A Meta-analysis of User Focused

Considerations in the Development

of Brain Computer Interface Technologies
to Enhance the User Experience

TM470
EMA
Project Report
E750136
 Contents

Section 1: Technical Report of Project

Page
Introduction: 1

Literature Review: 2

i: BCI Technologies and Development 2

ii: BCI System Testing 6
iii: Usability Design and Human-Computer Interactions 9

Legal, Ethical, Social & Professional BCI Concerns 12

Analysis and Discussion of Results 15

References 27

Bibliography 31

Section 2: Project Management, Development and Performance

Project Management Review 38

i: Review of Utilised Project Development Lifecycle 38

ii: Review of Associated Resources, Skills, Activities 45
iii: Project Risk Assessment 54

Review of Personal Performance and Future Project Development 59

i: Review of Personal Performance and Experience 59

ii: Future Project Development 60

Appendix i. Glossary 61
 Introduction

Brain computer Interfaces in development for a number of years have yet to transfer from
experimentation to a ubiquitous interface. However the classification accuracy of user generated
input commands is of particular issue due to the nature of the system to differentiate between a
command and unconscious, spurious input, impacting system performance.

Currently there has been research in the application of machine learning and artificial intelligence
to process and interpret user brain signals to generate a system input. Similar studies have
explored the utilisation of systems across varying groups such as age, sex and BCI experience to
monitor system performance. These all must be explored as the system utilises brain signals
to generate an input there is however difficulty harnessing this as the brain structure is unique
in that its structure and resultant generated signals can vary based on an individual’s physiology,
mental state and environmental inputs.

Development of BCI systems focused originally on their ability to allow those with a disability to
communicate, systems have further been developed for practical and novel applications however
a particularly prevalent issue has been one of consistent interpretation of a user generated signal
into a system output. This report will carry out a meta-analysis to explore the methods utilised to
improve system input classification accuracy, these systems, techniques and methods will be
discussed in relation to the changes in performance these conditions make. As in this analysis the
systems classification accuracy will relate to performance, Jennifer et al. (2019) attributes this
quality to a systems efficiency guiding a user through a task and effectiveness at carrying out the
desired task. These are referred to as usability criteria fundamental to the design and assessment
of system aspects as there is a relationship between usability and user experience goals.
A further usability criterion is safety, this will also be discussed in relation to various BCI
technologies, the social, medical, ethical and legal implications surrounding the use of such
devices. Key ICT aspects of the problem include the discussion of data input methods, processing,
the technologies supporting the interface and how these may be improved not only through
technological means but interaction design considerations to affect user system performance
as a human computer interaction to improve the usability of a system. By discussing this issue the
human element of human computer interactions and its potential to affect a systems performance
will be explored and compared to systems that rely on technological means. Such as machine
learning to improve performance with techniques such as combined technologies, guiding systems,
and prior user interface experience may be discussed and the best combination of techniques can
be interpreted from the meta-analysis to improve user performance and the user experience.

The report will compose of several sections including a literature review exploring BCI
advancements, discussion of a series of experimental BCI systems and a third section of the
literature report that examines usability concerns and user experience considerations with
respect to BCI development. A section will explore the legal, social, ethical and professional issues
associated with BCI technologies, its uses and the potential risks associated with integration of
these technologies to the user and society. A discussion utilising results from the systems
presented as part of the literature review will be compared, analysed and the results interpreted to
explore the ways in which guiding systems, experience, prolonged use, combination BCI system
and guidance system solutions affect user performance. From an analysis of the legal, social,
ethical, professional issues and the key metrics the optimum systems and practices can be

1
implemented to improve a user’s performance with these design considerations making this form
of user interface design suitable for the workplace.

Literature Review

This review will focus on Brain Computer Interface research with particular emphasis on
testing protocols, participant experience, user guidance systems, the combinations of various
technological and guidance systems and the impact on BCI system performance by focusing on
the human element of a human computer interaction.

This review will as such be divided into three sections:

Section one explores BCI technologies, investigating their development and providing an overview
of the techniques, technologies and processes involved as the BCI field has developed over time.
Section two will explore publications of BCI systems developments exploring technological and
user interaction considerations designed to improve user performance. The style of assessment
and the results will be discussed in detail here as these systems results and metrics will form the
analysis, discussion and results section of this report
Section three will explore human computer interface design and its relation to BCI.

The review focus’ on BCI technologies that passively and actively interpret instruction via a user’s
detectable brainwaves to operate a physical device or interact with a computer interface.
BCI technologies transcend Human Computer Interface design as it incorporates other aspects of
computing such as machine learning, software design and artificial intelligence.
It is interdisciplinary in nature covering areas such as engineering and neuroscience outside of
computing, with this consideration the report will touch upon these areas briefly while
discussing the topic of BCI testing and guidance protocols however these disciplines are not a focus
of the review.

Section 1: BCI Technologies and Development

This section of the literature review explores the development of BCI systems providing an
overview of BCI fundamental technologies and processes behind this interface type,
discussing benefits, drawbacks and areas in which future developments would improve their
overall usability enabling their transference from a research based to legitimate user interface.
This section explores these issues and how they may be overcome though there is a reliance on
technological means to provide a ubiquitous user experience.

Lin, Ko, Chang et al. (2010) review BCI integration with wireless and wearable EEG systems to
create wireless, real-time systems as they believe BCI systems are limited by bulky equipment
and signal processing of dense data caches hinders daily operation. Thirty-two BCI systems were
assessed of which seven systems featured wireless transmission. The authors discuss the reduced
time and complexity of installation, weight on the user and associated troubleshooting.
By exploring these systems Lin et al developed their own wearable, wireless BCI system to estimate
a user’s performance and drowsiness based on EEG readings it generates a warning tone alerting
the user. They believe that transferring to a wearable, wireless, real time BCI systems usability,
wearability, portability and reliability in non-controlled, laboratory-based environments could be
improved. The paper explores BCI technological developments with emphasis on utilisation

2
outside the laboratory setting and the user experience which previous papers had not discussed
however there was less detailed discussion of BCI technical specification, training, testing methods
their impact upon the user’s ability to operate a BCI system and subject results. The authors note
that subjects of the thirty-two papers may be subdivided into health volunteers, animals and
patient groups with a variety of injuries or disabilities, they note that there are no standardised
training protocols though some systems had eliminated training times. This overview of wireless,
wearable BCI systems link to this reports discussion of usability, training and guidance protocols
upon BCI performance however there was little detail exploring this and their designed system.

Zander, Thornsten and Kothe (2011) discuss the integration of a passive BCI system with
cognitive monitoring techniques, to develop a standard to interpret user commands, intention
and emotional state. The authors focus on the application and demand requirements of ’healthy’
user groups and application in activities such as gaming. They discuss non-invasive EEG based BCI
system benefits in relation to performance, ease of use and cost despite its vulnerability to
sporadic artefacts limiting research potential. They suggest that by combining techniques a
passive BCI could enhance the efficiency, effectiveness and usability of human system interactions
by enabling a system to adapt to a user by interpreting information gathered. The authors believe
that enabling a BCI to monitor user and system state human-machine system research and
methodology may develop.
The publication was intriguing as they explored novel BCI application to develop human-computer
interactions of healthy users. The piece was a review of advances in the BCI field with limited scope
in relation to the human component with no comparative data between passive and active BCI
systems or discussion of training technique of both BCI system types.

Nicolas-Alonso and Fernando (2012) discuss the advantages, disadvantages and advances in the
field, compiling an overview of BCI literature. The review includes a discussion of data acquisition
methods, user generated control signals, signal enhancement techniques to improve performance,
data processing of commands and finally a sample BCI system was selected for review.
This project will touch upon these areas though it focuses on training protocols, guiding systems
and user experience affecting BCI operation. The authors discuss the way in which research in the
field has expanded, accelerated potentially by medical field applications providing an impetus for
progress and social acceptance, interest has resulted in specialist companies developing their own
BCI systems. They discuss the need for BCI systems to simplify as they take an investment in time
to prepare, train users and calibrate. Due to their multidisciplinary nature they believe that BCI
research must establish a general framework as design variety hinders comparison and research
propagation, there are also ethical concerns surrounding the technologies ability to manipulate the
user’s brain and behaviour. Their conclusions indicate the likely possibility that with developments
the system may become useful and a practical method of human machine interaction facilitating
everyday use similar to current interfaces though the report does not explore how this could be
achieved.

Lebedev (2014) reviews BMI in relation to its potential treating neurological conditions
restoring motor function via prosthesis manipulation. The report explores BCI technologies
and the requirements such a system must facilitate to mimic human limbs however it does
not discuss how these commands would be generated by an individual though the report discusses
BCI system types in development enabling the operation of sensory, visual, auditory and motor
function restoring systems. The report provides context of BCI systems with a focus upon medical

3
application, this report helps to provide, clarity and context to my own research however it does
not discuss the human element of interaction and user design.

Cinel, Valeriani and Poli (2019) discuss the state of neurotechnologies in relation to human
cognitive enhancement, reviewing current neuroscience technologies utilised to both observe and
influence neural activities, focusing upon those related to cognitive augmentation with a focusing
application developments in the areas of communication, cognition, memory, attention, problem
solving and awareness. They focus on potential societal benefit from these technologies and
impact on future training protocols while simultaneously covering the associated ethical issues
and based upon their forecast potential future developments and areas of concern that must be
researched further. They discuss human enhancement as the techniques and approaches utilised
to augment ones bodily or cognitive functions to restore function or supersede human capability
arguing that communication is a key component of BCI systems theoretically allowing seamless
human interact with their environment and others at a ubiquitous level.
They discuss BCI applications enhancing or restoring cognitive functions in memory, learning,
reasoning, problem solving and decision making, as it develops and becomes socially acceptable
there is the potential for ‘healthy’ individuals to also benefit from the technology.
The BCI processes review covers signal recording, generation, observation and interpretation of
invasive and non-invasive systems. They state that as a communication tool further development
of methods and types of transmission are required though brain to brain communication is a viable
proof of concept which could potentially outperform traditional communication methods.
The authors found that there was a benefit of BCI usage to memory and learning capacity that
persisted up to six weeks positively impacting subject focus and situational awareness.
The authors theorise the possibility that job selection may be impacted as there is a variance in
individuals stimulation responses which can render a user incapable of operating a BCI currently
between 25-30% of users though depending upon the system utilised up to 80% of users can
reach a level of satisfactory control within a few training sessions, as user reactions to systems vary
impacting performance system, calibration developments could improve this.
Ethical issues discussed include, mind reading, privacy, safety, system invasiveness of
implementation and brain stimulation, societal effects of a technological reliance, issues
of agency, responsibility and liability are also discussed as concerns associated with the
implementation of BCI systems. The authors theorise that medical and technological and
social developments will occur creating an acceptance of invasive technologies.
This paper provided a comprehensive review of BCI developments and considerations of ethical
issues associated with BCI implementation and areas requiring further development providing
context of developmental trends for this report specifically the benefit of communication based
BCI. They discuss BCI illiteracy, how to overcome this by developing systems tailored to user
responses, when discussing this feature the authors mention projected user training timeline to
developed operational BCI skills to an acceptable level however they did not discuss training
methodologies, effects or how this aspect could be improved, they see the user as a passive
element of the HCI relationship.

Martins, Angelica, Chakravarthy et al. (2019) review the state of BCI system development and
theorise the integration of cloud based computing to create portable, user friendly devices, they
believe integration of these technologies will be safe to implement within the next twenty to thirty
years to implement a safe, secure, instantaneous BCI system utilising cloud technologies to enable
real time interfacing between the brain and computer systems implementing cloud based

4
information processing as recorded signals are transported wirelessly to cloud based architecture
and a resultant signal is returned triggering an order to the system. Currently much of the
discussion surrounding the paper involves theoretical nano devices to implement their fusion of
BCI and cloud-based systems however the implementation of cloud-based information processing
solves an issue relating to systems usability. The article theorises that the human element will be
nullified by technological advancement and acknowledges training and BCI illiteracy as a limiting
factor hampering the technologies potential though current training protocols are not assessed.

Mudgal, Sharma, Kumar et al. (2020) discuss the implementation of BCI technologies in the
medical field postulating its capacity to revolutionise healthcare systems as a communication
method non dependent upon muscle or neuromuscular paths. Their discussion reviews current
advancements, BCI system types and highlights the early detections of abnormalities from injury,
sickness and disease can be read from signals due to the prevalence of EEG based systems and
by learning to modulate neural outputs users can manage mental health symptoms improving
operation performance. Issues such as security, privacy and confidentiality of user data, ethical
issues discussed include risk of surgical complications, subjects physical and mental safety are of
concern as there is potential for altered behaviour affecting their emotion, personality, memory,
cognitive, emotional and physical states and how do medical practitioners acquire consent from
users with difficulty communicating. They discuss BCI illiteracy factors that can impact a user’s
ability to interact with a BCI including attention, memory, fatigue, stress, emotional state, age,
gender, lifestyle while also contending with the complexity and diversity of the human brain
structure. They find that though the technology is still experimental there is still great potential
with the capacity to change the services provided to patients though user compatibility issues
must further be mitigated.

This section of the literature review provided context of BCI system development over time the
literature focused away from restorative uses to a holistic system for work, communication,
augmentation and entertainment. This analysis allowed the comprehension of issues and
advances as time progressed, there are still several issues that prevent the establishment
of a ubiquitous BCI system as the dominant form of human-computer interaction. Issues include
system limitations such as portability, error rates and data processing though it has been theorised
that with a stable connection to cloud infrastructure to carry out processing this could be one
solution while facilitating the shrinking of non-invasive systems increasing portability however
preventing sensors from shifting on the user raises other concerns. Other issues include user
physiology due to BCI illiteracy, as their brains do not generate expected signals despite training
or external stimulation,
as such current research is exploring the utilisation of machine learning to tailor BCI system to the
user by self-calibrating to the individual to reduce the BCI illiteracy rate, while potentially reducing
or removing user training. Currently user training can take multiple training sessions over a
number of weeks, some experiments involve those with little to no experience of BCI systems and
a variety of training methodologies to test BCI systems, there have also been advances in the field
enabling the growth of systems capable of being tested outside of laboratory settings in settings
resembling real world scenarios.

5
 Section 2: BCI System Testing

In this section of the literature review I will be discussing the BCI systems I have chosen to include
as part of this reports examination of BCI user experience in relation to performance measured via
user classification accuracy, system training protocols, user experience and guiding systems impact
on BCI usability.

Technological Solutions to User Performance

Musk, Neuralink (2019) and Milekovic, Sarma, et al (2018) each explore issues regarding BCI use,
performance and limitations of the technology so far and present ways in which these may be
overcome through technological means. Musk, Neuralink (2019) discuss the implementation of a
high bandwidth invasive BCI utilising flexible polymer probe electrode arrays implanted by a
surgical robot. The proposed system has been utilised on rodents with the aim of future human
trials and eventual goal of a fully implantable system hermetically sealed with the capacity to
wirelessly power and transfer commands without exposing the system. The system would enable
custom tailoring of a user’s BCI sensors able to target specific regions of the brain and avoid
vasculature, reducing the bodies potential to reject the implant and risk to the user compared to
other invasive BCI systems. This could potentially improve the social acceptability of invasive BCI
and enables the targeted implantation of sensors that could lower the BCI illiteracy rate. Milekovic,
Sarma, et al (2018) instead discuss BCI limitations with regard to the recalibration and
maintenance required to retain system performance, the authors developed a BCI speller system
that while slow to operate proved capable of stable system requiring little intervention over an
extended period of time. The two individuals participating in the study underwent daily, hourly
training sessions over five days, one users system functioned for seventy-six, the other 138 days to
communicate via email, with family, and carers through an invasive BCI, demonstrating the
potential for a stable long term system with minimal interference and recalibration.
Choi and Bang (2013) discuss the implement integration of a gaze-tracking and an EEG based BCI
speller, there were ten study participants, it is not mentioned if they have prior BCI experience.
Training involved two sessions and forty tests of EEG, gaze-tracking and combined system trials.
Gaze based testing achieved 15.25% accuracy, the EEG based system achieved 62.25% accuracy
while the combined system achieved a success rate of 86.5%, considering these results and results
from the vibrotactile EEG based system developed by Chatterjee and Aggarwal (2007) there is a
performance enhancement and improved usability by overlapping BCI systems to establish an
accurate reading of user generated input data.

Training Protocols

Jeunet and Jahanpour (2016) and Rosca and Leba (2018) test three BCI systems to test users ability
to learn and retain operational control of these systems based on the utilisation of current BCI
training methodology. The BCI systems tested by Jeunet and Jahanpour (2016) are EEG and
other motor image based, fifty-four subjects took part in the first experiment while the second
comprised of the ten best and worst scoring subjects, they attempted to learn a new motor image
task.
The BCI system presented by Rosca and Leba (2018) applied an EEG based BCI enabled drone
translating EEG based brain signals into drone commands. Four mental tasks were trained to

6
define lift, drop and left/ right pitch they discuss the requirement of further discussion of training
protocols, drone usability, performance and continuous cognitive strain experienced.
Jeunet and Jahanpour (2016) theorised from the results that standard training protocols are
lacking and users individual spatial ability impacts their performance these issues tied with the BCI
illiteracy rate projected to be between 15-30% of the population was reflected in the pass and
failure rate of their initial experiment.

The Role of Guidance Systems

Chatterjee and Aggarwal (2007) present an EEG based BCI system that explores the effectiveness of
vibrotactile biofeedback in a BCI system involved six subjects, four were inexperienced.
The task involved the movement of a cue from high to low by using the predefined motor imaging
task, high involved visualised right hand movement while low visualised left hand movements with
the attempt begun at the midpoint users failed the attempt by not completing the task in 15s or
placing it in the opposite position. Users were initially trained with both vibrotactile and visual
interfaces however when collecting data only the vibrotactile interface was utilised.
The training session also served to calibrate their classifiers and tune signal modulation.
Performance was based upon accuracy, the percentage of completed trials and the latency
defined as the time taken to complete a task successfully. They found that by introducing the
system with visual feedback training performance was improved and speculated that if untrained
users had more training they could further improve the average accuracy of 56% which was
however much higher than the estimated 15% chance of random success. From their results
the authors discuss the need for appropriate training protocols, subject feedback and further
developed signal processing techniques to improve the system. It should be noted that from
the experiment results the two BCI experienced users had not outperformed any of the other
users this I hypothesis is due to their unfamiliarity to vibrotactile based BCI system feedback.

Combined Feedback

Ramos-Murguialday and Schurholz (2012), Choi and Bang (2013) and Gargiulo and Mohamed
(2012) utilise various BCI systems to observe any differences to user performance based on the
type of feedback presented. Ramos-Murguialday and Schurholz (2012) designed a non-invasive
EEG based BCI system along with a robotic exoskeleton hand capable of flexing and extending its
fingers, the experiment involved twenty-four subjects, split into three groups completing five tasks
involving differing combinations of motor imagery and movement to move the robotic hand.
Of the groups one group received positive, the second group received negative and the third sham
feedback, the group’s performance was then compared. Gargiulo and Mohamed (2012)
investigate the role of feedback presented to subjects in audible, visual and combined
auditory-visual forms as they operate a cursor, there were eleven subjects three of which had
previous BCI experience. Performance was based on the average successes, failures and timed out
trials then compared
to a hit rate of 50%. From subject results it was noted that combined audio-visual feedback had a
positive effect on performance and eased system frustration as 75% of subjects displayed an
improvement under testing. The article showed the importance of feedback and that there was a
holistic effect on the subject integral to BCI systems performance as performance is co-dependent
upon user mental state. Ramos-Murguialday and Schurholz (2012) theorised that though
system-user calibration and training feedback impacts BCI operational performance they found

7
those exposed to sham and negative feedback showed no significant improvement in performance
when compared to the group exposed to positive feedback as the holistic effect of accurate
feedback has a holistic reinforcing effect on the subject.

Human-Centric Design

Schreuder and Riccio (2013) discuss the importance of a user-centred approach to BCI system
implementation as each has their own strengths and limitations however the user themselves are
an integral component as they can in terms of performance react differently. In this article the
authors utilise a single subject with motor and communicational deficits, it took two studies to find
a system suitable for the subject to utilise. Both were ERP based, the first auditory system was ill
suited, however the visual ERP based system was successful. They discuss BCI system tailoring to
a user, making the final system choice based on the users clinical profile, feedback and testing
different BCI on a user to achieve an efficient, suited BCI for the end user.
Brouwer and Erp (2008) postulate that P300 based BCI are potentially better suited to tactile
over visual or auditory stimuli though most of these systems utilise visual stimuli.
To simulate a realistic setting eleven BCI inexperienced subjects were presented a target and
‘distractor’ stimuli, they must then indicate when a target stimulus is presented by pressing a
button. The authors stipulate that classification accuracy is of greater significance than the
generated signals strength. The procedure was explained prior to the beginning of the experiment
to subjects, there was no training session to orientate them to BCI system use, as a result subjects
average a 62% success rate however they feel that this could be increased with training and
calibrating systems to the users to increase performance a conclusion that Schreuder and Riccio
(2013) had also come to with regards to the performance benefit associated with tailoring a
system to the user’s needs. Tactile stimuli based BCI utilise differing neural pathways from auditory
and visual systems interference is minimal when exposed to these stimuli in realistic settings.
This paper was an interesting examination of interfering stimuli on users operating a BCI in a
simulated realistic scenario a user would encounter if the technology was implemented however
the experiment was carried out in a lab setting with users sitting and making minimal movements
to reduce interference counter-intuitive to their goals.

Impact of Prolonged BCI Use

Nierhaus, Vidaurre, et al (2019) and Mathe and Spyrou (2019) utilised EEG based Visual ERP and
motor imagery BCI to measure the physiological and cognitive strain presented by prolonged use
of BCI systems. Before and after Nierhaus, Vidaurre, et al (2019) hour long experiment the
twenty-three subjects had an MRI, the results had shown that where the neural signals had been
generated to operate the BCI there was a marked increase in induced brain plasticity, affecting
brain structure and function, they discuss the unknown implications of long term BCI use.
Mathe and Spyrou (2019) system operated a single input operated toy car, the experiment
evaluated twenty BCI inexperienced subjects to move the toy car. Subjects attenuated their
attention, surpassing a limit to initialise the toy, speed was then proportional to the operators
attention value. The goal to move the toy three metres there was no time limit and a two-minute
break between each of ten attempts. Mathe and Spyrou (2019) found that completion times
decreased until the seventh trial where there was a significant spike for two trials before the final
trial time again decreased. After discussion with the subjects they postulate that this spike was due
to fatigue and knowing they were completing the last trial as such they believe that while BCI

8
technologies have great potential associated fatigue and anxiety caused by continual use can
negatively impact some users in daily operation. They note however that 70% of subjects liked the
idea of interacting with devices through their minds with 75% of subjects wishing to participate in
future BCI experiments from the post experiment subject survey.

The Role of User Experience

Chen and Wang (2015) and Renton, Mattingley and Painter (2019) researched and implemented
BCI based spellers while Chen and Wang (2015) had utilised a mixture of naive and trained
subjects to simulate natural conditions, their results demonstrate the system achieving a character
selection rate approaching sixty per minute though the authors note subject bias as they were
aware of the technology and character placement Renton, Mattingley and Painter (2019) believe
that in such testing of Speller systems this bias is a problem as they found in their meta-analysis of
BCI spellers a regular occurrence of studies involving subjects repetitively typing single phrases
which they feel does not equate to realistic system application. As such they developed their own
non-invasive BCI speller to test utilising seventeen inexperienced BCI system subjects prompted
with a topic to express themselves through word association. The study had shown that their
words per minute were slower when tested in this manner, the authors theorise that future
developments require user centred design maximising reliability and control over performance
while considering the implementation of hybrid BCI systems to improve usability.
While Chen and Wang (2015) experiment demonstrates a bias I believe this replicates a subjects
habitual use and experience when utilising a system. Further developed user focused design
could see an improvement in user comfort to supplement protracted use potentially improving
performance and user experience over long periods as discussed by Mathe and Spyrou (2019) and
Nierhaus, Vidaurre, et al (2019).

Section 3: Usability Design and Human-Computer Interactions

This section of the literature review explores developments in human-computer centric design
and how to incorporate the user in the developmental phase of system development.
This also discusses the role of guidance systems on the user and training protocols to improve
the user experience which this has been defined based upon the literature to compare against
the BCI systems of Section Two.

Human-Computer Interaction

Kim(2015), Rocchi(2016) and Sharp, Rogers and Preece (2019) each discuss the fundamentals
of Human computer interaction design. They cover topics such as the human factor, system design,
interface design, the user experience so that user psychological, social and usability needs can
be transferred into system requirements that may be implemented throughout development
to deliver an interface that Kim (2015) describes as efficient with high usability as HCI desires
the consideration of system interactions and interfaces that are developed in a natural fashion
enabling truly ubiquitous unobtrusive interactions between human and interface, this can be
applied to BCI development as it utilises an individual’s brainwaves to interact with an interface
and could be construed as an example of an unrefined, naturalistic interaction however further
refinement in system performance could be found by integrating HCI fundamentals to affect

9
performance.

While Kim (2015) discussed the individual users importance in HCI development Rocchi (2016)
examines the users social needs and obligations in respect to HCI development by considering
the various use-cases under which an interaction may be instigated by transferring these user
needs into requirements and how this impacts an interface when individuals are interacting with
an interface for example in the work place or leisure the users goals, needs and expectations
change.
Sharp, Rogers and Preece (2019) provide an in-depth discussion of interaction design of human
computer interfaces with reference to best practice implementation of user experience design.
Concepts needed to be considered for HCI implementation in system development such as user
experience, accessibility, inclusivity, and ubiquity are discussed while also also covering topics such
as usability testing, selecting a type of evaluation and organising a study. These concepts will be
discussed and applied to BCI systems to discuss how system performance is evaluated, the impact
of prior BCI experience on performance and the role of guiding systems.

User Experience

Kogel, Jox and Friedrich (2020) and Bos, Gurkok, Laar et al. (2011) each discuss an issue with a
component of a BCI, the individual as though technology progresses and algorithms become more
complex to interpret an individual’s brain signals into commands there are across the field of BCI
many numbers of invasive and non-invasive systems that utilise markers from various areas of the
brain to interpret a command. Both papers discuss the need to implement the users experiences
and reactions to a BCI in the selection of a particular system as users may find an affinity to one
form of system over another due to their physiology, mental state or experiences as such Kogel, Jox
and Friedrich (2020) instigated a semi-structured interview study with BCI users utilising these
systems to restore function, to discuss the social and ethical issues while assessing the users
experiences, observations and attitudes toward BCI that are part of their daily lives and how they
have impacted them. This provided an interesting perspective that examines the users experience
with the technology and capture their thoughts on these systems. While Bos, Gurkok, Laar et al.
(2011) Based their assessment of a user’s experience on a meta-analysis of 105 papers to develop
key markers that BCI systems should be assessed, user experience being one that can be measured
based on a systems efficiency, learnability and satisfaction along with the typical measure of bit
rates and classification accuracy. The authors mention briefly that from their meta-analysis some
studies note a potential relationship between user motivation and performance considering it an
area requiring further exploration.

Feedback

Kosmyna and Lecuyer (2017) and Barbero and Grosse-Wentrup (2010) each discuss HCI principles
in relation to guiding system development, the integration of feedback bias and its effect on user
performance, at the time Barbero and Grosse-Wentrup (2010) found that positive or negative
feedback bias had a negative effect on individuals performance especially if there multiple types
of guiding system they suggest that guiding systems and bias should be tailored to a user’s level
of experience. They based their results on a study that found all but one subject had a negative
reaction to performance due to any form of feedback bias in the study had used a BCI and exposed
subjects to differing forms of feedback bias they found that this impacted their performance

10
negatively, though the paper did not discuss why these changes in feedback did not affect
the individual subject with no prior BCI experiences performance. While the Barbero and Grosse-
Wentrup (2010) study investigated guiding system bias Kosmyna and Lecuyer (2017) discuss the
implementation and user performance impact of feedback and feedforward guiding protocol
implementation in a BCI system. They note that there is a bias associated with these forms of
guiding system and that these HCI methods can be implemented to affect system performance.
The article explored feedback and feedforward guidance systems as a proposed method of
improving BCI performance and design, they explored the concept of integration of these guiding
solutions into an established system however there was no implemented testing to explore the
relationship between guiding solutions and performance.

Social Pressure

While Kogel, Jox and Friedrich (2020) and Bos, Gurkok, Laar et al. (2011) discuss the individuals
skills, experiences, physiology and mental states effect on performance and need for
consideration in BCI design and testing. Battarbee and Koskinen (2010) also explore the concept
of user experience being a key part of interface design to improve a products functionality and
usability, the paper further introduces effect of co-experience on the individual, how social
interactions impact a user’s own experience as this does not occur in a vacuum.
Co-experience can affect a user’s interaction with a BCI and their experience with a system.

Training

Lotte, Larrue and Muhl (2013) apply the educational psychology and instructional design concepts
to critique human training protocols in BCI testing as they believe it is not simply a matter of signal
quality and processing affecting performance but the human element as BCI use is considered a
skill. To train users they either learn to generate the necessary signals to interact with the system
or machine learning is utilised to optimise the system to the user, from assessing a number
of papers they found that instruction design guidelines for training approaches were flawed.
Their paper explores the roles of feedback, instructions and the associated tasks performed
by subjects. Their piece did not compare system performance across studies based upon their
hypothesis however they note the importance of a refined training protocol to improve BCI
performance.

11
 Legal, Ethical Professional and Social Issues Associated with BCI

This section will explore the legal, social, ethical and professional issues associated with BCI
technologies, the uses and potential risks associated with integration of these technologies
to the individual and society as this interface type develops to ubiquity. This will be done by
exploring each issue in turn, discussing the ethical and professional involved in the development
and testing and security requirements of BCI. Social considerations such as the potential abuses,
design considerations and the potential legal concerns associated with BCI technologies use in the
workplace.

Ethical and Professional Issues

Current research into BCI systems and uses have expanded tremendously since its implementation
as researchers explore the various ways in which the human brain may be harnessed to generate
a system output. This is done by measuring various signals generated by the brains various
neural systems both passively or invasively, these methods are further described in the Glossary
document. There have been developments in brain to brain and system based brain stimulation
along with brain to system interactions, these developments raise a number of ethical and
professional issues in relation to the research, development, testing and implementation of BCI
technologies.

When researching and developing a system utilising BCI technologies there are benefits to
utilising invasive systems such as improved classification accuracy of user input commands.
Higher information transfer rates with less interference can be achieved as signals are received
from sensors surgically implanted in the subject. While there are benefits to this method of input
data capture from the user Glannon (2014) expresses concerns as to the user’s safety as they
require neurosurgery to implant the sensors and at least a single incision from which data is sent
to the system and the sensors powered.
There are professional and ethical implications associated with this interface type with risks
associated with surgery and secondary issues such as infection and implant rejection. Due to these
risks there is especially in the development of novel and communication based system testing a
preference toward non-invasive, passive sensor based BCI systems to limit the potential harm to
those associated with the testing of BCI systems.

BCI represent a data protection and ethical concerns as implantable systems integrating biological
and technological components are developed that blurs the line between man and machine with
the potential to alter human culture as the technology reaches for ubiquity. When developing a
system in the modern age there are several unscrupulous individuals and groups that engage in the
distribution of malware, ransomware, phishing and cyberterrorism. Martins et al.(2019)
discuss the need for protections as security issues of this nature could have disastrous results for
the user.

As the technology utilises brainwaves to detect and interact with systems that are always scanning
for a command, this information could then be intercepted and interpreted by the system without
a user’s knowledge. As there are direct readings of brain waves taking place, data could be
required when processed when data is processed to be anonymised, preventing user tracking
similar to the way in which users opt in and out of HTTP cookies that create targeted advertising.

12
Some BCI technologies involve the bi directional flow of data and external stimulation of the brain
where a compromise or failure of the system will have lasting or fatal results without protections in
place.

Social Issues

Cinel et al. (2019) utilised forecasting techniques to discuss current neuroscience technology
developments and societal reactions and how this may change by 2040 speculating a positive
reaction to BCI systems that are non-invasive and wearable in nature becoming a common
occurrence. In this section the potential negatives associated with the commercialisation of
BCI systems will be discussed in relation to its widespread adoption into everyday use.

While those adopters of the technology will learn how to utilise these devices to interact with
a variety of interfaces in work and through their personal lives experiencing a system that provides
a universal interface operated through the generation of specific brain signals.
Enable those currently incapable of effectively utilising current system interfaces to their full
potential. Burwell et al. (2017) express concern for those incapable of utilising BCI or those
unwilling to operate BCI due to agency, trust and security issues. This could potentially deepen
the BCI illiteracy rate despite developments to increase user compatibility with a universal BCI as
many forms of technology are met with low confidence and scepticism such as societal reactions
to smart phones, tablets and wearables. Users may find that much like VHS and Betamax or Blu-ray
and HD-DVD they are left with expensive technologies that are considered legacy hardware that is
unsupported technologies or installation practices becoming ostracised from other users.

Research by Nierhaus et al. (2020) revealed the potential alteration of brain plasticity until its
extent can be established rules should be considered so that the development of children and
adolescents are not negatively impacted. BCI training could be considered a skill that would be
regulated. BCI use could be limited however by regulation much like a driving license with users
subject to assessment of their cognitive function so any incidents fault may be assigned to user
error, negligence or system fault. As automated work processes become a regular occurrence there
would be an increase in demand for those trained to operate BCI as a skill. Those going through
the employment process could find themselves excluded from selection due to inexperience, lower
performance utilising BCI in general or due to an employer utilising a particular BCI system.
While working the data interpreted by the system provides a time lapse of users brain activity
through the working day such that individual metrics may be monitored for periods of inactivity
which could result in employees becoming reprimanded and their rights to privacy being infringed.
By being able to interface with systems remotely the lines between work and social lives could be
blurred as processing of data and tasks are completed in the background.

13
 Legal Issues

Krausova (2018) discusses the way new technologies impact the established societal frameworks
of morality and value as they are confronted by the legal system, she describes its role as one of
technological risk management. When discussing the implication of implementing BCI in this way
there are conditions that must be met to ensure societal stability. As such when technologies are
developed ethical, professional and legal concerns can be applied and when necessary adapted to
mitigate the posed risk. These issues will be discussed briefly to explore the legal issues associated
with BCI in the workplace.

BCI particularly those that are being utilised as a restorative aid as opposed to a tool can find that
there is a conflict as to ownership and the self as the line between man and machine are blurred.
Drew (2019) described a patient suffering from chronic seizures that through a BCI monitoring
their condition would deliver a DBS neutralising an impending seizure, drastically improving the
patients’ quality of life. The company developing this treatment had become bankrupt as such
despite challenging the decision, the patients BCI was removed, the seizures returned, and she
grieved the loss of the existence taken from her.

For BCI to enter the workplace and society at large the technology must meet the standards and
regulations set in terms of performance and safety for the user and those around them with the
implementation of appropriate measures and practices. It is noted that this must be handled
delicately as over regulation can impact the dissemination of BCI while under regulating can result
in situations that present BCI negatively and its growth instead a balance must be found.

Bulbitz et al. (2019) discusses the distinction between human and computer merge with BCI and
the concept of utilising one’s mind to execute actions. If there is a fault or incident when
attributing blame there are issues associated with attributing the burden of truth, liability, and
establishing whether the operator had carried out and if they had intended to commit the crime.

Ofcom (2020) have stated that in the UK those with no internet access have limited opportunities
to participate in society fully as such it should be considered a basic right for all citizens.
Internet access as a basic right similarly as BCI technologies permeate society this too could
occur as traditional methods of interacting with systems become antiquated and phase out of use.
Leaving those unable or unwilling to harness a BCI ostracised from society without suitable
legislative protections.

There is an issue surrounding the concept of freedom of privacy, thought, speech and expression
as BCI propose the potential to monitor private thoughts altering the way an individual expresses
their inner self. Krausova (2018) notes that BCI require trained expressions of intent by the user
to operate a BCI impacting the way an individual thinks and expresses themselves, limiting their
latent capabilities to conform to societal pressures, seceding other individual freedoms in the
process.

14
 Analysis and Discussion of Results

In this chapter the performance of BCI systems discussed in the literature review will be utilised
to analyse the results and compare them to several metrics. These metrics and results will then be
analysed to present the impact of these systems and protocols to highlight the
best practices to improve the performance of the individual. This is in contrast to research into
the application and improvement of machine learning algorithms and artificial intelligence to
better process user input data to improve system performance.

The topics that will be discussed include the impact of experience on BCI performance, guiding
solutions utilised by BCI systems, guidance bias and the impact of combining guidance systems.
The types of assistive features built into these BCI systems including data processing techniques
and design considerations impact on user performance will be discussed. Finally a discussion into
the medical, professional and ethical concerns associated with prolonged BCI usage.

Number of System System

Study
Subjects Test Type
Lin, Z., et al. (2018) 13 BCI Speller P300

Naeem Mannan, M., et al. (2020) 20 Hybrid Speller SSVEP & Gaze

Nagel, S., Spuler, M. (2019) 9 BCI Speller SSVEP

Han, C., et al. (2018) 18 BCI Speller SSVEP

Ryan, D., et al. (2017) 36 BCI Speller P300

Xing, X., et al. (2018) 11 BCI Target Accuracy SSVEP – Dry Electrode

Speier, W., et al. (2018) 12 BCI Speller & Predictive Spelling P300

Volosyak, I., et al. (2017) 20 BCI Speller SSVEP

Renton, A., et al. (2019) 17 BCI Speller SSVEP

Chatterjee, A., et al. (2007) 6 Haptic Feedback EEG

Choi, J-S., et al. (2013) 10 Hybrid Speller P300 & Gaze

Chen, X., et al. (2015) 12 BCI Speller SSVEP

BCI Combined Audio-Visual Feedback

Gargiulo, G. D., et al. (2012) 11 EEG
Test

Jeunet, C., et al. (2016) 54 BCI Training Protocol Review EEG – Motor Imagery

Barbero, Á., & Grosse-Wentrup, M.

11 BCI Feedback Bias on Performance EEG – Motor Imagery
(2010).
Mathe, Eirini & Spyrou, Evaggelos.
20 Prolonged Use of BCI EEG
(2019)

15
Figure 1.: Presents the BCI systems that are being utilised for the analysis and discussion section.

Figure 1. contains a record of the BCI systems utilised in their respective studies detailing the type
of BCI utilised, the type of system test and the number of involved subjects testing the BCI.
The systems being analysed are predominantly BCI Spellers as these forms of system provide a
realistic use case as to the implementation and performance of a BCI in the workplace.
BCI spellers were also chosen as there was a consistency to presentation of system results as such
performance of subject BCI practices, protocols and system performance will be with reference to
subject classification accuracy.

The Role of Experience

This meta-analysis involved a sample size of sixteen studies incorporating 278 subjects testing
experimental BCI systems, predominantly spellers or target selection for menu navigation.
All systems included in the analysis were non-invasive BCI systems as this is the style that research
is more focused due to the difficulties associated ethically with the testing of invasive systems and
social acceptance. There was also no mixing of invasive and non-invasive systems so that the data
analysed would not be skewed as one of the benefits from invasive systems is a higher degree of
classification accuracy.

The subjects across the studies were separated into three groups, trained, untrained and
unspecified. There were thirty trained, ninety-four untrained and seventy-eight unspecified
experience subjects. Their performance was then compared to ascertain the impact BCI experience
has based on a user’s experience as such the unspecified experience subjects will not be examined.

Figure 2.: Depicts the classification accuracy of subjects that are either trained, untrained or unspecified.

16
From the results presented in Figure 2 it was found that the untrained subjects achieved an
average classification accuracy of 81.126% which was 4.91% lower than the trained subjects
attaining an average accuracy of 86.036%. Though in the trained subject data set there was an
outlier, subject Nine that had an accuracy of 40% which was less than half the group average,
other trained subjects had also achieved performances above 70%. This outlier brought the trained
group average performance down by 1.587%, the outlier also negatively impacted the standard
deviation of the dataset resulting in a SD of 11.61 instead of 7.829. Whereas the untrained group
attained a SD of 18.389 with subject accuracy ranging from 22.62-100%. These results show that
when considering experience affects user performance as the trained group achieved a higher
classification accuracy with a tighter spread of individual results.

The Role of Guiding Systems

Guiding systems are utilised in BCI design to guide the user through system interaction. This is
achieved either by interpreting user input and presenting this back to the user or by presenting the
available actions to the user for them to input. These forms of guiding system frameworks are
known as feedback and feedforward they present users two different ways in which to interact
with a system. The way a guiding system is designed and implemented can impact the usability of
the BCI systems in this meta-analysis by altering the human-computer Interaction this can be
presented visually, auditorily or as a tactile stimulus. Four metrics that will be discussed in relation
to guiding systems impacting usability includes bias, combined guiding systems, combined input
systems and assistive features to improve user performance. This is instead of utilising data
processing, interpreting and hardware to improve system performance.

Figure 3.: Performance change from testing experimental algorithms to improve usability.

17
There was an average classification accuracy improvement of 6.36% from 79.32% to 85.68%
through the use of experimental optimisation algorithms. All subjects had experienced the same
or improved performance except for subject Seven who experienced a performance drop of 0.5%.
Subject Twenty-four’s performance was an outlier who had achieved an unaltered classification
accuracy of 22.62% and 26.19% optimised. The optimisation techniques used resulted in a tighter
SD going from 19.93 to a SD of 18.42, this along with the group classification accuracy improves
when the outlier is removed from the data set. This data shows that through improved data
processing and interpretation techniques user performance and system usability can be positively
influenced.

Guiding System Bias

When providing a subject feedback the way in which it is presented to the user can affect their
performance as they operate a BCI system. These biases that may be implemented into a guiding
system can be described as neutral, strong positive, strong negative, weak positive and weak
negative.
Barbero and Grosse-Wentrup (2010) discussed how this bias in feedback influences the
performance of subjects naive to BCI, Figure 4. presents the results of this study.

Figure 4.: Results of the naive subject feedback bias study conducted by Barbero and Grosse-Wentrup (2010).

To differentiate between a command and guess a threshold was placed on subjects performance
as each bias type was tested, subjects successfully passed this threshold across each bias type.
It was found that when training naive subjects a strong positive bias provided the best
improvement to their overall performance followed by a weak negative feedback bias.

18
This could be due to a number of possibilities, subjects apprehensive and inexperienced with
BCI are quelled by a strong positive feedback bias while those exposed to a strong negative type
of feedback bias become disillusioned and is reflected in their performance.
Subject reaction to a weak negative bias was surprisingly positive as users were not dissuaded
by the feedback instead they achieved a classification accuracy of 67.90%.

Kosmyna and Lécuyer (2017) and Lotte, Larrue and Mühl (2013) discuss the importance of
feedback bias on user performance. As a user becomes more proficient with BCI operation
the role and bias of feedback utilised to train a user from novice, intermediate to an experienced
must adapt. Inaccurate positive feedback biases begin to negatively impact user performance
as their experience level increases.

Combining Guiding Systems

A strategy to analyse that could affect the performance of the user with a BCI would be the
combination of multiple forms of guiding system. This multi-layered solution would provide
users a comprehensive level of feedback as they assess their performance interacting with a BCI.

Gargiulo, G. D., et al. (2012) researched this in relation to a combined audio-visual feedback
guidance system and the performance changes of both naive and experienced BCI operators,
the results of which are presented in Figure 5. and Figure 6. for further discussion.

Figure 5.: Depicts the individual trained subjects average combined, best combined classification accuracy and the
best visual feedback accuracy and compares this to the group’s performance.

19
From the results depicted in Figure 5. the three subjects with prior BCI experience had achieved
classification accuracies utilising the combined feedback system that were lower than their best
single feedback system. Subject three had scored the closest best combined feedback score that
was less than two percent lower than their best single system performance. It is noted however
that the subjects found the additional feedback from the system helpful and found themselves less
cognitively strained by the sessions.

Figure 6.:Depicts the individual naive subjects average combined, best combined classification accuracy and their
average and best visual feedback accuracy then comparing this to the group’s performance.

From the results documented by the naive subjects all bar Subject 1 achieved an improvement of
their average classification accuracy score utilising a combined feedback instead of the single
feedback system. Also of note the subjects all achieved higher classification accuracies when
utilising a BCI with combined feedback systems, Subject 4 had achieved a best combined accuracy
score 35% higher than the single feedback system.

Naive subjects had with the extra information provided to them through the experiment had been
able to modulate their own performance as they interacted with the BCI. It was interesting to note
that the experienced subjects had achieved lower classification accuracies with the combined
feedback system. Combined feedback systems would be a useful training tool before moving users
to less intensive guidance systems as their experience grows.
Combined BCI Input Systems

A BCI utilises sensors to detect and interpret a user’s generated brain signals to execute a
command. System performance of a BCI is dependent on the accuracy to which a user’s command
is detected as such there has been research into multi-input system BCI. There are two hybrid BCI

20
spellers included as part of the meta-analysis, both track users eye placement combined with
SSVEP and P300 to improve the overall performance of these individual systems.

Figure 7.: Depicts the Classification accuracy of ten subjects utilising each system individually and their combined
system accuracy from Choi, J-S., et al. (2013).

As can be seen from Figure 7. gaze based systems produce low classification accuracy results with
none of the subjects achieving a result of 27.5% with this system type. Often system tests such as
Barbero and Grosse-Wentrup (2010) establish a limit to differentiate between guesses, spurious
input and commands, in their case it was 63.90%. Meanwhile the supplemental P300 system
enabled subject classification accuracies between 42.5-80%. When these systems were combined
however subjects achieved accuracies between 82.5-92.5%. This study exemplifies the potential
benefits associated with combining input systems to improve user performance however there are
issues with system bulk, troubleshooting and maintenance of such systems.

21
Figure 8.: Depicts the performance subjects utilising combined system inputs had achieved in comparison to the
average accuracy of single input BCI systems.

Subject 27 was the only combined system user to achieve a classification accuracy below the single
input system average of 80.48%. This average was also below the 88.73% average accuracy
achieved by those utilising combined BCI input systems. Though these results present the
advantage of higher accuracy rates associated with combined systems the economic cost
associated with installation, and maintenance of these systems. Users may find that their
performance varies based on the combination of BCI input systems utilised due to individual
neurological propensities to differing forms of BCI input monitoring.

Assistive Features

From the studies included in the meta-analysis there were two BCI speller systems that
implemented features meant to assist the user and improve their performance.
Ryan, D., et al. (2017) explored the impact of system interface changes when testing a
checkerboard style BCI speller. Their thirty-six mixed experience subjects used three styles of
speller interface, grey and white, grey and colour and a colour saturated checkerboard.
This change affected the groups average accuracy, the washed-out grey-white and colour saturated
interfaces achieved 81.17% and 81.81% with a standard deviation between 15.57 and 12.53
respectively. This showed that for BCI spellers interface design is instrumental to performance
and improving the user experience as the washed out and over stimulating colour saturated
interfaces performed poorly.
The combined interface however of grey and colour resulted in an average accuracy of 85.97%
with a standard deviation of 9.67. The group utilising this interface consistently achieved a higher
level of accuracy than the other two interfaces as the balance of colours selected to differentiate
between inputs was greater.

22
Speier, W., et al. (2018) implemented a feature into their BCI speller that has been implemented
in devices such as mobile phones, tablets and search features across services such as a Google
search, that being predictive text. The study compared the subjects average accuracy with and
without the predictive text suggesting the users following letter and word spelt.

Figure 9.: Twelve subjects average character accuracy with predictive text enabled and disabled, group averages of
each test are also presented.

From the results a quarter of subjects had seen a positive improvement to their average character
accuracy, another three were able to maintain their 100% accuracy rate. The second experiment
involved the implementation of predictive word suggestions based on user input.

Figure 10.: Average word spelling accuracy with predictive text enabled and disabled.

23
Subjects Two, Ten and Eleven retained their 100% classification accuracy, unlike the previous
experiment the implementation of word suggestions has resulted in four subjects improving their
accuracy. The group’s overall accuracy with the predictive text enabled was 2.7% higher than when
the feature is disabled. The poor result utilising predictive character selection due to a poor
learning model that is too generic whereas word and sentence possibilities are more structured.
Assistive features such as these can improve performance, productivity and system usability while
lightening the cognitive load of the user.

Cognitive load and Prolonged Use

The studies that have been a part of my analysis of BCI spellers had involved subjects with a limited
amount of time to attempt to complete their given task. Few of these studies explored the effect of
a sustained cognitive load on the users performance and ability to operate a BCI over a sustained
period of time or simulated operational environment.

Mathe & Evaggelos (2019) studied the effect of long-term BCI use, to simulate a representation of
the natural user experience the trials subjects attempted had no time limit. The time taken to then
complete each of the ten trials were then analysed, the results of the subject average are
presented in Figure 11.

As can be seen in Figure 11. user completion time except for trial nine were completed between
95-125 seconds. Subjects struggled with trial one as the inexperienced subjects encounter a BCI for
the first time. Their completion time improved before plateauing from trials three to six where
there was a decrease in completion time before trials eight and nine completion time spiked
before returning to a completion time similar to trials three and six.

Figure 11.: Average time taken to complete ten trials with no time limit and two minute interval between trials.

24
Trials eight and nine represent fatigue on the subjects as they attempt to complete the trials as the
average time to complete a trial was 114.8 seconds including trial intervals an average of 31.3
minutes has passed. Trial 10 then saw a reduction in completion time as subjects knew this was
the last trial this knowledge could impact performance as users’ mental states can affect user
performance with a BCI.

Discussion of Results

This report has explored the techniques and technologies as BCI systems have developed over
time, discussing the associated concepts and issues. Considering areas of current BCI research to
use non technological solutions to improve the user experience by considering the user experience
and interface design. User experience was equated to user performance for this review as I
hypothesize that there is a relationship between a user’s experience with a system and their
performance especially with BCI as the user’s mental state can impact performance. There is an
area of BCI research that concerns itself with improving BCI performance through technological
means whether it be hardware or software based. However when discussing the user experience
and interaction design there is a relationship between the user and the machine that cannot be
ignored. I analysed differing contributing factors that would improve system performance by
improving the way in which a user interacts with a BCI.

I utilised sixteen studies as part of my analysis of BCI systems, predominantly focusing on spellers
for integration into the workplace as a universal method of system input. My analysis covered user
experience, guidance systems, introduced guidance bias, combining guidance systems, combining
multiple user inputs, and implementation of user assistive systems in design.

Though there were far more inexperienced and unspecified experience levelled subjects in my
dataset than those with previous BCI experience, though they were marginally more accurate than
inexperienced users the consistency in their performances resulted in a lower standard deviation
showing that they had a consistent user experience.
Guidance systems either instruct a user in how to interact with a system or provide a critique of
their interaction performance. With this comes the potential introduction of biases and how this
impacted user performance. While strong negative bias was expected to impact performance
negatively it was interesting to see that weak negative biases were the second-best method of
feedback to improve user performance. It was noted however that as their experience grew the
feedback expected would also change to remain informative.
I believed that by providing the user additional information of subject performance their
performance would improve, I found however that this was not always correct as trained subjects
consistently performed worse when their best combined and visual feedback accuracy were
compared. Naive subjects however except Subject 6 attained an average and best accuracy rate
like if not higher than their single feedback system trials. Of note however trained users reported
that they found it less cognitively demanding to operate with the audio-visual feedback.
Combined user input systems had resulted as I expected in a higher average classification accuracy
than if subjects had utilised either user input system separately. However this analysis utilised two
studies exploring combined user input systems to improve performance as this is a growing area of
the field. Ideally more of my dataset would have incorporated combined systems for analysis
however the results currently are promising. There are concerns with implementation and upkeep
of such systems in a realistic environment however as the inputs of both systems must be

25
calibrated to discern user instructions. Despite this the average classification accuracy of combined
system inputs resulted in an improvement of performance of 8.25% over single user input systems
are significant to the users performance, accuracy and overall experience of utilising BCI.

By exploring the assistive features studies have implemented into their systems it became clear
that choices in interface design can affect user performance such as a faded out or over stimulating
interface, the use of submenus, and predictive text for both character and word selection these
findings were based upon two studies in the dataset that tested the implementation of assistive
features to improve user performance. One explored he use of interface colouring while the other
implemented predictive text in a BCI speller. I hypothesise that its poor performance was due to
the system utilising a poor learning model however these forms of assistive features have been
utilised in interfaces such as search engines, mobile phones and other websites.

These methods as demonstrated in the meta-analysis demonstrate the number of ways in which
the user can improve their user performance with a BCI instead of using technological means to
improve the user experience. These minor improvements should be further researched to divine
any potential compounding performance gains. The current knowledge lacking in human brain
mapping, requirement of tailored user systems, long-term BCI usage, performance and the legal,
social, ethical and professional concerns. Resulting in an interface system that is currently
financially, professionally and ethically unsuitable for implementation in the workplace in its
current state though this may change with future developments.

26
 References

Barbero, Á., & Grosse-Wentrup, M. (2010). Biased feedback in brain-computer interfaces.

Journal of NeuroEngineering and Rehabilitation, 7(1). https://doi.org/10.1186/1743-0003-7-34

Battarbee, K., & Koskinen, I. (2010). Co-experience: user experience as interaction.

In CoDesign, 1(1), 5–18. https://doi.org/10.1080/15710880412331289917

Bos, D. P. O., Gurkok, H., Van de Laar, B., Nijboer, F., & Nijholt, A. (2011). User Experience
Evaluation in BCI : Mind the Gap! International Journal of Bioelectromagnetism, 13(1), 48–49.
https://link-springer-com.libezproxy.open.ac.uk/book/10.1007%2F978-3-319-07638-6

Brouwer, A.-M., & Van Erp, J. B. F. (2008). A tactile P300 BCI and the optimal number of tactors:
Effects of target probability and discriminability. Proceedings of the 4th International
Brain-Computer Interface Workshop and Training Course 2008, November, 280–285.

Bublitz, C., Wolkenstein, A., Jox, R. J., & Friedrich, O. (2019). Legal liabilities of BCI-users:
Responsibility gaps at the intersection of mind and machine? International Journal of Law and
Psychiatry, 65, 1–12. https://doi.org/10.1016/j.ijlp.2018.10.002

Burwell, S., Sample, M., & Racine, E. (2017). Ethical aspects of brain computer interfaces: a
scoping review. BMC Medical Ethics, 18(1), 60. https://doi.org/10.1186/s12910-017-0220-y

Chatterjee, A., Aggarwal, V., Ramos, A., Acharya, S., & Thakor, N. V. (2007). A brain-computer
interface with vibrotactile biofeedback for haptic information. Journal of NeuroEngineering and
Rehabilitation, 4. https://doi.org/10.1186/1743-0003-4-40

Chen, X., Wang, Y., Nakanishi, M., Gao, X., Jung, T. P., & Gao, S. (2015). High-speed spelling with a
non-invasive brain-computer interface. Proceedings of the National Academy of Sciences of the
United States of America, 112(44), E6058–E6067. https://doi.org/10.1073/pnas.1508080112

Choi, J. S., Bang, J. W., Park, K. R., & Whang, M. (2013). Enhanced perception of user
intention by combining EEG and Gaze-tracking for brain-computer interfaces (BCIs).
Sensors (Switzerland), 13(3), 3454–3472. https://doi.org/10.3390/s130303454

Cinel, C., Valeriani, D., & Poli, R. (2019). Neurotechnologies for human cognitive augmentation:
Current state of the art and future prospects. In Frontiers in Human Neuroscience
(Vol. 13). Frontiers Media S.A. https://doi.org/10.3389/fnhum.2019.00013

Drew, L. (2019). The ethics of brain–computer interfaces. Nature 2019 571:7766.

Gargiulo, G. D., Mohamed, A., McEwan, A. L., Bifulco, P., Cesarelli, M., Jin, C. T., Ruffo, M.,
Tapson, J., & van Schaik, A. (2012). Investigating the role of combined acoustic-visual feedback in
one-dimensional synchronous brain computer interfaces, a preliminary study.
Medical Devices: Evidence and Research, 5(1), 81–88. https://doi.org/10.2147/MDER.S36691

Glannon, W. (2014). Ethical issues with brain-computer interfaces. Frontiers in Systems

27
Neuroscience, 8(JULY). https://doi.org/10.3389/fnsys.2014.00136

Grau, C., Ginhoux, R., Riera, A., Nguyen, T. L., Chauvat, H., Berg, M., Amengual, J. L., Pascual-
Leone, A., & Ruffini, G. (2014). Conscious brain-to-brain communication in humans using non-
invasive technologies. PLoS ONE, 9(8). https://doi.org/10.1371/journal.pone.0105225

Han, C., Xu, G., Xie, J., Chen, C., & Zhang, S. (2018). Highly Interactive Brain-Computer Interface
Based on Flicker-Free Steady-State Motion Visual Evoked Potential. Scientific Reports, 8(1).
https://doi.org/10.1038/s41598-018-24008-8

Jennifer, Sharp; Helen, Rogers; Yvonne, P. (2019). Interaction Design: Beyond Human-Computer
Interaction (5th ed.). John Wiley & Sons, Inc., Indianapolis.

Jeunet, C., Jahanpour, E., & Lotte, F. (2016). Why standard brain-computer interface (BCI) training
protocols should be changed: An experimental study. Journal of Neural Engineering, 13(3).
https://doi.org/10.1088/1741-2560/13/3/036024

Kim, G. (2015). Human–Computer Interaction. CRC Press.

https://link-springer-com.libezproxy.open.ac.uk/chapter/10.1007/978-3-319-58634-2_40

Kögel, J., Jox, R. J., & Friedrich, O. (2020). What is it like to use a BCI? - Insights from an interview
study with brain-computer interface users. BMC Medical Ethics, 21(1), 1–14.
https://doi.org/10.1186/s12910-019-0442-2
Kosmyna, N., & Lécuyer, A. (2017). Designing guiding systems for brain-computer interfaces.
Frontiers in Human Neuroscience, 11. https://doi.org/10.3389/fnhum.2017.00396

Krausova, A. (2018). Legal aspects of brain-computer interfaces. Masaryk University Journal of

Law and Technology, 8(2), 199–208. https://www.researchgate.net/publication/29
2846508_Legal_aspects_of_brain-computer_interfaces

Lebedev, M. (2014). Brain-machine interfaces: An overview. In Translational Neuroscience (Vol. 5,

Issue 1, pp. 99–110). Versita.https://doi.org/10.2478/s13380-014-0212-z

Lin, C. T., Ko, L. W., Chang, M. H., Duann, J. R., Chen, J. Y., Su, T. P., & Jung, T. P. (2010). Review of
wireless and wearable electroencephalogram systems and brain-computer interfaces - A mini-
review. In Gerontology (Vol. 56, Issue 1, pp. 112–119). https://doi.org/10.1159/000230807

Lin, Z., Zhang, C., Zeng, Y., Tong, L., & Yan, B. (2018). A novel P300 BCI speller based on the Triple
RSVP paradigm. Scientific Reports, 8(1). https://doi.org/10.1038/s41598-018-21717-y

Lotte, F., Larrue, F., & Mühl, C. (2013). Flaws in current human training protocols for spontaneous
Brain-Computer interfaces: Lessons learned from instructional design. Frontiers in Human
Neuroscience, SEP. https://doi.org/10.3389/fnhum.2013.00568

28
Martins, N. R. B., Angelica, A., Chakravarthy, K., Svidinenko, Y., Boehm, F. J., Opris, I., Lebedev, M.
A., Swan, M., Garan, S. A., Rosenfeld, J. V., Hogg, T., & Freitas, R. A. (2019). Human Brain/Cloud
Interface. Frontiers in Neuroscience, 13, 112. https://doi.org/10.3389/fnins.2019.00112
Mathe, E. & S, Evaggelos. (2019). Assessment of User Experience with a Commercial BCI Device. In
Research Gate https://www.researchgate.net/publication/335757897_Assessment
_of_User_Experience_with_a_Commercial_BCI_Device
Milekovic, T., Sarma, A. A., Bacher, D., Simeral, J. D., Saab, J., Pandarinath, C., Sorice, B. L., Blabe,
C., Oakley, E. M., Tringale, K. R., Eskandar, E., Cash, S. S., Henderson, J. M., Shenoy, K. V.,
Donoghue, J. P., & Hochberg, L. R. (2018). Stable long-term BCI-enabled communication in ALS and
locked-in syndrome using LFP signals. Journal of Neurophysiology, 120(1), 343–360.
https://doi.org/10.1152/jn.00493.2017

Mudgal, S. K., Sharma, S. K., Chaturvedi, J., & Sharma, A. (2020). Brain computer interface
advancement in neurosciences: Applications and issues. Interdisciplinary Neurosurgery:
Advanced Techniques and Case Management, 20(February), 100694.
https://doi.org/10.1016/j.inat.2020.100694

Musk, E. (2019). An Integrated Brain-Machine Interface Platform With Thousands of Channels.

Journal of Medical Internet Research, 21(10), e16194. https://doi.org/10.2196/16194

Naeem Mannan, M. M., Kamran, M. A., Kang, S., Choi, H. S., & Jeong, M. Y. (2020). A hybrid
speller design using eye tracking and SSVEP brain–computer interface. Sensors (Switzerland), 20(3).
https://doi.org/10.3390/s20030891

(i)Nagel, S., & Spüler, M. (2019). World’s fastest brain-computer interface: Combining EEG2Code
with deep learning. PLoS ONE, 14(9). https://doi.org/10.1371/journal.pone.0221909

Nicolas-Alonso, L. F., & Gomez-Gil, J. (2012). Brain computer interfaces, a review.

In Sensors (Vol. 12, Issue 2, pp. 1211–1279). https://doi.org/10.3390/s120201211

Nierhaus, T., Vidaurre, C., Sannelli, C., Mueller, K., & Villringer, A. (2019). Immediate brain
plasticity after one hour of brain–computer interface (BCI). The Journal of Physiology, JP278118.
https://doi.org/10.1113/JP278118

OfCOM. (2019). Online Nation 2019. Ofcom. May, 164. https://www.ofcom.org.uk/research-and-

data/internet-and-on-demand-research/online-nation/narrative ?

Ramos-Murguialday, A., Schürholz, M., Caggiano, V., Wildgruber, M., Caria, A., Hammer, E. M.,
Halder, S., & Birbaumer, N. (2012). Proprioceptive Feedback and Brain Computer Interface (BCI)
Based Neuroprostheses. PLoS ONE, 7(10). https://doi.org/10.1371/journal.pone.0047048
Renton, A. I., Mattingley, J. B., & Painter, D. R. (2019). Optimising non-invasive brain-computer
interface systems for free communication between naïve human participants. Scientific Reports,
9(1), 1–18. https://doi.org/10.1038/s41598-019-55166-y

29
Rocchi, P. (2016). Introducing human-computer interaction: A didactic experience. IEEE Global
Engineering Education Conference, EDUCON, 10-13-Apri(April), 992–996.
https://doi.org/10.1109/EDUCON.2016.7474673

Rosca, S., Leba, M., Ionica, A., & Gamulescu, O. (2018). Quadcopter control using a BCI. IOP
Conference Series: Materials Science and Engineering, 294(1).
https://doi.org/10.1088/1757-899X/294/1/012048

Ryan, D. B., Townsend, G., Gates, N. A., Colwell, K., & Sellers, E. W. (2017). Evaluating
brain-computer interface performance using color in the P300 checkerboard speller. Clinical
Neurophysiology, 128(10), 2050–2057. https://doi.org/10.1016/j.clinph.2017.07.397

Schreuder, M., Riccio, A., Risetti, M., Dähne, S., Ramsay, A., Williamson, J., Mattia, D., &
Tangermann, M. (2013). User-centered design in brain-computer interfaces-A case study. Artificial
Intelligence in Medicine, 59(2), 71–80. https://doi.org/10.1016/j.artmed.2013.07.005

Speier, W., Arnold, C., Chandravadia, N., Roberts, D., Pendekanti, S., & Pouratian, N. (2018).
Improving P300 spelling rate using language models and predictive spelling. Brain-Computer
Interfaces, 5(1), 13–22. https://doi.org/10.1080/2326263X.2017.1410418

Volosyak, I., Gembler, F., & Stawicki, P. (2017). Age-related differences in SSVEP-based BCI
performance. Neurocomputing, 250, 57–64. https://doi.org/10.1016/j.neucom.2016.08.121

Xing, X., Wang, Y., Pei, W., Guo, X., Liu, Z., Wang, F., Ming, G., Zhao, H., Gui, Q., & Chen, H.
(2018). A High-Speed SSVEP-Based BCI Using Dry EEG Electrodes. Scientific Reports, 8(1).
https://doi.org/10.1038/s41598-018-32283-8

Zander, T. O., & Kothe, C. (2011). Towards passive brain-computer interfaces: Applying brain-
computer interface technology to human-machine systems in general.
In Journal of Neural Engineering, 8(2). https://doi.org/10.1088/1741-2560/8/2/025005

30
 Bibliography

Adams, S. S., Arel, I., Bach, J., Coop, R., Furlan, R., Goertzel, B., Hall, J. S., Samsonovich, A.,
Scheutz, M., Schlesinger, M., Shapiro, S. C., & Sowa, J. F. (2012). Mapping the landscape of
human-level artificial general intelligence. AI Magazine, 33(1), 25–41. https://doi.org/
10.1609/aimag.v33i1.2322

Barbero, Á., & Grosse-Wentrup, M. (2010). Biased feedback in brain-computer interfaces. Journal
of NeuroEngineering and Rehabilitation, 7(1). https://doi.org/10.1186/1743-0003-7-34
Battarbee, K., & Koskinen, I. (2010). Co-experience: user experience as interaction. CoDesign, 1(1),
5–18. https://doi.org/10.1080/15710880412331289917
Bos, D. P. O., Gurkok, H., Van de Laar, B., Nijboer, F., & Nijholt, A. (2011). User Experience
Evaluation in BCI : Mind the Gap! International Journal of Bioelectromagnetism, 13(1),
48–49.
Brouwer, A.-M., & Van Erp, J. B. F. (2008). A tactile P300 BCI and the optimal number of tactors:
Effects of target probability and discriminability. Proceedings of the 4th International Brain-
Computer Interface Workshop and Training Course 2008, November, 280–285.
Bublitz, C., Wolkenstein, A., Jox, R. J., & Friedrich, O. (2019). Legal liabilities of BCI-users:
Responsibility gaps at the intersection of mind and machine? International Journal of Law
and Psychiatry, 65, 1–12. https://doi.org/10.1016/j.ijlp.2018.10.002
Burwell, S., Sample, M., & Racine, E. (2017). Ethical aspects of brain computer interfaces: a
scoping review. BMC Medical Ethics, 18(1), 60. https://doi.org/10.1186/s12910-017-0220-y
Chatterjee, A., Aggarwal, V., Ramos, A., Acharya, S., & Thakor, N. V. (2007). A brain-computer
interface with vibrotactile biofeedback for haptic information. Journal of NeuroEngineering
and Rehabilitation, 4. https://doi.org/10.1186/1743-0003-4-40
Chen, X., Wang, Y., Nakanishi, M., Gao, X., Jung, T. P., & Gao, S. (2015). High-speed spelling with a
noninvasive brain-computer interface. Proceedings of the National Academy of Sciences of
the United States of America, 112(44), E6058–E6067.
https://doi.org/10.1073/pnas.1508080112
Choi, J. S., Bang, J. W., Park, K. R., & Whang, M. (2013). Enhanced perception of user intention by
combining EEG and Gaze-tracking for brain-computer interfaces (BCIs). Sensors (Switzerland),
13(3), 3454–3472. https://doi.org/10.3390/s130303454
Cinel, C., Valeriani, D., & Poli, R. (2019). Neurotechnologies for human cognitive augmentation:
Current state of the art and future prospects. In Frontiers in Human Neuroscience (Vol. 13).
Frontiers Media S.A. https://doi.org/10.3389/fnhum.2019.00013
Drew, L. (2019). The ethics of brain–computer interfaces. Nature 2019 571:7766.

31
Ferrucci, D., Brown, E., Chu-Carroll, J., Fan, J., Gondek, D., Kalyanpur, A. A., Lally, A., Murdock, J.
W., Nyberg, E., Prager, J., Schlaefer, N., & Welty, C. (2011). Building watson: An overview of
the deepQA project. AI Magazine, 31(3), 59–79. https://doi.org/10.1609/aimag.v31i3.2303
Folcher, M., Oesterle, S., Zwicky, K., Thekkottil, T., Heymoz, J., Hohmann, M., Christen, M., Daoud
El-Baba, M., Buchmann, P., & Fussenegger, M. (2014). Mind-controlled transgene expression
by a wireless-powered optogenetic designer cell implant. Nature Communications, 5.
https://doi.org/10.1038/ncomms6392
Forno, R. F. (2019). Risk awareness and the user experience. 3. In SIGDOC'19: The 37th ACM
International Conference on the Design of Communication Proceedings, October 04–06, 2019,
https://doi.org/10.1145/3328020.3353918
Fuhrmann, G., Segev, I., Markram, H., & Tsodyks, M. (2002). Coding of temporal information by
activity-dependent synapses. Journal of Neurophysiology, 87(1), 140–148.
https://doi.org/10.1152/jn.00258.2001
Gargiulo, G. D., Mohamed, A., McEwan, A. L., Bifulco, P., Cesarelli, M., Jin, C. T., Ruffo, M.,
Tapson, J., & van Schaik, A. (2012). Investigating the role of combined acoustic-visual
feedback in one-dimensional synchronous brain computer interfaces, a preliminary study.
Medical Devices: Evidence and Research, 5(1), 81–88. https://doi.org/10.2147/MDER.S36691
Glannon, W. (2014). Ethical issues with brain-computer interfaces. Frontiers in Systems
Neuroscience, 8(JULY). https://doi.org/10.3389/fnsys.2014.00136
Grau, C., Ginhoux, R., Riera, A., Nguyen, T. L., Chauvat, H., Berg, M., Amengual, J. L., Pascual-
Leone, A., & Ruffini, G. (2014). Conscious brain-to-brain communication in humans using
non-invasive technologies. PLoS ONE, 9(8). https://doi.org/10.1371/journal.pone.0105225
Guduru, R., Liang, P., Hong, J., Rodzinski, A., Hadjikhani, A., Horstmyer, J., Levister, E., & Khizroev,
S. (2015). Magnetoelectric “spin” on stimulating the brain. Nanomedicine, 10(13), 2051–2061.
https://doi.org/10.2217/nnm.15.52
Guger, C., & Allison, B. Z. (2013). Brain – Computer Interface Research A State-of-the-Art
Summary. Springer Science and Business Media LLC.
Han, C., Xu, G., Xie, J., Chen, C., & Zhang, S. (2018). Highly Interactive Brain-Computer Interface
Based on Flicker-Free Steady-State Motion Visual Evoked Potential. Scientific Reports, 8(1).
https://doi.org/10.1038/s41598-018-24008-8
Hoffmann, Ulrich., Vesin, Jean-Marc., Ebrahimi, T. (2007). Recent advances in brain-computer
interfaces. 2007 IEEE 9Th International Workshop on Multimedia Signal Processing, MMSP
2007 - Proceedings, May 2014, 17. https://doi.org/10.1109/MMSP.2007.4412807
Jackman, S. L., & Regehr, W. G. (2017). The Mechanisms and Functions of Synaptic Facilitation. In
Neuron (Vol. 94, Issue 3, pp. 447–464). Cell Press.
https://doi.org/10.1016/j.neuron.2017.02.047

32
Jennifer, Sharp; Helen, Rogers; Yvonne, P. (2019). Interaction Design: Beyond Human-Computer
Interaction (5th ed.). John Wiley & Sons, Inc.

Jeunet, C., Jahanpour, E., & Lotte, F. (2016). Why standard brain-computer interface (BCI) training
protocols should be changed: An experimental study. Journal of Neural Engineering, 13(3).
https://doi.org/10.1088/1741-2560/13/3/036024
Jorfi, M., Skousen, J. L., Weder, C., & Capadona, J. R. (2015). Progress towards biocompatible
intracortical microelectrodes for neural interfacing applications. In Journal of Neural
Engineering (Vol. 12, Issue 1). Institute of Physics Publishing. https://doi.org/10.1088/1741-
2560/12/1/011001
Katona, J., & Kovari, A. (2016). A Brain-Computer Interface Project Applied in Computer
Engineering. IEEE Transactions on Education, 59(4), 319–326.
https://doi.org/10.1109/TE.2016.2558163
Kim, G. J. (2015). Human–Computer Interaction: Fundamentals and Practice. CRC Press. Boca
Raton.
Kober, S. E., Witte, M., Ninaus, M., Neuper, C., & Wood, G. (2013). Learning to modulate one’s
own brain activity: The effect of spontaneous mental strategies. Frontiers in Human
Neuroscience, OCT. https://doi.org/10.3389/fnhum.2013.00695
Koch, C., Poggio, T., & Torre, V. (1983). Nonlinear interactions in a dendritic tree: Localization,
timing, and role in information processing. Proceedings of the National Academy of Sciences
of the United States of America, 80(9 I), 2799–2802. https://doi.org/10.1073/pnas.80.9.2799
Kögel, J., Jox, R. J., & Friedrich, O. (2020). What is it like to use a BCI? - Insights from an interview
study with brain-computer interface users. BMC Medical Ethics, 21(1), 1–14.
https://doi.org/10.1186/s12910-019-0442-2
Kosmyna, N., & Lécuyer, A. (2017). Designing guiding systems for brain-computer interfaces.
Frontiers in Human Neuroscience, 11. https://doi.org/10.3389/fnhum.2017.00396
Kosmyna, N., Tarpin-Bernard, F., Bonnefond, N., & Rivet, B. (2016). Feasibility of BCI control in a
realistic smart home environment. Frontiers in Human Neuroscience, 10, 10.
https://doi.org/10.3389/fnhum.2016.00416
Kothe, C. (2012). Introduction To Modern Brain-Computer Interface Design – SCCN. Swartz Center
for Computational Neuroscience, UCSD, Teaser, Overview, Lecture 1, Lecture 1.1 What is a.
http://sccn.ucsd.edu/wiki/Introduction_To_Modern_Brain-Computer_Interface_Design
Kozai, T. D. Y., Jaquins-Gerstl, A. S., Vazquez, A. L., Michael, A. C., & Cui, X. T. (2015). Brain tissue
responses to neural implants impact signal sensitivity and intervention strategies. ACS
Chemical Neuroscience, 6(1), 48–67. https://doi.org/10.1021/cn500256e

33
Krausova, A. (2018). Legal aspects of brain-computer interfaces. Masaryk University Journal of
Law and Technology, 8(2), 199–208. https://www.researchgate.net/publication/
292846508_Legal_aspects_of_brain-computer_interfaces
Kwak, N. S., Müller, K. R., & Lee, S. W. (2017). A convolutional neural network for steady state
visual evoked potential classification under ambulatory environment. PLoS ONE, 12(2),
e0172578. https://doi.org/10.1371/journal.pone.0172578
Landreth, A., & Silva, A. J. (2013). The need for research maps to navigate published work and
inform experiment planning. In Neuron (Vol. 79, Issue 3, pp. 411–415).
https://doi.org/10.1016/j.neuron.2013.07.024
Lazar, J., Feng, J. H., & Hochheiser, H. (2017). Research Methods in Human-Computer Interaction.
In Research Methods in Human-Computer Interaction (2nd ed.). Elsevier.
https://doi.org/10.1016/b978-0-444-70536-5.50047-6
Lazarou, I., Nikolopoulos, S., Petrantonakis, P. C., Kompatsiaris, I., & Tsolaki, M. (2018).
EEG-based brain–computer interfaces for communication and rehabilitation of people with
motor impairment: A novel approach of the 21st century. Frontiers in Human Neuroscience,
12(January), 1–18. https://doi.org/10.3389/fnhum.2018.00014
Lebedev, M. (2014). Brain-machine interfaces: An overview. In Translational Neuroscience (Vol. 5,
Issue 1, pp. 99–110). Versita. https://doi.org/10.2478/s13380-014-0212-z
Lebedev, M. A., & Nicolelis, M. A. L. (2006). Brain-machine interfaces: past, present and future. In
Trends in Neurosciences (Vol. 29, Issue 9, pp. 536–546).
https://doi.org/10.1016/j.tins.2006.07.004
Lebedev, M. A., Opris, I., & Casanova, M. F. (2018). Editorial: Augmentation of brain function:
Facts, fiction and controversy. In Frontiers in Systems Neuroscience (Vol. 12). Frontiers Media
S.A. https://doi.org/10.3389/fnsys.2018.00045
Lewis, P. M., Ayton, L. N., Guymer, R. H., Lowery, A. J., Blamey, P. J., Allen, P. J., Luu, C. D., &
Rosenfeld, J. V. (2016). Advances in implantable bionic devices for blindness: a review. In ANZ
Journal of Surgery (Vol. 86, Issue 9, pp. 654–659). Blackwell Publishing.
https://doi.org/10.1111/ans.13616
Li, G., & Zhang, D. (2016). Brain-Computer Interface Controlled Cyborg: Establishing a Functional
Information Transfer Pathway from Human Brain to Cockroach Brain. PloS One, 11(3),
e0150667. https://doi.org/10.1371/journal.pone.0150667
Lin, C. T., Ko, L. W., Chang, M. H., Duann, J. R., Chen, J. Y., Su, T. P., & Jung, T. P. (2010). Review of
wireless and wearable electroencephalogram systems and brain-computer interfaces – A
mini-review. In Gerontology (Vol. 56, Issue 1, pp. 112–119).
https://doi.org/10.1159/000230807
Lin, Z., Zhang, C., Zeng, Y., Tong, L., & Yan, B. (2018). A novel P300 BCI speller based on the Triple
RSVP paradigm. Scientific Reports, 8(1). https://doi.org/10.1038/s41598-018-21717-y

34
Lotte, F., Larrue, F., & Mühl, C. (2013). Flaws in current human training protocols for spontaneous
Brain-Computer interfaces: Lessons learned from instructional design. Frontiers in Human
Neuroscience, SEP. https://doi.org/10.3389/fnhum.2013.00568
Martins, N. R. B., Angelica, A., Chakravarthy, K., Svidinenko, Y., Boehm, F. J., Opris, I., Lebedev, M.
A., Swan, M., Garan, S. A., Rosenfeld, J. V., Hogg, T., & Freitas, R. A. (2019). Human
Brain/Cloud Interface. Frontiers in Neuroscience, 13, 112.
https://doi.org/10.3389/fnins.2019.00112
Mathe, E. (2019). Assessment of User Experience with a Commercial BCI Device. Researchgate,
September.https://www.researchgate.net/publication/335757897_Assessment_of_User_Exp
erience_with a_Commercial_BCI_Device

Milekovic, T., Sarma, A. A., Bacher, D., Simeral, J. D., Saab, J., Pandarinath, C., Sorice, B. L., Blabe,
C., Oakley, E. M., Tringale, K. R., Eskandar, E., Cash, S. S., Henderson, J. M., Shenoy, K. V.,
Donoghue, J. P., & Hochberg, L. R. (2018). Stable long-term BCI-enabled communication in
ALS and locked-in syndrome using LFP signals. Journal of Neurophysiology, 120(1), 343–360.
https://doi.org/10.1152/jn.00493.2017
Millán, J. del R., & Mouriño, J. (2003). Asynchronous BCI and local neural classifiers: An overview
of the adaptive brain interface project. IEEE Transactions on Neural Systems and
Rehabilitation Engineering, 11(2), 159–161. https://doi.org/10.1109/TNSRE.2003.814435
Mudgal, S. K., Sharma, S. K., Chaturvedi, J., & Sharma, A. (2020). Brain computer interface
advancement in neurosciences: Applications and issues. Interdisciplinary Neurosurgery:
Advanced Techniques and Case Management, 20(February), 100694.
https://doi.org/10.1016/j.inat.2020.100694
Musk, E. (2019). An Integrated Brain-Machine Interface Platform With Thousands of Channels.
Journal of Medical Internet Research, 21(10), e16194. https://doi.org/10.2196/16194
Naeem Mannan, M. M., Kamran, M. A., Kang, S., Choi, H. S., & Jeong, M. Y. (2020). A hybrid
speller design using eye tracking and SSVEP brain–computer interface. Sensors (Switzerland),
20(3). https://doi.org/10.3390/s20030891
(i)Nagel, S., & Spüler, M. (2019). World’s fastest brain-computer interface: Combining EEG2Code
with deep learning. PLoS ONE, 14(9). https://doi.org/10.1371/journal.pone.0221909
(ii)Nagel, S., & Spüler, M. (2019). Asynchronous non-invasive high-speed BCI speller with robust
non-control state detection. Scientific Reports, 9(1). https://doi.org/10.1038/s41598-019-
44645-x
Nakanishi, M., Wang, Y., Chen, X., Wang, Y. Te, Gao, X., & Jung, T. P. (2018). Enhancing detection
of SSVEPs for a high-speed brain speller using task-related component analysis. IEEE
Transactions on Biomedical Engineering, 65(1), 104–112.
https://doi.org/10.1109/TBME.2017.2694818

35
Nicolas-Alonso, L. F., & Gomez-Gil, J. (2012). Brain computer interfaces, a review. In Sensors (Vol.
12, Issue 2, pp. 1211–1279). https://doi.org/10.3390/s120201211
Nierhaus, T., Vidaurre, C., Sannelli, C., Mueller, K., & Villringer, A. (2019). Immediate brain
plasticity after one hour of brain–computer interface (BCI). The Journal of Physiology,
JP278118. https://doi.org/10.1113/JP278118
Norton, J. J. S., Lee, D. S., Lee, J. W., Lee, W., Kwon, O., Won, P., Jung, S. Y., Cheng, H., Jeong, J. W.,
Akce, A., Umunna, S., Na, I., Kwon, Y. H., Wang, X. Q., Liu, Z. J., Paik, U., Huang, Y., Bretl, T.,
Yeo, W. H., … Bao, Z. (2015). Soft, curved electrode systems capable of integration on the
auricle as a persistent brain-computer interface. Proceedings of the National Academy of
Sciences of the United States of America, 112(13), 3920–3925.
https://doi.org/10.1073/pnas.1424875112
OfCOM. (2019). Online Nation 2019. Ofcom, May, 164. https://www.ofcom.org.uk/research-and-
data/internet-and-on-demand-research/online-nation/narrative?
Palaniappan, R. (2005). Brain computer interface design using band powers extracted during
mental tasks. 2nd International IEEE EMBS Conference on Neural Engineering, 2005, 321–324.
https://doi.org/10.1109/CNE.2005.1419622
Ramos-Murguialday, A., Schürholz, M., Caggiano, V., Wildgruber, M., Caria, A., Hammer, E. M.,
Halder, S., & Birbaumer, N. (2012). Proprioceptive Feedback and Brain Computer Interface
(BCI) Based Neuroprostheses. PLoS ONE, 7(10).
https://doi.org/10.1371/journal.pone.0047048
Renton, A. I., Mattingley, J. B., & Painter, D. R. (2019). Optimising non-invasive brain-computer
interface systems for free communication between naïve human participants. Scientific
Reports, 9(1), 1–18. https://doi.org/10.1038/s41598-019-55166-y
Rezeika, A., Benda, M., Stawicki, P., Gembler, F., Saboor, A., & Volosyak, I. (2018).
Brain–computer interface spellers: A review. Brain Sciences, 8(4).
https://doi.org/10.3390/brainsci8040057
Rocchi, P. (2016). Introducing human-computer interaction: A didactic experience. IEEE Global
Engineering Education Conference, EDUCON, 10-13-Apri(April), 992–996.
https://doi.org/10.1109/EDUCON.2016.7474673
Rosca, S., Leba, M., Ionica, A., & Gamulescu, O. (2018). Quadcopter control using a BCI. IOP
Conference Series: Materials Science and Engineering, 294(1). https://doi.org/10.1088/1757-
899X/294/1/012048
Ruf, S. P., Fallgatter, A. J., & Plewnia, C. (2017). Augmentation of working memory training by
transcranial direct current stimulation (tDCS). Scientific Reports, 7(1).
https://doi.org/10.1038/s41598-017-01055-1

36
Ryan, D. B., Townsend, G., Gates, N. A., Colwell, K., & Sellers, E. W. (2017). Evaluating brain-
computer interface performance using color in the P300 checkerboard speller. Clinical
Neurophysiology, 128(10), 2050–2057. https://doi.org/10.1016/j.clinph.2017.07.397
Schreuder, M., Riccio, A., Risetti, M., Dähne, S., Ramsay, A., Williamson, J., Mattia, D., &
Tangermann, M. (2013). User-centered design in brain-computer interfaces-A case study.
Artificial Intelligence in Medicine, 59(2), 71–80. https://doi.org/10.1016/j.artmed.2013.07.005
Speier, W., Arnold, C., Chandravadia, N., Roberts, D., Pendekanti, S., & Pouratian, N. (2018).
Improving P300 spelling rate using language models and predictive spelling. Brain-Computer
Interfaces, 5(1), 13–22. https://doi.org/10.1080/2326263X.2017.1410418
Volosyak, I., Gembler, F., & Stawicki, P. (2017). Age-related differences in SSVEP-based BCI
performance. Neurocomputing, 250, 57–64. https://doi.org/10.1016/j.neucom.2016.08.121
Xing, X., Wang, Y., Pei, W., Guo, X., Liu, Z., Wang, F., Ming, G., Zhao, H., Gui, Q., & Chen, H. (2018).
A High-Speed SSVEP-Based BCI Using Dry EEG Electrodes. Scientific Reports, 8(1).
https://doi.org/10.1038/s41598-018-32283-8
Zander, T. O., & Kothe, C. (2011). Towards passive brain-computer interfaces: Applying brain-
computer interface technology to human-machine systems in general. Journal of Neural
Engineering, 8(2). https://doi.org/10.1088/1741-2560/8/2/025005
Zhu, D., Bieger, J., Garcia Molina, G., & Aarts, R. M. (2010). A survey of stimulation methods used
in SSVEP-based BCIs. In Computational Intelligence and Neuroscience (Vol. 2010, p. 702357).
https://doi.org/10.1155/2010/702357

37
 Project Management Review

Review of Utilised Project Development Lifecycle

During project development I had chosen to utilise the iterative waterfall lifecycle. This approach
involves the separation of development into four stages, analysis, design, implementation,
evaluation and iteration within themselves and amongst these stages.

My research phase represented by my literature review was a significant portion of my report

iterated upon throughout project development. This iteration occurred as other elements of the
project report were further developed to further research a topic to close a gap in knowledge
which required the literature review to be further developed. Iteration also occurred as sections
were completed for quality of clarity, spelling and content.

Utilising the iterative waterfall project lifecycle during development I completed the necessary
research, analysis and synthesis for my project report, currently in the final days of the project.
I am completing the title, contents and numbering the pages with a final review of the completed
project report into BCI and the user experience before submission. I had found the requirement of
iteration throughout project development useful as I could review my progress and set future tasks
adjusting to account for unexpected disruptions and breaks in development to meet the project
deadline.

I have included the projected and actual weekly task breakdown as my project was developed,
throughout development this was revised upon as prospective completion times shifted and
external events occurred. These will be discussed further in the review of my personal
performance and future project development. However, as these events occurred, and tasks were
attempted being able to review and reiterate upon my progress I was able to retain momentum
and progress development.

Prospective Timeline Actual Timeline

Wee Weekly Key dates and Time Weekly Key dates and Time
k Objectives events Allotte Objectives events Allotte
d d
(hours) (hours)
1   Call scheduled 15 Collate plans on Call scheduled 9
with supervisor OU forum with supervisor
modules into
prospective ideas
to discuss with
my project Tutor
2 Finalise   20 Finalise details,   20
details, begin to look at
begin to look TMA01
at TMA01 objectives and
objectives research proposed

38
 and research topic
proposed
topic
3 Start   10 Start working on   10
working on TMA01, explore
TMA01, information
explore sources and begin
information to read journals
sources and
begin to read
journals
4 Complete 3rd March 7 Complete 3rd March 7
TMA01, TMA01Deadlin TMA01, schedule TMA01Deadlin
schedule call e call with e
with supervisor with in
supervisor two weeks to
with in two discuss proposal
weeks to
discuss
proposal
5 Complete 12th March 25 Complete 12th March 25
TMA03 for TMA03 TMA03 for TMA03
TM356 deadline for TM356 deadline for
TM356 due TM356 due
6 Continue   15 After finishing   0
research and both TMA’s the
write preceding two
reviews of weeks I took a
each source week off from
for the OU studies to
literature refocus
review
7 Continue   15 Completed two   12
research and weeks study of
write TM356
reviews of
each source
for the
literature
review
8 Continue   15 Continue research Covid-19 has 17.5
research and and the literature disrupted work
write review as I am placed
reviews of on furlough
each source
for the
literature
review
9 Continue   15 Continue research Discuss project 19
research and and the literature with Tutor and

39
 write review explain the need
reviews of for a TMA
each source extension until
for the the 16th April
literature for TMA02 of
review TM470
10 Continue   15 Continue research   20
research and and write reviews
write of each source for
reviews of the literature
each source review and work
for the on TM470
literature TMA02
review
11 Continue   15 Complete   15
research and TMA02 of
write TM470
reviews of
each source
for the
literature
review
examine
TMA02 and
start TMA04
12 Work on   15 Work on TM356   13
TMA02 and TMA04 and
TMA04, review project
begin structure
planning
Chapter
contents of
research
project
13 Complete 7th May 15 Complete and 7th May 15
TMA04, TMA04 TM356 submit TM356 TMA04 TM356
continue due TMA04 completed
working on
TMA02 and
begin
writing first
section of
the report

14 Complete 14th April 15 Organise papers 14th April 15

TMA02 and TMA02 due for review and TMA02 due
submit, complete TM356
finish TMA02
writing first
section,

40
 reference,
check and
update
literature
review if
necessary

15 Plan out the   10 Review literature   10

next two review structure
sections and examine data
content for suitable BCI
systems
16 TM356   20 From data found   17
exam in research my
revision initial plan was
not suitable. The
plan was
refocused
17 TM356   20 Write literature   18
exam review section
revision one
18 TM356 12th June 20 Write literature TM356 exam 14.5
exam TM356 exam review section cancelled due to
revision and two Covid-19
exam
19 Begin   15 Go through Received 15
writing research to collate literature review
following the data from feedback from
sections that papers so that Tutor on the 21st
have been these may be of June.
planned. manipulated in a
suitable way for
the discussion
section.
20 Explore   15 Incorporate   13
requirements changes to lit
of TMA03, review
continue recommended by
work on my project
sections supervisor.

21 Begin   18 Begin TMA03   20

TMA03 and and continue
continue writing sections
writing
sections

41
22 Finalise and 7th July TMA03 12 Finish literature 7th July TMA03 16
submit due review alterations due
TMA03, for the TMA
finish
sections
23 Plan   10 Finalise and   5
remaining submit TMA03.
sections,
find
supplementa
l material if
required and
begin
writing these
sections
24 Continue to   15 Utilised   8
write these manipulated data
sections to structure the
discussion section
25 Continue to   15 Complete   18
write these synthesis section
sections and
review the
content and
referencing
of each
26 Complete   10 Complete legal,   15
section on moral, ethical,
ethics, societal section
medical and
legal
concerns
27 Complete   10 Begin and   10
section on Complete
ethics, conclusion also
medical and integrate
legal feedback from
concerns supervisor as to
the introduction
and conclusion

28 Review   8 Review project   10

project and and ensure
construct a formatting and
suitable referencing etc is
introduction in order
and
conclusion
that
represents

42
 the reports
content
29 Review   8 Finish project and   23
project and EMA part 1
construct a
suitable
introduction
and
conclusion
that
represents
the reports
content
30 Review   8 Review project,   11
project, update
update referencing,
referencing, create cover page,
create cover contents,
page, supplemental
contents and material and page
numbering numbering
31 Review   8 Complete part 2   17
project, of EMA: Project
update Management and
referencing, performance
create cover review and
page, finalise project
contents and report formatting
numbering
32 Review and 14th September 2 Submit EMA on 14th September 2
finalise 12pm project 13th September 12pm project
before deadline. as I have work on deadline.
submission the 14th
 Keep a project journal as items are completed documenting Keep a project journal as items are completed
progress or lack thereof, update the project plan as required documenting progress or lack thereof, update the project
and document time actual in comparison to the plan. plan as required and document time actual in comparison
10 to 15 minutes each day project work is carried out to the plan.
10 to 15 minutes each day project work is carried out

43
 Figure 1: Presents a visual representation of the prospective and actual project timeline associated with the chart
to visualise project time usage.

Associated Project Activities, Skills, Risks & Resources

44
This section includes a chart below demonstrating the associated tasks, skills, risks and resources
necessary for the completion of this project. There is some overlap due to the nature of my
research-based project however they have been included for completeness of each activity.

Activity Skills Risks Resources

Determining the feasibility of Planning Project not feasible PubMed
project (contacted tutors for
Organisation Not related to a WebofScience
planning and support/
previous module
Suggested back up
Communication Poor quality IEEE Explorer
projects/presented literature of
literature
a reputable source)
Critical reading Out of date Mendeley
literature
Information Non open access Open
technology literature University
library
  Lack of studies from Previous
your investigative module
aspect resources
    Tutor
    Open
University
project forum
Project proposal (define & Efficient reading Little relation IEEE Explorer
elucidate the area of between BCI and
interest/Define project type; module topics
Research, Development or Critical reading Low quantity of Tutor
investigative/determining high-quality
criteria for success) literature
Communication Project rejection Mendeley
Computationally Literature Module
adept inaccessible Documentation
Source Different criteria of PubMed
evaluation success of a BCI
Referencing Determination of a MS Office
suitable time frame
Application of   Tutor TMA
previous feedback
learning
Adaptability    

45
Obtaining project approval Communication Rejection Tutor
Effective No relation to L3 Previous
presentation of module module
project ideas resources

Evaluation of    
risks defined
above
Background reading into the Organisation Inefficient tracking IEEE Explorer
Development of BCI of literature could
lead to information
overload
Critical analysis Timescale may not Mendeley
permit enough
background
research to critically
evaluate current BCI
projects
Efficient reading Project highly MS Office
interdisciplinary
which would
contribute to the
above risk
Evaluation of Word count Open
literature restrictions propose University
a significant limiting library
factor (background
knowledge:
synthesis ratio)
Referencing Retaining relation to PubMed
computing and
previous modules
Computational Hard to integrate  
literacy and assess multiple
metrics and
differing criteria of
BCI success
Know when to    
stop reading
background info
Development and improvement Critical analysis Timescale to IEEE Explorer
of search statement of literature adequate search
strategy
Referencing Information Mendeley
overload
Keeping track of Keeping it related Open
results University

46
 library
Identification of   PubMed
project related
key words
Source    
evaluation
Development of Glossary Organisation Not understanding Mendeley
the
multidisciplinary
content outside of
computing
Developed Unable to discuss Pubmed
understanding the topics,
of the content strategies and
reasoning
conventions of BCI
research to inform a
computing-based
reader
Tailor glossary Too much time Open
to inform a focusing on the University
computing glossary, detracting library
focused reader from the final
report timeline
Investigative   Tutor
skills
Developed aid   SPSSTutorials
while carrying
out critical
readings
Examination of the types of BCI Analytic ability Not understanding Mendeley
the concepts and
fundamentals of BCI
understanding Spend too much Pubmed
the concepts time examining the
and different types of
fundamentals of BCI
BCI
Develop an Poorly presenting Open
understanding the differences University
of the traits between BCI library
associated with technologies
varying types of
BCI technology
and technique

47
 Organisation The project Excel
disproportionately
describes BCI with
little synthesis

Referencing   Word
Evaluate the    
presentation of
results
Assessment of User Experience Progress Poorly applying the Mendeley
and Interaction Design independent L3 module material
reading and to the project topic
understanding
of UX & ID
research
beyond L3
module
materials
Critical analysis Assessment and Pubmed
of literature application of these
topics takes up a
significant portion
of the timeline
Organisation   Open
University
library
Evaluation of   OU L3 ux & id
literature module site
Application of   IEEE Explorer
these concepts
to BCI
Referencing   Word
Legal, Professional and Ethical Critical analysis Too much time Mendeley
considerations of BCI of literature focused on
developing an
understanding of
LPE BCI concerns
Application of Poorly considering Pubmed
LPE concerns in LPE concerns in
the research, relation to BCI
development technologies
and
implementation
of BCI
technologies

48
 Organisation   Open
University
library
Referencing   OU L3 UX & ID
module site
    IEEE Explorer

Choosing a type of BCI for meta- Consider the Multidisciplinary Mendeley

analysis literature nature of BCI
examined as to research increases
how BCI type the risk of
data is discontinuity
presented between
experiment
presentation of
results
Select a BCI Discontinuity Pubmed
type for meta- between study
analysis with a results could result
satisfactory in difficulty
number of synthesising the
studies studies selected
Consider the Not able to find a Open
timeliness of suitable BCI type for University
the BCI type analysis library
chosen for
meta-analysis
Organisation Unable to find a Excel
significant number
of similar studies for
comparison and
synthesis
Being decisive in Excluding significant Word
selecting an studies in the field
appropriate of BCI to support
type of BCI for the project due to
meta-analysis narrowly focusing
research
Evaluation of Meta-analysis Tutor
studies utilises too many
dissimilar studies
  Poorly synthesising  
studies selected for

49
 meta-analysis
resulting in a report
with little critical
thinking skills
Conducting meta-analysis Keep track of Poorly presented Mendeley
results results
Organisation Ineffectively Excel
communicate
metrics and findings
Evaluation of Select inappropriate Word
studies metrics that do not
support the aims of
the project
Referencing Poor amount of  
project dedicated to
synthesis of results
Create metrics    
that will be
measured
against those
studies selected
to be part of the
meta-analysis
Assess the    
results of the
studies against
those metrics to
arrive at a
unique result
Present the    
results in text
and utilise
graphs and
charts to
support findings
Project management Keep project Project not Word
journal completed in time
Select a project Rushed Excel
development development with
lifecycle little time to edit
the final report
Consider the Unprepared for Day planner
benefits of project milestones
differing
lifecycle
Manage my Loss of project Email
time and progress via system

50
 consider the failure, overwriting
time of others save data etc.
such as my tutor
as they are a
valuable
resource
Review the   OU materials
project on project
timescale and management
adjust as
necessary to my
progress to
meet the
deadline

Carry out a risk   TMA feedback

assessment and
mitigation
strategies for
the project
Communicate   Google Drive
difficulties and
unexpected
issues with the
project tutor
Retaining   OneDrive
suitable filing
systems and
back-ups of
project progress
from system
failures and
version control
Organisation   Mendeley
ecosystem
    SVN project
management
tool (version
control)
Project Journal and tracking Make notes of Not sure of stage in Project journal
work the project
completed,
difficulties, tasks
to be
completed,
areas of interest

51
 and thoughts on
progress

  Inability to Notes and

complete section of reminder
EMA requiring a function on
recounting of phone and
project laptop
management
experience
  Forgotten thoughts, OU materials
tasks and on project
investigation of management
areas of interest
    TMA feedback

Review of project management Be introspective Presenting my Word

and follow review incoherently
professional
guidelines as to
accurately
describing my
experience
completing this
project
Present an Not providing Excel
accurate enough discussion
representation
of the thought
involved in
planning this
project
Evaluate the Unable to be critical OU materials
planning and about my on project
reality that performance management
occurred planning and
through project executing a project
development
Analyse my Poor project Project journal
personal progress tracking
strengths and
weaknesses and
any external
factors that
impacted
project

52
 development

Assess if given Poor organisational Day planner

the opportunity skills
what would be
done if the
project were to
be reattempted
From my project   Emails
discuss future
developments
and
improvements
that could be
made with
further research
Methodical   Phone call
notes

Critical of   TMA feedback

performance
and event
through
development
Organisation    
Submit completed project Organisation Not completed on The resources
report time mentioned
above
Evaluation of Little time to The OU
literature present the report website
in the desired
fashion
Critical reading Failure to meet the  
deadline
Synthesis Presenting a poor  
review of the work
done during project
development
Follow The EMA system  
instruction not fully functional
Complete Missing the  
project report deadline
requirements

53
 Present project    
report as
required
Utilise EMA    
system
Project Risk Assessment

The associated project activities, skills, risks and resources reviewed in the previous section are
further discussed in relation to the project activity any associated risks and a discussion of the risk
mitigation strategies implemented to ensure project completion.

Activity Project Risks Risk Mitigation Strategies for Project Activity Risks
Project not feasible Begin considering project topics before the module start date
Not related to a Discuss potential project topics with peers and tutors on module
Determining the previous module forum
feasibility of project Poor quality Develop several potential project topics in case first choice is
(contacted tutors for literature rejected
planning and support/ Out of date Carry out initial literature search to establish quality and
Suggested back up literature timeliness of literature
projects/presented Non open access Review and adapt the investigative aspect of the project as it
literature of a literature progressed
reputable source) Lack of studies  
from your
investigative aspect
Little relation Discuss project topics with tutor and how they relate to prior L3
between BCI and modules
Project proposal module topics
(define & elucidate the Low quantity of Utilise feedback from tutor to define and consider the potential
area of interest/Define high-quality of the project topic
project type; literature
Research, Project rejection Review available literature for quality, quantity and timeliness
Development or
investigative/determini Literature Background reading to develop understanding of the literature
ng criteria for success) inaccessible
Different criteria of Establish the criteria from which the performance of a BCI is
success of a BCI defined
Determination of a Create a weekly task breakdown until the project due date that
suitable time frame is reviewed to monitor progress and adjust
Rejection Devote time to review L3 module content studied to relate the
Obtaining project content to the desired project
approval No relation to L3 Discuss with module supervisor thoughts, concerns associated
module with the project proposal
Inefficient tracking Utilise Mendeley ecosystem to track and organise literature that
of literature could can be accessed across devices
lead to information
overload
Background reading
into the Development
of BCI Timescale not Establish a timescale, review and adjust it as required
permitting enough

54
 background
research to
critically evaluate
BCI research

Project highly Develop a quick reference glossary to utilise as engaging the

interdisciplinary literature
which would
contribute to the
above risk
Word count Discuss and review content, report breakdown and content
restrictions propose periodically with my tutor
a significant limiting
factor (background
knowledge:
synthesis ratio)
Retaining relation Ensure that the topics from a L3 module are applied to BCI
to computing and
previous modules
Hard to integrate Select sources for comparison and synthesis that utilise similar
and assess multiple metrics without omitting those that are significant yet differ
metrics and
differing criteria of
BCI success
Timescale to Review progress against the project timescale periodically
adequate search accounting for sickness, distraction and unforeseen difficulty
Development and strategy
improvement of search Information Retain a glossary that explains terms and phrases that are
statement overload unfamiliar and research topics
Keeping it related When necessary halt secondary and tertiary readings that are
detracting from the established project topic
Not understanding Spend time researching the content that I am unfamiliar with to
the establish a basic understanding to then build
multidisciplinary
content outside of
computing
Unable to discuss Tailor my project, establishing that its focus is routed in
the topics, computing as such some of this may be overlooked however
Development of strategies and readers may utilise the glossary to further explore the non-
Glossary reasoning computing BCI content
conventions of BCI
research to inform
a computing-based
reader

Too much time The glossary is for repeated, specialist terms and is built on as
focusing on the literature is analysed
glossary, detracting
from the final

55
 report timeline

Not understanding Cursory reading of other BCI meta-analysis from prior years to
the concepts and understand the fundamentals and how research has adapted
fundamentals of
BCI

Spend too much As carrying out this task evaluate the time devoted to
Examination of the time examining the developing an understanding of BCI technologies
types of BCI different types of
BCI
Poorly presenting Establish an understanding of BCI technologies and develop a
the differences quick reference summary
between BCI
technologies
The project Establish a firm size to project sections and establish key metrics
disproportionately to compare via meta-analysis
describes BCI with
little synthesis
Poorly applying the Ensure that I am approaching the project from a user experience
L3 module material and interaction design perspective
to the project topic
Assessment of User Assessment and Utilise Tutor feedback and compare progress to the prospective
Experience and application of these project timeline to recover from any setbacks
Interaction Design topics takes up a
significant portion
of the timeline
Too much time Ensure these issues of LPE computing and information systems
focused on are discussed in relation to BCI by developing current research
developing an
understanding of
Legal, Professional and LPE BCI concerns
Ethical considerations Poorly considering Establish a timeline for this section and adapt as necessary
of BCI LPE concerns in
relation to BCI
technologies
Multidisciplinary Establish a BCI type to focus analysis and compare key metrics
Choosing a type of BCI nature of BCI
for meta-analysis research increases
the risk of
discontinuity
between
experiment
presentation of

56
 results
Discontinuity Review the way in which studies present their results and
between study discuss them to establish data compatible for synthesis
results could result
in difficulty
synthesising the
studies selected

Not able to find a There are several BCI types implemented for several functions,
suitable BCI type select one mimicking an established interface type
for analysis
Unable to find a Review studies researching interface types without omitting
significant number important studies that consider the UX or ID
of similar studies
for comparison and
synthesis
Excluding Ensure the reasoning and goal of a study is considered when
significant studies researching the literature
in the field of BCI to
support the project
due to narrowly
focusing research

Meta-analysis Balance the number of unique studies to the interface type

utilises too many being examined
dissimilar studies
Poorly synthesising The discussion was carried out in the literature review leaving a
studies selected for complete section of the project review for synthesis
meta-analysis
resulting in a report
with little critical
thinking skills
Poorly presented Utilise graphs, charts and establish key metrics to assess the
results data gathered by the meta-analysis
Ineffectively Curate the studies that have been included in the meta-analysis
communicate
metrics and
Conducting findings
meta-analysis Select Review UX & ID research and establish the way in which this is
inappropriate compared to BCI metrics
metrics that do not
support the aims of
the project

Poor amount of Negotiate with Tutor the structure and presentation of the
project dedicated project report
to synthesis of
results
Project not Utilise suitable time management skills to manage my progress,
completed in time review this regularly to remain on course

57
 Rushed Account for sickness, life events, distractions, breaks and
development with consider other issues
little time to edit
the final report
Unprepared for Utilise SVN, save differing versions before commencing work and
Project management project milestones backing up data to Google Drive and OneDrive for version
Loss of project control and data recovery
progress via system
failure, overwriting
save data etc.

Not sure of stage in Note work done at the end of each session and difficulties faced
the project
Inability to Establish work done and consult prospective tasks on the
complete section of timeline
EMA requiring a
recounting of
Project Journal and project
tracking management
experience
Forgotten thoughts, Keep project notebook and pen nearby or create electronic
tasks and reminders to be compiled.
investigation of
areas of interest
Presenting my Use charts and establish structure to provide a clear
review incoherently presentation of my project management skills
Not providing Retain notes of events, times and their impact through
enough discussion development
Review of project
management Unable to be  
critical about my
performance
planning and
executing a project
Poor project  
progress tracking

Poor organisational  
skills

Not completed on Ensure all work is submitted irrespective of its completion rather
time than not submit at all
Little time to Utilise time management which is reviewed continuously
present the report throughout project lifecycle to ensure project progression
in the desired
fashion
Failure to meet the Review EMA guidelines to establish the method of presenting
deadline the project work

58
Submit completed Presenting a poor Upload to the EMA system well in advance of the due time to
project report review of the work ensure there is significant time available for unforeseen issues
done during project
development
The EMA system  
not fully functional

Review of Personal Performance & Future Project Development

Personal Performance

In the end I feel that I have ultimately delivered what I had intended. A report examining methods
of improving the user experience of BCI by examining methods of improving user performance
through considerations of Interaction design and the user experience. Instead of exploring
technological and analytical methods to improve interface performance. As interface design
involves the understanding that user, system and the interface are involved in a symbiotic
relationship.

I have a tenancy to be particularly verbose with a style of grammar that is wanton, I have tried to
reign this in and be more concise and less flamboyant which would be more suited to this form of
project report. The project itself was a good experience to select, develop, research and create a
project report into a specialist topic. Doing so with less help from tutors than prior more
traditional modules studied yet using the skills and knowledge developed to apply them to BCI
and the user experience was memorable. This was due to the unique and highly specialised,
multidisciplinary field of BCI and all that this entailed in the research and development phases of
the project.

I like many others had been impacted by Covid-19 being placed on furlough, living under lockdown
and the impact this was going to have when I finish my degree as I hoped to finish my studies and
change the area of my profession. I am typically goal orientated, planning tasks and executing
them finding satisfaction completing tasks. Throughout my OU experience I studied part to full
time while working full time, commuting and being part of a number of sporting groups and
teams, I very much manage my time as much as possible to get the most out of it.
I had thought during the project I had all the time in the world especially with lockdown and a
cancelled exam however with the huge break in routine I became aimless with a severe mental
health decline that took some time to recover from. While feeling good mentally and enjoying the
research towards the end of the project at the synthesis and the project management review
stage I became more apathetic and lethargic losing some motivation. I had however constantly
been reviewing my progress and created a to-do list to manage my time and remaining milestones
as completing the project report would ultimately reap greater levels of satisfaction.

59
I enjoyed researching and carrying out synthesis in a highly specialised and multidisciplinary field
such as BCI as there was a wide variation in the methods results discussion, data and report aims
were presented. This provided a unique challenge where I had to carryout secondary research to
understand the results to compare to other studies and then combine this to discuss my own
topic. Though this dissonance resulted in my need to adapt the content of my project numerous
times through development though It proved to be an interesting experience to engage a topic
and carry out a more challenging and substantial report than I have on any other computing
module.

Future Project Development

If the project were to be developed further in the future there are a few issues and areas that I
would like to explore and improve to provide a more substantial report.

I had discussed the effect training had on performance however an issue I found when conducting
research was that often users experience levels were listed as mixed or unspecified though this
can be expected due to the multidisciplinary nature and focus of reports.

I found that there was often a substandard amount of information describing training protocols,
system alignment to the user which would be an interesting area of exploration into improving the
user experience.

Studies had no singular method of presenting results due to the multidisciplinary field of BCI
research. As such this made data analysis, manipulation and presentation more complicated than
intended.

The report explored the impact on performance and the user experience based on a number of
key metrics however there were no studies to examine the effect of combined key metrics to see if
there is a cumulative impact to user performance.

Due to the specialist area of research there was a limited number of studies though my reports
number of unspecified experience users should be minimised to improve the training impact to
user performance metric.

60
 Appendix i. Glossary

Action Potentials: Describe the potential between the rise and fall in voltage along a neurons
membrane due to the flow of ions across it.

Analogue-to-Digital Converters: a system which converts analogue signals such as brain patterns
into a digital signal

Application Specific Integrated Circuit: Integrated circuit chips specifically designed for a
specific purpose.

Artefact Rejection Thresholds: When recording a signal received by the brain through the sensors
predetermined limits are set, any artefacts received exceeding these limits results in the readings
rejection.

Bayesian Linear Discriminant Analysis: Is a method of dimensionality reduction, reducing the

amount of random variables by defining principle variables before classification, this technique is
used in areas such as machine learning.

BCI Decoder: In real time the brain signal received is through the use of machine learning methods
compared to decode the users intended actions.

Biocompatibility: Refers to the properties of a material and how they interact with living tissue this
includes toxicity and liability to generate an immune response when introduced.

Biofeedback: Through connection to sensors placed upon the user feedback may be presented so
that the user may gain both awareness of and control over functions.

Blood-brain Barrier: This is a barrier between the blood vessels servicing the brain and brain
tissue it protects the brain from external substances, other substances required by the rest of the
body and aims to maintain an optimum environment to sustain the brain.

Body Schema: An inter disciplinary term, it is a model which tracks in real time limb position, the
bodies spatial properties and the signals that result in sustaining or changing these properties.

61
Brain Computer Interface: A communication method enabling the conversion of brain signals
into commands which may be actioned upon by a computer for a desired effect.
Invasive: Technologies enable direct communication between the brain and computer, through
invasive techniques such as surgery they enable high quality data to be recorded.
Electrocorticography (ECoG): Through a craniotomy electrodes are surgically placed upon the
brain surface to capture high quality electrical activity from the cerebral cortex.
Microelectrodes: miniature electrodes are inserted into brain cell membranes to record neural
signals across an area of the brain without damaging the cell
Non-invasive: These methodologies enable the communication between the brain and computer
interface through non-invasive means, resulting in a lower accuracy signal readings due to the skull
interfering with signal readings though its non-invasive properties make these methodologies the
more popular practice.
Electroencephalogram (EEG): A monitoring method where by electrodes are noninvasively
placed on the scalp and electrical activity generated by the brain is recorded.

Functional Magnetic Resonance Imaging (fMRI): Neuronal activation is coupled along with
cerebral blood flow, these changes are then measured.
Magnetoencephalography (MEG): A technique that maps brain function and activity by recording
the magnetic fields generated by the brain.
Near infrared Spectroscopy: This method of non-invasive recording of brain activity utilises the
electromagnetic spectrum, near infrared to measure its absorption rate by brain tissues as an area of
the brain is stimulated this changes and the difference measured.
Visual Evoked Potentials: A potential which is generated through a visual stimulus this can be
monitored via Electroencephalogram along the occipital cortex which receives and interprets visual
stimuli.
Type:
Active: Outputs are derived from consciously controlled and directed brain activity independent of
external stimulus while reactive systems derive output from external stimulus, passive systems
operate independently of control and stimulus instead generating outputs based upon unmodulated
brain activity.
Bidirectional: Brain Computer Interfaces involve the input of brain signals generating a computer
interface based output, these systems involve the reading of brain signals through sensors that are
decoded and an encoder writes directly to the nervous system.
Sensory: By measuring neural stimulation the system enables the execution of decisions in real
time.

Synchronous: These systems only measure brain signals generated in a specified time frame, while
asynchronous systems continuously measure the brains generated signals offering a more natural
human computer interaction.

BCI illiteracy: BCI’s are complicated to operate often requiring a degree of specialised training or
practice to operate however there are individuals that possess physiological traits precluding them
from effective utilisation of the technology.

Brain Stimulation Technologies

Deep Brain Stimulation: Device surgically implanted deep into the brain which can generate
electrical pulses, stimulating areas of the brain.
Electroconvulsive Therapy (ECT): An electrical current is passed through the brain, this can be
manipulated through voltage dosage and electrode placement.
Focused Ultrasound: Ultrasonic waves are utilised to stimulate areas of the brain by promoting

62
blood flow to the desired area.
Implanted Microelectrodes: Are implanted into the subject to stimulate brain tissue and or record
neural signals.
Transcranial Electrical Stimulation: An electrical current is passed between two electrodes
targeting an area of the brain which is of interest to trigger brain activity.
Transcranial Magnetic Stimulation: An electrical current is induced through the generation of an
alternating electromagnetic current to trigger brain activity.

Canonical Correlation Analysis: Is the study of the linear relations between two sets of variables.
It is the multivariate extension of correlation analysis.

Contralateral Sensorimotor Cortex: Relates to aspects of the sensorimotor cortex that affect the
opposite side of the body from which stimulation occurs.

Cortical Layers: The Neocortex may be separated into six layers each with their own unique
distribution of neurons, connections and corresponding cortical regions.

Decoding Efficacy: Under controlled circumstances brain signals received from a user is
interpreted by the system measures the efficiency of this process to produce the desired result.

Decoding Methods:
Artificial neural networks (ANNs) These systems are based loosely upon he biological neural
networks, each artificial neuron is connected similarly to a synapse, signals may be received,
processed and transmitted to other connected neurons.
Adaptive Logic Network (ALN): This ANN can be described as a set of binary trees with digital
inputs. Every input has connections with any number of tree inputs .These types of trees can then
perform any form of Boolean function.
Bayesian Logistic Regression Network: Can be used to restrict the conditional probabilities for
discrete variables by modelling each variable a logistic regression model can be applied to the
network, the parents of a variable define the covariates. When the interactions between the parent
variables are incorporated as covariates any conditional probabilities may be estimated with little
restriction.
Multilayer Perceptron (MLP): A class of ANN consisting of a minimum of three layers including
an input, processing and output layer, it is a form of feedforward network.
Learning Vector Quantization Network (LVQ): The algorithm prepares a set of prototype
vectors in the observed input data sample domain using these to classify unseen examples, these
samples are prepared and the network then exposed. A winner-take-all approach is used where the
most similar vectors to an input pattern are selected and adjusted to the input vector.
Recurrent Neural Networks (RNN): Node connections form a temporal sequence allowing the
network to exhibit dynamic behaviour. the network can use its internal state to process variable
length sequences of inputs including previous outputs which may be reprocessed as an input or
reserved as a hidden state under direct control of the neural network.
Detection Algorithms:
Kalman Filter: Utilises a series of measurements over time, containing both statistical noise and
inaccuracies to produce estimates of unknown variables. The filter is a two stage verification system
that predicts a value and then updates the measurement continuously to improve the estimated
value.
Population Vector: The sum of the preferred direction of a population of neurons.
Unscented Kalman Filter: Uses a sampling technique to pick a number of sample points around

63
the mean. These points are then propagated through nonlinear functions producing a new mean and
covariance estimate.
Wiener Filter: A filter to compute an estimate of an unknown signal using a similar signal as a
filter, the unknown signal can be filtered and an estimated output produced.
Discrete Classifiers:
Fuzzy Logic Decoder: Based on human observation of decision making being based upon
imprecise information the model is a mathematical means of representing vague and imprecise
information and have the capacity to recognise, represent, manipulate, interpret, and utilise both
vague and imprecise data.
Gaussian Classifier: A generative approach that attempts to model class posterior and input
class-conditional distribution enabling the generation of new samples in the input space.
Hidden Markov Models: A statistical model in which the system is assumed to be a Markov
process whereby the past, present and future state of the system is independent enabling future
states to be predicted not solely upon its full history.

Linear Discriminant Analysis: A method used in machine learning to explore a linear

combination of features that characterizes or separates a set number of classes, the resulting
combination may be used as a linear classifier or reduction tool.

Nearest Neighbours Algorithm: An algorithm that stores all available cases and classifies new
data based on a measure of its similarity in relation to how its neighbours are classified.
Support Vector Machine: An algorithm that sorts data into two categories. it is a form of
non-binary linear classifier. Trained with a data sample classified into two categories, new data
points must then be assigned a category to which they belong.

Dielectric: insulating material or a very poor conductor with high polarization properties when
interacting with an electrical field.

Electrophysiological Recordings: In neuroscience, it involves measurement of physiological

properties and changes such as neuronal electrical activity and action potentials while recording
these results

Event Related Potentials (ERP): The measured brain response that is of direct result of either a
specific sensory, cognitive, or motor event or stimulus.
Contingent Negative Variation: A slow negative cortical potential appearing during reaction time
tasks which develops between the warning and impending stimulus.
Error Related Negativity: An event-related negative potential generated which peaks shortly after
an error is recorded by a subject
Event-Related Desynchronization: Short-lasting attenuation or blocking of rhythms within the
alpha band, this may occur before and during visual stimulation.
False-Positive Spike Events: Is a result inferred from a recorded event that indicates that a
condition has not been met, when in fact it has.
Mental Imagery: To visualise an experience without the appropriate stimulus or outputs presence.
P300 Wave: This is an event related potential linked to the process of decision making.
Slow Cortical Potentials: An ERP that may be naturally triggered or stimulated generating electrical
activity typically in the range of less than 1Hz these shifts in cortical electrical activity last between
several hundred milliseconds to several seconds.
Steady-State Visual Evoked Potentials: Signals that are natural responses to a given visual
stimulus, this may be implemented for other forms of sensory input.

64
feedback:
Abstract: Lacks the specifics to guide forward behavioural progress of test subjects.
Accuracy: Indicates the correctness of a subjects response.
Explicit: The subjects attention is drawn to errors.
Haptic: A generated experience of touch through the application of forces, vibrations, or motions
to the subject
Implicit: There is no overt attention drawn to an error.
Instantaneous: Feedback is provided instantaneously at each stage as the subject progresses.
Motivation: The subject is given approval for progressing through tasks as a method of reinforcing
the desired behaviour.
Proprioceptive: A neural representation of body mechanics.
Sensory: provided by sensory receptors within the sensory systems feedback returned along the
afferent pathways so the brain can monitor the consequence of actions

Feature Vector Construction: In machine learning it is the application of constructive operators to

existing vectors resulting in the construction of new vectors.

Field-Programmable Gate Array: An integrated circuit designed to be configured post

manufacture.

Flip-chip: A method for interconnecting semiconductor devices to external circuitry with solder
bumps located on the chip pads.

Frequency Domain Features: As opposed to using time frequency domain refers to the analysis of
signals with respect to frequency.

Functional Electrical Stimulation (FES): A technique whereby low-energy electrical pulses are
utilised to artificially generate body movements in individuals.

Grandmother Neurons: Hypothetical neurons that represent complex and specific concepts or
objects they only activates when a person experiences or triggers a particular concept or object.

Guiding Systems: How subjects are guided through BCI use

Feedback: An event that occurs when a system output is used as an input as part of a chain of cause
and effect.
Feedforward: Context is provided to the subject prior to the desired action being attempted.
Multimodal: The utilisation of at least two guiding systems to direct a subject.
Descriptive: Subjects are given a series of instructions endeavouring to guide them to a correct
interaction.
Repeat Until Correct: The operation will be repeated until the expectant operation is completed.
Task Specific: Feedback that corresponds specifically to a task.
Taxonomy: A hierarchy of feedback modalities and types based upon the BCI and subjects
abilities.
Operant Conditioning: This method produces good results however extensive training that can be
several months long is required that the subject must learn.

Hemodynamic Response: The delivery of blood to activated neuronal tissues.

Hyperscanning: A method by which multiple subjects in separate MRI scanners can interact while

65
their brain scans are simultaneously captured.

Instructional Design: Involves the development and delivery processes of instructional

specifications to ensure quality instruction of a product, process or system.

Interstimulus Interval: The interval between the offset of one stimulus to the onset of another.

Intracortical Micro stimulation: A technique that stimulates a cross-section of neurons by passing

a small electrical current through a microelectrode.

Kinematics: Calculates the range of motion of points, bodies, and systems without consideration of
the forces that cause their movement.

Kinesthetic Imagination: imagining the sensation of a movement.

Laplacian Spatial Filter: A filter sensitive to noise used to find areas of rapid change such as
edges in an image.

Machine learning: A branch of artificial intelligence it is based on the idea that systems can learn,
identify patterns and make decisions from data with minimal human involvement.

Meninges: The three layers of membranous coverings over the brain and spinal cord.

Multivariate Synchronization Index: A multichannel frequency recognition method.

Muscle Contractions: The activation of tension-generating sites within muscle fibres

Neuroergonomics: Applies neuroscientific knowledge to ergonomic considerations during

development to address the biological level of ergonomic design emphasising the human nervous
system.

Neurofeedback: A form of biofeedback that presents a live display of brain activity.

Neuronal Ensemble Activity: A neuronal ensemble is a collection of nervous system cells

involved in a particular computation and engage as they attempt their particular activity.

Neuronal Rates: The recorded rate as neuronal communication intensifies along with the stimulus
intensity.

Neuroprosthetic: A device that enhances the input or output of a subjects neural system.

Operant Conditioning Approach: A method of learning that utilises rewards and punishments for
recorded subject behaviour.

Oscillatory Brain Activity: The rhythmic or repetitive electrical activity generated by neural tissue
in the central nervous system

Percutaneous Lead: Commonly utilised for permanent implants the device and respective leads are
placed via needle puncture rather than surgical incision.

66
Procedure:
Active Movement: The subject carries out the motor task replicated by the BCI output.
Motor Imagery With Direct Control: The subjects neural activity directly produces an output.
Motor Imagery Without Direct Control: The subject attempts the desired output without a
corresponding motor function.
Passive Movements: The BCI output is triggered without subject input.
Rest: Time between structured tasks and activities.

Prostheses: An artificial device replacing an organic body part.

Polyimide: A plastic known for its thermal stability, good chemical resistance, and excellent
mechanical properties, these may be used in the construction of multielectrode arrays.

Reflex Arc: Reflexes are an involuntary and nearly instantaneous movement in response to a
stimulus made possible by neural pathways which act on an impulse before reaching the brain.

Sensorimotor-Rhythm: Recorded oscillations in magnetic

or electric fields from the sensorimotor cortex.
Amplitude: A measure of periodic variables change over a period.
Peak Alpha Frequency: Utilising an EEG there is a relationship between PAF and physical
fatigue.
Peak Gamma Frequency: Responsible for triggering cognitive functioning, learning, memory,
and information processing.

Slow Cortical Potentials: Shifts in cortical electrical activity lasting from several hundred
milliseconds to several seconds.

Somatosensory Cortex: Part of the brain that receives and processes sensory information.

Spatial Domain Features: A plane where a digital image is defined by the spatial coordinates
of its pixels.

Spatio-Temporal Resolution: To look at data in terms its spatial resolution and temporal
resolution, the detail resolved over a unit of time.

Spike Detection Parameters: Determining the electrical activity triggered close to the
electrodes has higher amplitude than those further away to filter received noise from the desired
area parameters are set.

Stimulus-Locked Intertrace Correlation: Classification using EEG time-locked to the onset of

stimulus as opposed to traditional frequency-domain-based methods, the method is found to be
more robust and suited to experimentation requiring irregular stimulus patterns.

Time Domain Graph: Presents how a signal changes with time.

Training Protocols
Incremental: Classifier are used online directly with training examples added as the system is
being operated as a result this can shorten BCI training sessions.
Traditional: Signals are registered offline and used to train the classifier before users use it.

67
Vasculature: The arrangement of blood vessels in the body.

Windsorizing: transformation of statistical data by limiting extreme values to reduce the effect of
spurious outliers.

68