Brain-Computer-Interface

1st Pilla Chaitanya 2nd Majjari Lakshmi Gowtham

School of Computer Science Engineering and Technology School of Computer Science Engineering and Technology
Bennett University,NOIDA Bennett University,NOIDA
Noida,India NOIDA,INDIA
[email protected] [email protected]

3rd Guvvala Bhargav Raja Reddy

School of Computer Science Engineering and Technology
Bennett University,NOIDA
NOIDA,India
[email protected]

Abstract—Brain impulses are received, processed, and trans- b) This work presents a sophisticated algorithmic frame-
formed by brain-computer interfaces (BCIs) into orders that are work that enhances the accuracy and adaptability of signal
transmitted to output devices to carry out predetermined tasks. interpretation specific to motor imagery tasks, placing it in a
BCIs do not employ regular neuromuscular output channels.
The main goal of brain-computer interface (BCI) is to replace good position to solve these basic problems. By concentrating
or restore functional capacity to those who have been rendered on three key areas for development, the initiative aims to
incapable due to neuromuscular disorders including stroke, amy- dramatically improve the functional efficacy and user-centric
otrophic lateral sclerosis, cerebral palsy, or spinal cord damage. performance of BCIs. This will help those with severe motor
Since the first demonstrations of electroencephalography-based disabilities communicate better and have better lives overall.
spelling and single-neuron-based device control, researchers
have used electroencephalographic, intracortical, electrocortico- :
graphic, and other brain signals for increasingly complex control
of cursors, robotic arms, prostheses, wheelchairs, and other
devices. Non-invasive BCIs monitor brain activity using external
sensors, negating the need for surgical operations. Electroen-
cephalography (EEG) is one of the most extensively utilized non-
invasive BCI technologies, measuring electrical activity through-
out the scalp caused by neuron firing in the cerebral cortex. II. I NTRODUCTION
Invasive BCIs entail inserting electrodes directly into brain tissue
via surgery. While they provide high-resolution information, they
also pose surgical risks. Non-invasive BCIs use external sensors,
such as EEG (electroencephalography), to monitor brain activity
without requiring surgery. They are safer and more accessible
than invasive BCIs, but the spatial resolution may be reduced. Spinal cord injury (SCI), brainstem stroke, and amyotrophic
Non-invasive BCIs are used in assistive technology, communica- lateral sclerosis (ALS) are examples of severe neuromuscular
tion assistance, and numerous research sectors that require direct illnesses or motor system impairments that typically result in
brain-to-computer communication without intrusive treatments. the loss of upper limb function, including the ability to reach
I. P ROBLEM S TATEMENT and grasp things and carry out other daily tasks. Still, most
affected individuals may be able to use their brains to operate
a) Brain-Computer Interfaces (BCIs) based on motor
external devices through function-related neural processes.
imagery have the potential to completely change the fields
Robots, which are generally developed as neuroprosthetic
of rehabilitation therapy and assist individuals with motor
devices, can enhance their ability to sense their environment.
function deficiencies in communicating. Despite these tech-
However, coordinating the many degrees of freedom (DOFs) to
nologies’ transformative promise, persistent technical chal-
achieve anthropomorphic control continues to be a significant
lenges severely restrict their mainstream clinical adoption.
challenge for these severely afflicted patients. Expanded use
Among these are the following limitations: real-time data
of BCI systems has been seen in control applications, such
processing slows; brain signals matching to imagined motor
as motorized wheelchairs that take the place of mobility in
motions cannot be accurately recognized; and systems cannot
cases when limb control is lost or neuroprosthetic devices that
dynamically adjust to each user’s unique neural profile. These
help people with spinal cord injuries regain hand function. The
defects not only reduce the functional reliability of BCIs but
development of new CGI interfaces and their combination with
also limit their practical usability and human acceptance. :
BCI technology have made it easier to analyze how subjects
Identify applicable funding agency here. If none, delete this. behave in virtual environments.
 III. C OMPONENTS OF A BCI S YSTEM

