Computer-Brain Interface Device Experimental Research

Study of Ergonomy and Usability in Business Environment

Radim Brixí, Jaroslav Kalina and Pavel Němec

Department of System Analysis, University of Economics in Prague, nám. W. Churchilla 4, Prague, Czech Republic

Keywords: BCI, EEG, Interface, Brain Waves, Commercial Devices, Epoc Headset, Business Application.

Abstract: This paper presents results of experimental research from 2011, performed on a group of volunteers and

which aim was to properly explore the maturity of BCI devices to be integrated in business environment,

e.g. office work. The testing scenarios and focus was aimed to verify if the current capabilities of the

selected devices were sufficient in their ergonomic features and spectrum of functions to properly

complement and in some case substitute the traditional ways of utilising keyboard and mouse by the users.

The research focused also on ability to learn to use BCI device. The study led to identification of several

ergonomic and hygienic areas which act as a hindrance for wide utilisation of these devices as a feasible

way to provide an users input for the computer over a prolonged period of time.

1 INTRODUCTION

1.1 Aim of the Paper

The utilisation of electroencefalograhical (EEG)

signals as a mean of communication between human

and computer represents one of the greatest tasks

within the borders of theory of signals research

(Bazztarica, 2002). The crucial part of systems of

communication known as BCI (Brain-Computer

Interface) is the interpretation of EEG signals.

The electrical nature of human brain is known

for more than one century and the first successful

recording of human brain electric activity was

performed by a German psychiatrist and neurologist,

Hans Berger in 1924 (Berger, 1940). Berger utilised

a galvanometer and electrodes created from a

platinum wire, implanted under the test subject’s

skin.

Technology accessible to a broader spectrum of

population is still in it’s infancy, first economically

affordable device together with dedicated software

for the end users became available in 2009 (Dillow,

2010); (Hanlon, 2008).

These days with a 3 years long development and

creation of several companies pioneering this

marker, we have a spectrum of cheap devices that

are affordable almost for everyone.

This paper summarises results of a brief study

performed on a group of students during 2011 at

University of Economics, Prague. The aim of this

study was to assess the feasibility of this innovative

way of interfacing with computer to substitute the

traditional ways of interacting with computer. The

research also focused on ability of the students to

fulfil given tasks and we focused on the learning

ability of the students.

1.2 BCI

We consider as a BCI technology such one that

possesses an interface through which human EEG

signal can be captured and transmitted to further

processing (Berger, 2008). Alternative labels are NI

– Neural Interface and BMI – Brain Machine

Interface. (Hatsopoulos, 2009)

We distinguish two types of this technology.

Invasive version requires to be physically implanted

into the brain to allow capturing the EEG signal.

Non-Invasive types attempt to capture the set of

signal by using external devices that do not need to

be implanted into the subject’s body. The current

trend is to develop more advanced non-invasive

types of devices that will reach the quality and

precision of their invasive counterparts (Graimann,

2010).

The current trend in the area of commercial BCI

technologies is to penetrate into the already

established segments and provide and alternative

method of controlling the computer for the purpose

Brixí R., Kalina J. and N

emec P..

Computer-Brain Interface Device Experimental Research - Study of Ergonomy and Usability in Business Environment.

DOI: 10.5220/0003990200860090

In Proceedings of the 14th International Conference on Enterprise Information Systems (ICEIS-2012), pages 86-90

ISBN: 978-989-8565-12-9

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

of entertainment or to substitute classical ways of

input by assigning certain functions to defined brain

activity patterns (Beschizza, 2006).

The utility of alternative means of machine

control and possibly in the future on higher

interconnection or merging between the machines

and their users, and the diminishing of the clear

borders between then may lead to potential ethical

problems as well and redefinition of the relationship

between humans and machines (Kotchetkov, 2010).

Other implied ethical issues into business

environment also are necessary to take in account

(Sigmund, 2010). BCI technologies are innovative

approaches facing problems described in Mildeová

and Brixí (2011).

