremendously top pleasant
sign. The waves we will be most without difficulty capable to distinguish are alpha and beta waves -- alpha waves
arise at round 8-12 Hz and when measured from the frontal lobe provide an estimate of how at ease a person is, at
the same time as beta waves are around 12-30 Hz and correspond to how someone is concentrating or how alert they
may be.

3.1 Software
When the hardware is completely prepared, person can put the electrodes at the scalp as a way to get the mind wave
indicators as an enter. These signals could be then processed within the device built. For further processing, i.e. to
expose these indicators at the laptop display we require the software program component. This system evolved at the
start degree may be sincerely only a records acquisition/ visualization one.

Even if desktop is capable of get the mind waves as an input, these alerts will now not be repetitive. For this reason,
a system studying platform is developed with a purpose to be easy to classify the ones alerts the use of a quality
perfect algorithm. So the type step could be very vital to recognize the user’s intentions. Regression or type
algorithm can help to gain this purpose however presently, classification is the maximum famous method.

Type algorithms have traditionally been calibrated via customers thru supervised learning the use of a labeled data
set. It is assumed that the classifier is able to locate the patterns of the brain signal recorded in on line classes with
feedback.

3.2 Classification algorithm

The motive of the classification step in a BCI system is recognition of a user’s intentions on the basis of a feature
vector that characterizes the brain activity provided by the feature step. Classification algorithms can be developed
via either offline, online or both kinds of sessions. The offline session involves the examination of data sets, such as
BCI competitions data sets which are collected from an adaptive or closed-loop system.

The feature vector is low (two) dimensional, that’s why good classification results can be obtained with simple
algorithms. The KNN algorithm is one of the simplest algorithms to classify objects in machine learning. All feature
vectors which are extracted from the sub-training set are located in the feature space. A feature vector which belongs
to the test data is then classified with the class of the majority of the k-nearest neighbors among all the located
feature vectors. The performance of a nearest neighbor classifier depends on the distance metric and the value of the
neighborhood parameter k, which controls the volume of the neighborhood. The most common choice for the
distance function is the Euclidean metric. The KNN algorithm is really very sensitive to the local distribution of
feature vectors. EEG signals are naturally non-stationary, subject-specific, distorted by various artifacts like EMG
and EOG, that’s why this kind of information is usually not available. Consequently, the neighborhood parameter
must be estimated from the training data at hand by using a heuristic technique. The cross-validation technique
described above can be used to select the optimal neighborhood parameter that yields the highest accuracy rate for
the test data. The k-nearest neighbors (KNN) algorithm finds the k nearest neighbors among the training set, and the
categories of the k-nearest neighbors are used to weigh the category candidates. The performance of this algorithm
greatly depends on two factors: a suitable similarity function and an appropriate value for k. If k is too large, big
classes will overwhelm the small classes whereas if k is too small, the advantage of KNN algorithm cannot be
traced.

4. Applications
The main target populations for BCI packages fall into 3 categories. The primary group consists of complete
Locked-In nation (CLIS) sufferers who have misplaced all motor manage, due to the fact they'll be at a terminal
degree of ALS or suffer severe cerebral palsy. The second one organization incorporates Locked-In country (LIS)
patients who are nearly absolutely paralyzed, however with residual voluntary movement, along with eye
movement, eye blinks, or twitches with the lip. The third organization of ability BCI customers includes abled
bodied humans and people with giant neuromuscular manage, especially speech and/or hand manage.

8415 www.ijariie.com 321

Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

Fig 2-BCI applications

Nowadays, there are a considerable variety of very different BCI packages, which includes word processors, tailored
net browsers, mind manage of a wheelchair or neuroprostheses, and games, amongst others. But, maximum
packages have completely been designed for education or demonstration purposes.