A. Signal Acquisition
Signal acquisition is the process of measuring brain sig-
nals using a particular sensor modality, such as scalp or
intracranial electrodes for electrophysiologic activity or fMRI
for metabolic activity. The signals are amplified to levels
suitable for electronic processing after being filtered to remove
electrical noise and other undesirable signal characteristics,
such as 60-Hz power line interference. After the impulses are
digitalized, they are routed to a computer.
B. Feature Extraction
Feature extraction is the process of distilling meaningful
signal characteristics—that is, signal characteristics associated
with the individual’s intent—from extraneous content in digital
signals and encapsulating them in a compressed format that
may be translated into output commands. The user’s objective
ought to be closely linked to these characteristics. Since most
of the relevant (i.e., most strongly correlated) brain activity
is either transient or oscillatory, time-triggered EEG or ECoG
response amplitudes and latencies, power within specific EEG
a) A brain-computer interface (BCI) aims to detect and or ECoG frequency bands, or firing rates of individual cortical
quantify features of brain activity that correspond with the neurons are the most commonly extracted signal features
user’s objectives, and then translate those aspects into com- in current BCI systems. The reduction or elimination of
mands for the device that are executed in real time to fulfill physiologic artifacts, such as electromyographic signals and
the user’s wish. To do this, a BCI system consists of the environmental artifacts, ensures a trustworthy evaluation of the
following four components: [1] device output, [2] feature brain signal properties.
extraction, [3] feature translation, and [4] signal capture. C. Feature Translation
These four elements are governed by an operational protocol
that details the start and stop of operation, the details of The user’s intention is then fulfilled by the feature transla-
signal processing, the kinds of commands the device gives, tion algorithm, which takes the signal features and uses them
and performance monitoring. An effective operating protocol to convert them into commands that the output device can
allows a BCI system to be versatile and tailored to each user’s comprehend and execute. For example, a power drop within
specific needs. : a specific frequency band could be understood as moving the
computer cursor upward, or a P300 potential could be regarded
as selecting the letter that triggered it. It is imperative that the
b) components of a BCI system. It is possible to record translation algorithm be dynamic in order to accommodate
electrical impulses originating from brain activity using and respond to learned or spontaneous changes in the signal
electrodes positioned inside the brain, on the cortical surface, properties, so ensuring that the user’s available range of feature
or on the scalp. Brain impulses are converted to digital values spans the full range of device control.
format and enhanced in strength. Vital signal characteristics
are extracted and transformed into commands that drive an D. Device Output
output device, such as a motorized wheelchair, prosthetic limb, The feature translation method generates commands that
or spelling software. For best device functioning, the user are used to operate the external device. You can utilize these
can modify brain signals based on feedback from the device. : commands for a variety of tasks, such as selecting letters,
controlling a robotic arm, and more. When the user receives a hand orthosis, functional electrical stimulation (FES) of a
feedback regarding how the device is operating, the control patient’s hand, robotic and prosthetic devices, 31, 32.33, 34
loop closes. Two-dimensional cursor control has also been achieved using
attention modulation. :
IV. L ITERATURE R EVIEW: ]The visual P300 speller’s comparatively simple installation
a) BCI technology explores the development and history and performance make it a highly explored BCI paradigm.
of BCIs, covering invasive and non-invasive techniques such It has been effectively proven for a range of applications in
as hybrid approaches, fNIRS, and EEG. The paper assesses both healthy and impaired users, including typing, surfing the
classification methods including support vector machines and Internet, steering a wheelchair along preset routes, and more.
deep learning models used for brain signal interpretation, in Like the P300 evoked response, steady-state visual evoked
addition to signal processing techniques like filtering, artifact potentials are inborn and do not require training, but they can
removal, and feature extraction. It highlights effective case elicit faster response times. P300-based BCIs, however, need