2 STUDY

2.1 Study Design

The study was conducted on a group of 10 separate

volunteers in the age of 20-26 years by using

commercial end-user EPOC headset BCI device

from EMOTIV. Each participant went through 12

separated testing sessions invoking selected set of

functions through contraction of facial muscles,

simulating mouse movement by inbuilt gyroscope

and providing user input solely through cognitive

functions of the brain.

All participants went through 12 session, each

long 10 minutes and focused on invoking predefined

functions in the testing program interface that were

bind to combinations of facial expressions.

Table 1: Overall cognitive training time per participant.

Test subject

Overall cognitive training duration in

minutes

1M 480

2F 840

3M 850

4M 610

5M 800

6F 840

7M 840

8F 840

9M 840

10F 860

Table 2: Overall time summary of the study.

Participants 10

Sessions 12

Avg. total time per person 935 min.

Avg. session time per person 78 min.

Total time of sessions 9350 min.

Mouse movement simulation took altogether

time of avg. 35 min. spread across testing sessions

during the study.

2.2 Facial Expression Recognition

The training scenario practiced in this section

consisted of mapping facial expression recognizable

by the neuroheadset to 4 alphanumerical characters

to write a 3x word “Hello” by their invoking,

without making a mistake. Each tested subject was

allowed 25 attempts. Table 3 summarizes the

amount of successful attempts.

Table 3: Writing through facial expression letter invoking.

Participant No. of successful attempts. Attempts total.

15 25

7 25

17 25

4 25

0 25

10 25

0 25

7 25

2 25

10F

0 25

Total

62(25,5 %) 250

Additionally two tests were performed on the set

of participants, during which the participants had to

invoke randomly one of the four previously bound

letters through their facial expressions. A simple

program was built to randomly generate the required

letters and the users had to react with appropriate

facial expression to write the required letter. The

first contained a series of 25 attempts and users were

allowed see virtual representation of their faces on

the screen. The anticipated success rate was decided

upon 70 %.

Test no. 2 was a variation on the first one and

users were not allowed to see the virtual

representation of their faces on the screen. The

anticipated success rate was lowered to 60%.

2.3 Cognitive Activity Recognition

The series of tests focused on usability of computer

functions keybinded to a recognized patterns of

brain activity is of utmost importance from the point

of view of this study, as it is the testing of proper

BCI part of the Emotiv Epoc device.

Participants were asked to focus on a virtual cube

model and to imagine specific movement patterns

(rotation and movement in a specific direction). The

Computer-BrainInterfaceDeviceExperimentalResearch-StudyofErgonomyandUsabilityinBusinessEnvironment

patterns captured by during this calibration were

assigned to defined movement types of the virtual

model.

The aim of this series of tests was to assess the

reliability of this way of exchange of information

with the computer.

Figure 1: Random facial expression function invoking test

results.

First test consisted of asking the participants to

attempt to invoke a demanded movement of the 3D

model cube through the previously bound brain

patters as set of random 25 instructions were

generated for the users to attempt to follow.

Required success rate was decided on 67%.

Test no. 2 required users to again invoke 25

times, in randomly generated order, one the

previously trained brain activity patterns and

through binding with selected keyboard button

combinations call functions assigned to these

buttons in MS Project Professional 2010, an office

application for planning of projects. Decided success

rate was decided on 60%.

Figure 2: Random brain activity pattern invoking test

results.

3 RESULT DISCUSSION

3.1 Functional Attributes

The obtained data from observation suggest that the

main issue that may hinder utilization of these

devices for a serious work is precision and

reliability.

The mouse movement simulation through the

inbuilt gyroscope proved to be the only type of

activity the where the users scored success all of the

times. Also the both series of facial expressions

control tests proved successful in case 90% of the

users against decided threshold. However this high

success rate is only for a set of discrete attempts

from test 1 and 2. When faced with a more complex

task (writing “Hello” through invoking keys bound

to certain facial expressions) only 25,5 % attempts

led to a flawless execution.