Leisure-orientated BCI programs have commonly had a lower precedence in this discipline. Till now, research into
BCI generation has generally focused on assistive packages, inclusive of spelling gadgets, wheelchair control or
neuroprostheses rather than applications with enjoyment purposes. Leisure-orientated BCI programs have adapted
very well-known video games inclusive of Pac-man, Pong and comparable games so that they'll be played thru
motor imagery.

Here the game of ping pong is advanced to expose the primary running of BCI where you manipulate the paddles
together with your concentration of alpha waves. The very best way to manipulate the controls is to shut your eyes
and loosen up to make the paddle circulate up, and to open your eyes and attention to make it move down.

5. Future scope
The facts switch charge provided by BCIs is simply too low for herbal interactive communication, even for
experienced topics and nicely-tuned BCI systems.

the excessive mistakes charge in addition complicates the interaction.

BCI systems cannot be used autonomously by u (D.S. Tan, 2010)sing disabled people, because BCI systems require
assistants to use electrodes or signal-receiving gadgets earlier than the disabled individual can talk.

A BCI user can be capable to turn the BCI gadget off by means of brain pastime as input, but generally can't flip it
again on once more, that is termed the “Midas touch” hassle dealing with BCI applications demands a high cognitive
load which could normally be carried out by way of users in quiet laboratory environment, however not inside the
actual global.

6. Conclusion
This above paper indicates the details of mind computing interface and how it may be very useful if programs for
real life may be evolved. It additionally suggests the hardware and software development of a small device and
related software so as to reveal how BCI can be an open source platform the use of machine learning. Various
programs and challenges are also given inside the paper. Those challenges can be overcome with time the use of
extra advanced software as well as hardware answers. The achievements and the ability for brand new BCI packages
have manifestly given a great raise to BCI research concerning multidisciplinary scientists e.g., neuroscientists,

8415 www.ijariie.com 322

Vol-4 Issue-3 2018 IJARIIE-ISSN(O)-2395-4396

engineers, mathematicians, and medical rehabilitation professionals, amongst others. Interest in the BCI discipline is
predicted to increase and BCI design and improvement will in all opportunity hold to carry blessings to the daily
lives of disabled people.

7. References
1. Antoni MalkP, C. Y. (2015). Mind guided Motion Control of Robot Manipulator using EEG signals.
International Conference on Information Science and Technology (IClST) . Hunan, China.

2. D.S. Tan, A. N. (2010). Brain-computer interfaces: applying our minds to human-computer interaction.
springer .

3. Gomez-Gil, L. F.-A. (2012). Brain Computer Interfaces, a Review.

4. J.S. Brumberg, A. N.-C. (2010). Brain-computer interfaces for speech communication. Speech Commun, 52
(4) (2010),, (pp. 367-379).

5. Luis Fernando Nicolas-Alonso and Jaime Gomez-Gil. (2012). Brain computer interfaces, a review. NCBI ,
13.

6. Mellinger J., S. G. (2007). An MEG-based brain-computer interface (BCI) NeuroImage., (pp. 581-593).

7. Sarah N.Abdulkader, A.-S. M. (July-2015). Brain computer interfacing: Applications and challenges.
Egyptian informatics journal , 213-230.

8. Wolpaw J.R., B. N. (2002). Brain-computer interfaces for communication and control. Clin.Neurophysiol ,
767-791

9. Blankertz B., Tangermann M., Vidaurre C., Fazli S., Sannelli C., Haufe S., Maeder C., Ramsey L., Sturm
I., Curio G., Mueller K.R. The Berlin Brain-Computer Interface: Non-Medical Uses of BCI
Technology. Front. Neurosci. 2010 doi: 10.3389/fnins.2010.00198.

10. Teplan M. Fundamentals of EEG measurement. Meas. Sci. Rev. 2002;2:1–11

11. Lebedev M.A., Nicolelis M.A.L. Brain-machine interfaces: Past, present and future. Trends
Neurosci. 2006;29:536–546.

8415 www.ijariie.com 323