studies and clinical trials while covering a variety of BCI significantly less on eye movement control than do steady-
applications, such as neurofeedback, gaming, assistive tech- state visual evoked potentials-based BCIs. Steady-state visual
nology, and neurorehabilitation. Future prospects including evoked potentials have been used for environmental control,
hybrid BCI systems and brain-inspired computing are con- spelling, prosthesis control, binary selection, FES, and both
sidered, along with challenges like signal variability and user discrete and continuous two-dimensional cursor control.
adaption. :
b) All things considered, the literature assessment offers B. BCIs That Use ECoG Activity
a basis for comprehending the present situation, pointing out
areas in need of more study, and suggesting creative paths for
developing BCI technology and its uses. The Brain-Computer
Interface (BCI) research literature review is an extensive
analysis of the body of knowledge now available and the
developments in BCI technology. The historical development
and evolution of BCIs is first described, covering the transition
from invasive methods like as intracranial recordings to non-
invasive techniques like EEG and fNIRS and hybrid systems
that combine multiple modalities for optimal performance. :
c) The study provides a comprehensive overview of signal
processing methods necessary for BCI operation, including
feature extraction, artifact handling, and noise reduction.
These methods are essential for deriving meaningful infor- 1) BCIs That Use ECoG Activity:
mation from complex brain signals. It also discusses the a) Since the cortical surface is where ECoG activity
classification techniques—like machine learning models—that is acquired, an electrode array—either subdural or epidu-
are applied to decipher these signals and convert them into ral—must be inserted. When recording signals with greater
commands that can be executed or control signals that can be amplitudes, ECoG offers superior spatial resolution and spec-
sent to other devices. In addition, a comprehensive review of trum breadth than EEG. Together with the lower-frequency
various BCI uses in domains like gaming, neurofeedback, as- (¡40 Hz) activity that dominates the EEG, ECoG also includes
sistive technology, neurorehabilitation, and cognitive improve- higher-frequency (i.e., ¿40 Hz gamma band) activity up to
ment is part of the analysis. It includes positive case studies 200 Hz and possibly higher. Gamma activity is significant
and clinical trials that demonstrate the effectiveness of BCIs because of its exceptionally accurate functional localization,
and their prospective benefits for enhancing communication, strong correlation with specific domains of motor, linguistic,
restoring motor function, and improving quality of life for and cognitive function, and association with signals that are
those with neurological diseases or disabilities. : reliant on blood oxygen levels and individual neuron firing
rates that have been found by fMRI. :
V. C URRENT E LECTROPHYSIOLOGIC BCI S YSTEMS
b) Individual hand 84–85, arm, and finger movements
A. BCIs That Use Scalp-Recorded EEG 80–83 have all been correctly decoded from ECoG. ECoG-
a) Noninvasive EEG-based BCIs have been the subject based BCIs use motor or sensory imagery, or working memory
of the most research due to their minimal risk and relative (dorsolateral prefrontal cortex) to control 1- or 2-dimensional
ease of study design and participant recruitment. Thus far, cursor movements.87–91 An ECoG-based BCI can use motor
the primary applications have been in discrete selection and imagery or the P300 event-related potential to enable users to
continuous movement control with a decreased degree of choose characters or use a prosthetic hand.12,92–94 Online
freedom. Sensorimotor rhythms have been used to control cursor control was the most recent application of ECoG
cursors in 1, 2, 10, 23, 24, and 325, 26 dimensions, a signals captured over speech cortex during perceived or ac-
wheelchair, a spelling assistance, traditional assistive devices, tual phoneme and word articulation95. Additionally, accurate
offline decoding was attained for potential application in voice and microcontroller are two examples of electrical circuits.
synthesis control. :
]It has been shown that motor imagery-based BCI control
that uses locations over motor cortex can produce ECoG
changes larger than those resulting from actual movements,
that ECoG-based BCI performance with fixed parameters is
stable for at least five days, and that epidural ECoG can be
used to achieve BCI control.100 It was found that across
several months, both the ECoG data and the model’s capacity
to interpret movement remained stable in a monkey investiga-
tion.101–102 These results suggest that ECoG is likely to be
effective for long-term BCI applications.

VI. H ARDWARE
B. Selection of EEG device
a) In order to actually implement a brain-computer in-
]Every BCI system needs the sensors and related equipment
terface (BCI), a device that records brain activity information
to gather brain signals. The development of hardware is critical
is needed. The electroencephalogram is the suggested method
to brain-computer interactions in the future. The ideal EEG-
for acquiring the brain activity signal (EEG). Consequently,
based (noninvasive) BCIs should have electrodes that don’t
selecting the best EEG gear for this endeavor is essential. The
require skin abrasion or conductive gel (also referred to as ”dry
commercialized market is home to a variety of EEG gadgets.
electrodes”); they should also be small and entirely portable;
They go by the names of Emotiv EPOC, Neurosky Mindset,
they should have easy-to-install, comfortable, and aesthetically
Mattel Mindflex, and OpenEEG. :
pleasing mountings; they should be easy to set up; they should
b) There are a few requirements that the EEG device
function well in all environments; they should run on telemetry
must meet in order to execute the BCI for the electrical
rather than wires; and they should be easily interfaced with a
wheelchair. The gadget needs to be convenient and portable
variety of applications. In theory, existing technology may be
first. Because the electronic boards required a lot of room,
able to meet several of these requirements, and solutions for
the OpenEEG is not appropriate in this situation[16]. The
dry electrodes are beginning to become available.
other devices, except from the OpenEEG gadget, are wear-
]Brain-computer interfaces with implanted electrodes en- able and portable. The software development tool is another
counter a number of difficult issues. Hardware for these prerequisite for this BCI. Since EEG is a complex electrical
systems needs to be fully implantable, safe, dependable, and waveform, creating software that can analyze and handle these
intact for decades; record signals steadily over time; transmit waveforms from the ground up will take a lot of effort. The
the recorded signals via telemetry; have long-lasting batter- developer’s workload can be lessened by using the offered
ies or be refilled in-situ; have external components that are developer tool. This is so that the developer tool, which
comfortable, convenient, and unobtrusive; and easily interface includes the Fast Fourier Transform algorithm, can process
with high-performance applications. Though there have been the raw EEG waveform. :
some amazing recent developments and the rare case of c) The developer can get the required data depending on
microelectrode implants continuing to function over time, it the BCI purpose after completing the basic signal processing.
is uncertain which treatments will be the most successful. Only Neurosky Mindset and Emotiv EPOC offer developertools
Instead of using neural action potentials, BCIs that use local in this situation. The Mattel Mindflex headset was therefore
field potentials or electroencephalograms (ECoG) may provide not selected for this project. Upon comparing the Neurosky
more consistently steady performance over time. But before Mindset and Emotiv EPOC, the primary distinction between
invasive BCIs realize their full potential, it’s possible that the two is the quantity and placement of the sensors. The
significant, as-yet-undefined advances in sensor technology are Mindset features a single sensor on the forehead, but the
required. The majority of the necessary research will still rely Emotiv EPOC has fourteen sensors spread across the head.
on studies done on animals before human trials begin. Emotiv EPOC can measure the signal strength in each distinct
VII. EEG C ONTROLLED W HEELCHAIR area of the brain thanks to these sensors. The Mindset is
limited to monitoring activity in a single region of the brain
A. Methodology because it only uses one sensor to record the EEG data. In
light of this, Emotiv EPOC can offer data that is more accurate
For this project, an electrical wheelchair is necessary. than that of Neurosky Mindset. Moreover, Emotiv EPOC has
The electronic circuitry for controlling the wheelchair’s the capacity to track the entire pattern of brain activity. This
electric motor are created, and the electrical wheelchair characteristic is highly helpful for this study since it allows us
frame is constructed entirely from scratch.The motor driver to distinguish between the patterns of brain activity associated
with the user’s mental commands, such as thinking. The best by directly controlling the motion commands that correspond
EEG equipment for this study is therefore Emotiv EPOC. : to the user’s brain signals while imagining movements of their
left or right limbs. The robotic platform is illustrated.Choi
C. Key Techniques of BRAIN-CONTROLLED Mobile Robots and colleagues [23], [28] also used aBCI based on motor
The main idea behind a brain-controlled mobile robot is to imagery to create a brain-controlled mobile robot. This robot
use the user’s brain impulses to operate the robot and get can execute three motion instructions, including turning left,
it where it’s going securely and effectively.Brain-computer right, and forward, and they validated this robot in a real-
interface (BCI) is the fundamental technology used to con- world setting. :
struct brain-controlled robots. It converts EEG data into user b) There is no need for further robot intelligence with
intentions and is essential to the operation of any mobile brain- these kinds of machines. Their computing complexity and
controlled robot. Apart from brain-computer interface (BCI), cost are hence minimal.Furthermore, users have the ability to
additional methods encompass 1) robotintelligence approaches control their movements as much as they like using personal
for detecting environmental conditions, localization, route motion control. Nonetheless, the overall efficacy of these brain-
planning, and avoiding obstacles and collisions, and 2) shared controlled mobile robots primarily hinges on the noninvasive
control methods that integrate BCI and robot intelligence to BCIs’ performance, which is now erratic and slow. Stated
distribute control over the robot [33], [34].We do not review differently, the robots’ performance is limited by the BCI
robot intelligence in this study because it is a common problem systems. Moreover, users frequently become fatigued from
that affects all autonomous mobile robots, in order to keep the issuing commands for motor control. :
scope of the paper as broad as possible manageable. ]The second category of brain-controlled robots has been
1) Brain-Controlled Activites: The main idea behind a developed from the perspective of shared control, where a
brain-controlled mobile robot is to use the user’s brain im- user (using a BCI) and an intelligent controller (like an
pulses to operate the robot and get it where it’s going securely autonomous navigation system)share control over the robots.
and effectively.Brain-computer interface (BCI) is the funda- This approach aims to address the two questions that the
mental technology used to construct brain-controlled robots. robots that are directly controlled by a BCI meet and to
It converts EEG data into user intentions and is essential to enable the user to control the robot for an extended period
the operation of any mobile brain-controlled robot. Apart from of time. In contrast to the robots under direct direction of
brain-computer interface (BCI), additional methods encompass the BCI, the second group is able to rely on the robots’
1) robotintelligence approaches for detecting environmental intelligence. As a result, it will be possible to better guarantee
conditions, localization, route planning, and avoiding obstacles user safety when operating these robots and even increase the
and collisions, and 2) shared control methods that integrate precision of human intention inference. Furthermore, users
BCI and robot intelligence to distribute control over the robot are less likely to experience weariness and use the BCI for
[33], [34].We do not review robot intelligence in this work shorter periods of time. Compared to the previous category,
in order to keep its scale reasonable, as it is a common issue these robots have the drawback of being more expensive and
affecting all autonomous mobile robots. computationally difficult due to the usage of a variety of
sensors, particularly laser sensors. created a shared-control
approach to create a brain-controlled robotic wheelchair. The
intelligent controller is meant to filter out potential erroneous
commands (like moving forward and turning left or right)
obtained by analyzing brain signals of motor imagery. As
a result, the patient can continuously control the wheelchair,
enabling a more natural interaction with it. The outcomes of
the trial indicate that there is a chance that the shared control
system can enhance overall driving performance.

VIII. M ETHODOLOGY
a) This kind of research aims to formulate the research
2) Brain-Controlled Mobile Robots: question within a specific framework by looking at upcoming
a) Based on their modes of operation, we separate mo- technology. It also aims to show future possibilities for the
bile robots controlled by the brain into two categories. One technologies, given that their growth and practical applica-
type of control is referred to as ”di-rect control by the BCI,” bility are both helpful and realistic, and that their results
whereby the robots are directly controlled by the BCI through are relevant to a certain social group. In a similar vein, it
the translation of EEG data into motion directives.Ta-ble I aims to understand the possible benefits and drawbacks of
displays several ways to put this idea into practice. Tanaka et incorporating such technology into society, from the viewpoint
al.’s work [21] is a prime example; they developed and tested a of end users to the core beliefs of humanity and our social
brain-controlled robotic wheelchair that can turn left or right structure. :
 b) To perform an exhaustive, careful, and critical as- c) Kennedy et al105 are conducting continuing clinical
sessment and achieve the finest results possible, reasonable studies on a technology that uses intracortical microelectrodes
constraints must be specified. The methodology and techniques contained in glass cones, where the neurites grow into the
of the approach, as well as the time horizon, scope, and cones to provide stable and strong long-term recording. This
strategic aim, must all be well-defined and substantiated by gadget was implanted in a patient who suffered a brainstem
innovation and scientific importance. It’s also necessary to stroke in 1998 and has locked-in syndrome. Over the four-
select the stakeholders who will be consulted. : year experiment, the patient learned how to use a computer
c) The strategic focus is on non-invasive BCI using EEG cursor. Current study intends to restore speech by implanting
with respect to three specific medical applications: motor sub- the device in the speech motor area and decoding phonemes
stitution, motor recovery, and communication/control. Because from fake speech. :
it’s non-invasive and used in medical applications, a lot of
people are involved. It is also one of the combinations that
combines the three medical applications and the most cutting-
edge technology uses: non-invasive. Knowledge becomes more
readily available as a result, which is crucial for the Construc-
tive Technology Assessment procedure. :

A. BCIs That Use Activity Recorded Within the Brain

a) A patient who is tetraplegic has an intracortical

microelectrode array implanted. A US cent with an array
of 100 microelectrodes on top. B: A microelectrode array
seen in a scanning electron micrograph. C: An axial T1-
weighted magnetic resonance image of the patient taken before
surgery. The red square in the precentral gyrus represents
the approximate location of the array. D. The patient is
working with a technician while seated in a wheelchair and
using a brain-computer interface. The gray arrow points to
a percutaneous pedestal that houses the amplifier and related
signal-acquisition hardware. :
b) In clinical trials, Hochberg et al. are using a 96-
electrode microarray implanted in the right precentral gyrus
of tetraplegic patients. These trials have demonstrated control
over a robotic arm, computer cursor, lighting, and television
by imagined arm movements. It was recently demonstrated that
accurate cursor control performance may still be attained a
thousand days after implantation.103 Currently under inves-
tigation is the use of this approach to brain-actuated FES of
paralyzed muscles and prosthetic limb control. :