Test series focused on true BCI aspects of the

device were even less positive. Invoking movements

of the model 3D cube, over desired threshold, was

successful only for 50% of the tested subjects. The

most important part of this study, working with an

existing office application, e.g. invoking 3 functions

(calling the menu, escaping the application, opening

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

settings, etc.) by binding them to brain activity

patterns proved to be above the agreed threshold

only in case of 30% of the users.

These low rates basically inhibit the utilization of

these devices for office work, acting in the current

state more as a hindrance, than as a tool to improve

working with the computer in office conditions.

Each tested subject spent in avg. 935 minutes

training with various aspects of the device. We do

consider this time to be an unbearable obstacle in

real business environment, especially accompanied

with the low reliability the test data have uncovered.

Even though 935 minutes is a relatively long time

for learning, it is matter of further testing how much

can one improve when working with BCI device

very much longer (for example as long as one learns

to write by hand).

3.2 Non-functional Attributes

3.2.1 Subjective Repulsion from Wet

Sensors

Tested users reported a continuous repulsive feeling

accompanied with application of wetted sensors on

their skin. The ends of sensors made from foam or

dross need to be drenched properly in conductive

salt fusion to increase the quality of received EEG

signals by the sensors.

From practicality point of view this has been

issue for people with longer hair, especially female

participants. The ideal user for employing the non-

invasive class of EEG sensors is ideally free of hair

cover completely.

3.2.2 Hygiene Issues

Prolonged contact with foam or dross drenched in

the conductive fusion on the skin can lead to skin

irritation over time. Especially individuals with more

sensitive types of skin are susceptible to this effect.

From hygienic point of view, it is recommended

to have a spare set of sensor endings to avoid

possible transmission of skin diseases, lice and

dandruffs.

Keeping the ends of sensors in humidor to

preserve the wetness needed to properly transmit the

signals creates an ideal environment for growing of

funguses and mould.

3.2.3 Drying of Sensors

The necessity of keeping the ends of the sensors

drenched in the conductive fusion brings a problem

for scenarios that take a prolonged period of time,

e.g. a simulation of a typical 8 hours office work

day. The drying during the scenario leads to a

reduction of the quality of perceived signal and

deteriorates the capability of the BCI device perform

it’s function as a computer interface. Additional

reapplications of the fusion, lead to a necessity to

stop working and take of the headset, re-apply the

fusion and arrange the sensors once again on the

designed spots of the head. This re-application is a

source additional time-loss.

The usage of conductive salt fusion leads in time

to creation of salt sediments on the surface of the

sensors and the foam ends. Newly available dry

sensors should solve most of the problems, but the

testing of proper functionality of these dry sensors is

a matter of other future research.

3.2.4 Pain, Frustration and Loss of

Concentration

Tested users have reported a growing pressure due

the frame of the sensor device on their heads,

leading to feeling of physical pain.

Table 4: Avg. times of continuous wearing of the

neuroheadset without perceived pain.

Test Subject Avg. time without interruption (minutes)

1M 60

2F 38,21

3M 60

4M 53, 08

5M 30

6F 60

7M 60

8F 60

9M 42,5

10F 43,75

Our prior aim was to test if the users can wear

the neuroheadset at least 60 minutes continuously.

Table 4 summarizes the time during which were the

users able to suffer the device on their heads. After

this time the perceived pain and loss accompanied

loss of concentration grows rapidly, making further

work unfeasible. The 60% success rate indicates that

the device is in a significant number of cases

problematic for long term wearing ergonomically.

3.2.5 Preparation Time

Preparing and proper positioning the device on

user’s head requires certain amount of time as well-

mostly to drench the sensor ends in the conductive

fusion and set the ends on designed points on the

head. Table 5 provides the average preparation time

per tested subject over all 12 sessions.

Computer-BrainInterfaceDeviceExperimentalResearch-StudyofErgonomyandUsabilityinBusinessEnvironment

Table 5: Avg. preparation time per users.

Subject Avg. preparation time (minutes)

1M 3:13

2F 4:05

3M 2:16

4M 1:15

5M 2:29

6F 3:57

7M 1:43

8F 2:47

9M 2:39

10F 4:02

It is noticeable that the avg. time to set the device

properly on the head was noticeably longer with

females, mostly due to complications accompanied

by longer hair. All tested users had to set-up the

device on their heads by themselves.

The whole sample of volunteers was tested on

exactly one device. This traffic generated enough

strain on the BCI device’s cover to lead to chipping

off several parts used to fixate the device properly

on the skull.

4 CONCLUSIONS

The study has identified several areas ranging for

hygiene, ergonomic aspects, and conformability to

concentration issues that from our point of view

represent an obstacle for serious utilisation of these

devices in day to day life of end users. However

there is significant success in achieving successfully

not all but at least some tasks by users. This fact

opens possibilities for future improvements of

ergonomic issues and preciseness based on number

of sensors of BCI device together with more time for

learning to use such a device. If development is

focused especially on presented issues we believe,

that success in the tests should get noticeable better.

Regardless of the critique backed by the study’s

results, and remaining lack of proper robustness and

spectrum of capabilities of the device used during

the study, we are enthusiastic about the

developments in this area and it’s future implications

for the computer industry, user experience and the

implications for development of business software

applications.

In the perspective of the upcoming decades BCI

devices may contribute to fundamental changes in

several areas:

 Layout changes of the distribution of control

elements in graphical user interfaces or web pages

that will be better handled by users accessing then

by using BCI device as their primary way of

interacting with the computer.

 Gaming industry is currently the most promising

are, where BCI devices can be used even with

relatively limited spectrum of different inputs they

can record. The potential of BCI lays in enhanced

immersion of the virtual experience.

 Training and therapeutic utilisation of the device

for practicing concentration and self-control.

ACKNOWLEDGEMENTS

Authors would like to thank the faculty of

Informatics and Statistics, University of Economics

in Prague, to allow this branch of research to be

followed within the borders of the programme of

applied information science and institutional

research funding on Department of System Analysis.

Authors also thank all participating volunteers

for cooperation and time dedication which was

provided without any financial compensation.

REFERENCES

Baztaricca, O. J., 2002, EEG Signal Classification for

Brain Computer Interface Applications. Ecole

Polytechnique Federal De Lausanne.

Berger, H., 1940. Psyche. Jena: Gustav Fischer, p.6.

Berger, T. W., Chapin, J. K., Gerhardt, G. A., McFarland,

D. J., 2008 Brain-computer Interfaces: An

International Assessment of Research and

Development Trends. Springer.

Beschizza, R., 2006, Hitachi: Commercial Mind-Machine

Interface by 2011. In Wired.

Dillow, C., 2010, The World's First Commercial Brain-

Computer Interface. In Popsci.

Graimann, B., Allison, B., Pfurtsheller, G., 2010, Brain-

computer Interfaces: Non-invasive and Invasive

Technologies. Springer.

Hanlon, M., 2008, The first commercial Brain Computer

Interface. In GizMag.

Hatsopoulos, N. G., Donoghue, J. P., 2009, The Science of

Neural Interface Systems. In Annu Rev Neurosci.

NBCI.

Kotchetkov, I. S. et al., 2010, Brain-computer Interfaces:

Military, Neurosurgical, and Ethical Perspective. In

Neurosurg Focus.

Mildeová, S., Brixí, R., 2011 The Limits of ICT for

Innovations and Economic Growth. In: IDIMT-2011.

Linz: Trauner Verlag universität, 2011, p. 157–164.

ISBN 978-3-85499-873-0.

Sigmund, T., 2010. Is business ethics possible? In: IDIMT-

2010 Information Technology – Human Values,

Innovation and Economy. Linz: Trauner, 2010, p.

303–310. ISBN 978-3-85499-760-3.

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems