Comfort Evaluation from EEG Dipole Imaging

Yuna Shigematsu, Yuta Ueji and Atsushi Ishigame

Graduate School of Engineering, Osaka Prefecture University, 1-1, Naka-ku gakuen-cho, Sakai, Osaka

Keywords: Brain Activity, Comfort, EEG, Dipole Imaging, Evaluation.

Abstract: Different people may have different feelings even in the same environment. However, most of the evaluation

index in comfort are based on a fixed standard without considering individual differences. In this study, we

focus on the preference of comfort, and discussed the dipole imaging of brain waves to evaluate the comfort.

The amygdala is said to be one of the parts of the brain related to comfort. In this paper, we stated the

relationship between comfort and the area around the amygdala by dipole imaging.

1 INTRODUCTION

For humans, it is important to prepare an environment

where they spend a lot of time on a daily basis to lead

a comfortable and healthy life. If environmental

comfort can be improved, it can play a major role in

improving quality of life, and further improve

learning efficiency and work productivity (Vernon,

1919).

Depending on the surrounding environment, we

have various feelings such as heat cold, glare, and

noisiness as shown in figure 1. In addition, different

people may have different feelings even in the same

environment due to individual differences, gender

and age differences. Thus it is necessary to provide

each individual with an appropriate environment so

that everyone can have a comfortable life.

Figure 1: Conceptual diagram (Feeling comfort).

Most of the evaluation index are based on a fixed

standard without considering individual differences.

When it comes to a thermal environment, it is

evaluated by 4 environmental elements (the indoor

temperature, humidity, air flow velocity, radiation)

and 2 human body elements (human clothing amount,

metabolism). By using these, we evaluate a thermal

environment such as WBGT (Wet-Bult Globe

Temperature) for environmental elements, PMV

(Predicted Mean Vote), SET (Standard New Effective

Temperature), for human body elements. These index

are created based on the rule of thumb and the

questionnaire, which is a type of the subjective

evaluation.

The method using brain information can be

mentioned as a method of considering individual

differences. Research focusing on the brain, which

controls the majority of biological reactions, requires

expensive equipment. So those research was focused

on reports of applications in the medical field such as

epilepsy and sleep disorders. However, in recent

years, with the sophistication and price reduction of

devices, we have been actively researched on human

sensation by measuring brain function.

There are various types of devices that measure

brain function, such as MRI (Magnetic Resonance

Imaging), MEG (Magnetoencephalography), fNIRS

(functional Near-Infrared Spectroscopy), and EEG

(Electroencephalogram). MRI and MEG have high

spatial resolution as a merit, but they have low time-

resolving ability and restrain the body. On the other

hand, although NIRS and EEG have the

disadvantages that the measurement site is rough due

to the low spatial resolution and it is difficult to

measure deep brain, the device is simpler than MRI

and MEG, and can measure simply covering the

subject with a headgear-like device without restraint

of the body (Teodore, John, Greg, Dennis, Jose,

424

Shigematsu, Y., Ueji, Y. and Ishigame, A.

Comfort Evaluation from EEG Dipole Imaging.

DOI: 10.5220/0010243404240429

In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 5: HEALTHINF, pages 424-429

ISBN: 978-989-758-490-9

2009). For this reason, we evaluate comfort

considering individual differences with EEG in this

paper.

The amygdala is one of the parts of the brain

related to comfort, and it is said that the amygdala is

excited in an uncomfortable state. In this paper, we

infer that there is a difference around the amygdala by

using EEG when comfortable and uncomfortable.

Frequency analysis is mainly used for EEG

analysis. However, detailed analysis can be difficult

due to the complexity of brain function. In addition,

EEG has a high temporal resolution, but has a low

spatial resolution. Therefore, it is difficult to directly

identify the electrical activity in the brain. Thus, in

this study, we decided to use dipole imaging to

identify the signal source by estimating the equivalent

dipole signal intensity distribution on the virtual

surface in the brain from the scalp potential. Then, we

hypothesized that comfort can be evaluated from the

difference around the amygdala when comfortable

and uncomfortable by using dipole imaging. To test

this hypothesis, we conducted an experiment showing

images that give comfortable feelings and

uncomfortable feelings, and analyzed EEG by dipole

imaging. By proving this hypothesis, we reveal that

comfort can be evaluated from amygdala information

by dipole imaging.

By further applying this, it is possible to consider

individual differences in a comfortable environment

for the current control of the environment such as air

conditioning and lighting. By adding the comfort

evaluation index that takes individual differences into

consideration as the element of environmental

control, we perform environmental control that takes

into consideration differences in comfort due to

differences in gender, age, amount of exercise, and

the situation that was placed until just before. By

doing so, we can create a comfortable environment

that suits each situation, and we believe that we can

improve QOL, reduce fatigue, and improve

productivity accordingly.

2 COMFORT AND DISCOMFORT

Comfort and discomfort are one of the most basic

psychological attributes for understanding behavior,

and it approaches a stimulus that causes pleasure but

tries to move away from a stimulus that causes

discomfort.

The amygdala is an important component of the

limbic system located inside the temporal lobe. The

amygdala is thought to play a central role in

controlling emotional behavior (Olds and Milner,

1954; Klüver and Bucy, 1937). It is expressed in

determining the behavior by judging the external

situation by judging whether it is advantageous for

the survival of the individual, the maintenance of the

species or not, and specifically the autonomic nervous

function, awakening, sleep, and attention. It is

considered to have a decisive influence on the

regulation of motor control. The amygdala is agitated

when it becomes psychologically burdensome such as

an unpleasant scene. The prefrontal cortex suppresses

amygdala excitement, but if the load continues to

occur, the amygdala remains agitated, resulting in

increased blood pressure and insomnia. By touching

the body, it is synthesized in the hypothalamus, and

oxytocin is secreted from the pituitary gland, whereby

the amygdala excitement can be sedated.

3 EEG ANALYSIS

The brain is a group of innumerable nerve cells,

which is said to have 14 billion cells, and is said to be

the highest center that not only controls human

thoughts and behaviors but also controls their

emotional and autonomic functions. Nerve cells

communicate with each other by weak electricity via

dendrites emerging from them. This phenomenon

occurs in the pyramidal cells of the cerebral cortex,

and their electrical activities are superimposed on

each other and transmitted to the surface of the head.

The EEG is a measurement of this transmitted

electrical activity.

3.1 Source Imaging

The electroencephalogram is an effective method to

elucidate the brain function in an environment close

to nature because the measurement environment is

not limited and can be easily measured non-

invasively. However, the spatial resolution of EEG is

low due to the limited number of electrodes and the

low conductivity of the skull. Therefore, it was

difficult to identify the electrical activity in the brain

directly from the potential distribution on the scalp

surface. As a method to solve this problem, brain

dipole imaging has been proposed in which the

equivalent dipole signal strength distribution on the

virtual surface in the brain is estimated from the scalp

potential and the signal source is specified. According

to this method, the signal source generated in the

brain can be equivalently expressed by the

distribution of multiple dipole signal intensities on the

virtual surface in the brain, without being limited in

the number and direction. The solution to this inverse

Comfort Evaluation from EEG Dipole Imaging

425

EEG problem is affected by noise due to

measurement and errors in the transfer matrix caused

by distortion during model design. The measurement

noise is caused by measurement environment such as

electrode impedance and artifacts such as blink and

body movement. On the other hand, the error of the

transfer matrix is caused by the distortion in the

model design such as the displacement of the

electrode attachment position, the individual

difference in the head shape, and the variation in

conductivity. Therefore, it is important to consider the

influence of noise in the solution of the inverse EEG

problem, due to estimate brain dipole imaging with

high accuracy (Rush and Driscoll, 1969; Ary, Klein

and Fender, 1981; Salu, Cohen, Rose, Sato, Kufta,

and Hallett, 1990).

3.2 Event-related Potential

Among the observed EEGs, those that spontaneously

and continuously appear on spruce are called

background EEG. Background EEG occurs because

the activity of neurons on the surface of the cerebral

cortex is constantly occurring throughout the cortex.

On the other hand, the brain potential that occurs after

stimulation of receptors and events related to

psychological processes such as perception, attention,

cognition, and memory is called event-related

potential (ERP). Since ERP is a minute potential

change of about 0.1 μV to several tens of μV

compared with the background EEG, multiple

waveforms measured under the same conditions are

arithmetically averaged to identify the ERP

component. In addition, the positive wave that

appears at about 300 ms is called P300 among ERP

(Sidman, Ford, Ramsey and Schlichting, 1990). P300

is thought to be involved in stimulus comparison,

evaluation, judgment, selective attention, and

cognitive context updating.

4 VERIFICATION EXPERIMENT

ON THE COMFORT

EVALUATION

4.1 Experiment Outline

As described in the previous section, the authors

believe that there is the difference of comfort and

discomfort in the amygdala using EEG source

imaging. To verify the hypothesis, we conducted an

experiment to show images that are thought to give

comfortable and uncomfortable feelings to the

subject. We show a conceptual diagram of the

experiments in Figure 2.

For the subject, 10 comfortable images and 10

uncomfortable images (20 in total) specified by

GAPED (Details explain in Chapter 4.2) were

randomly displayed for 3 seconds each. Figure 3

shows the examples of the displayed image. The

procedure was performed for 10 times. We show the

flow of the experiment in Figure 4.

Figure 2: Conceptual diagram of the experiments.

(a) (b)

Figure 3: The examples of the displayed image.

(a) The example of the comfortable image

(b) The example of the uncomfortable image.

Figure 4: The flow of the experiment.

HEALTHINF 2021 - 14th International Conference on Health Informatics

426

4.2 Used Equipment and Measuring

Method

The used instruments were as follows.

-Electroencephalogram measurement system

EMOTIV EPOC+ (14 Channel)

Figure 5 shows EPOC+ used in the experiment.

Figure 5: EMOTIV EPOC+.

-GAPED(The Geneva Affective Picture Database)

A Database of emotional visual stimuli. Those

with great valance and dominance and little

arousal are comfortable images, and the opposite

is uncomfortable images (Dan-Glauser and

Scherer, 2011).

The measured EEG was averaged to remove the

influence of background EEG. Then it was analyzed

by the method described below for the measured

brain waves.

-Dipole imaging

To evaluate the comfort state of the subject, we

determined the content of the dipole imaging(Cuffin,

1998).

A head model is set to estimate the signal intensity

distribution in the brain from the scalp potential. Head

models include sphere models, FEM (finite element

model)(Awada, Jackson, Williams, Wilton, Baumann

and Papanicolaou, 1997), and BEM(boundary

element model)( Fuchs, Drenckhahn, Wischmann

and Wagner, 1998.). However, since BEM and FEM

require MRI images of each individual, the head

model is the one-layer sphere in this study as shown

in Figure 6 (Baillet, Mosher, Leahy, 2001).

Figure 6: Head model.

A dipole layer was virtually placed in the brain of

this head model. On this layer, multiple dipole signal

sources in the radiation direction were installed at

equal intervals. The signal sources generated in the

brain can be equivalently represented by dipoles on

this layer, regardless of the number or direction.

Using the transfer matrix L from this dipole layer to

the scalp surface, the process of observing the scalp

surface potential Φ was modeled by the following

equation.

ΦL∙jn

(1)

j is the dipole signal strength distribution and n is

the noise. The transfer matrix L is determined by the

shape of the head model, conductivity, and electrode

placement.

That equation is a forward problem for dipole

imaging, and solve this inverse problem. In this

study, we solved the inverse problem by mne

(minimum norm estimation) as in the following

equation (Pascual-Marqui, 1999; Gramfort,

Luessi, Larson, Engemann, Strohmeier, Brodbeck,

Goj, Jas, Brooks, Parkkonen, Hämäläinen, 2013).

min |

ΦLj|



(2)

5 RESULT

5.1 Simulation Result

The result of simulation by dipole imaging is shown

in Figures 7 and 8. Figures show the parietal

hemisphere after normalizing the dipole signal

intensities to size 1. The upper part shows the frontal

region. Looking at Figures 7 and 8, there is a strong

dipole only in the simulation result at the time of

discomfort in the right central part (Broken line area),

that is, around the amygdala. That is, it is considered

that the signal source exists around the amygdala only

when the user is uncomfortable.

Figure 7: Simulation result (Comfortable images).

Comfort Evaluation from EEG Dipole Imaging

427

Figure 8: Simulation result (Uncomfortable images).

5.2 Discussion

From the results of this simulation, there was a

difference between the presentation of comfortable

images and the presentation of uncomfortable images.

When the uncomfortable image was presented, the

source signal was seen in the right central part, that is,

around the amygdala. However, in this study, the

results of the comfort image presentation and the

uncomfortable image presentation are added and

averaged to remove noise. Looking at the simulation

results for each image, there were cases where the

source signal was found around the amygdala even

when the comfort image was presented, and no signal

source was found around the amygdala even when the

uncomfortable image was presented. Figure 9 shows

the simulation result when a certain uncomfortable

image is presented.

Figure 9: Simulation result (A certain uncomfortable

image).

According to this result, a weak reaction is seen

around the amygdala, but the reaction is not strong

enough to be called a signal source. From this, it is

considered that the accuracy that can be evaluated in

consideration of individual differences, which was

the initial objective, has not been reached the level at

which individual comfort evaluation can be

performed in consideration of individual differences.

As a method for improving the accuracy, there is an

improvement of the head model. This time, a single-

layer sphere model was used as the head model. There

are various things such as the skull and the brain in

the head and it is not uniform. Therefore, we use a

three-layer sphere model in which the conductivity of

the scalp, the skull, and the brain are separately set

(Sidman, Ford, Ramsey and Schlichting, 1990). It is

possible to perform imaging in consideration of noise

caused by the scalp and skull, and it is considered that

the accuracy can be further improved. In addition, we

used mne as the solution of the inverse problem.

However, when mne is used, the current value may be

estimated over a wide range. Therefore, it is

considered that the accuracy of the signal source can

be further improved by using another method such as

MCE (minimum current estimation), LASSO (least

absolute shrinkage and selection operator), or

hierarchical variational Bayes estimation method.

6 CONCLUSION

In this paper, we examined the evaluation of comfort

and discomfort using dipole imaging. We presented

images that gave comfortable and unpleasant

emotions, and analyzed them using EEG dipole

imaging. As a result, it was found that the source

signal was found around the amygdala when the

uncomfortable image was presented. By adding the

proposed comfort evaluation index, that takes into

account individual differences, for environmental

control such as air conditioning and lighting as an

element of environmental control, we can control the

environment according to the differences in gender,

age, amount of exercise, and the situation that was

placed until just before. By doing so, we can create a

comfortable environment that suits individual

situations, improve QOL, reduce fatigue, and

improve productivity accordingly.

Therefore, the goal is to evaluate comfort

considering individual differences, the head model is

changed to a three-layer sphere model, and the

solution of the inverse problem is improved to

improve accuracy.

In the future, we would like to verify whether the

proposed comfort evaluation is practical. To that

purpose, we would like to compare the currently used

comfort evaluation indexes such as air conditioning,

lighting, and noise with the methods that add the

proposed method. And by adding the comfort

evaluation index proposed as an element of

environmental control such as air conditioning and

lighting as an element of environmental control, we

HEALTHINF 2021 - 14th International Conference on Health Informatics

428

would like to create a comfortable environment that

suits each individual.

REFERENCES

Vernon, H. M., 1919. The influence of hours of work and of

ventilation on output in tinplate manufacture, H.M.

Stationery Off. London

Teodore, W. B., John, K. C., Greg, A. G., Dennis, J. M.,

Jose C. P., 2009. Brain-Computer Interfaces: An

international Assessment of research and development

trend / Edition 1, Springer Nertherlands

Olds, J., Milner, P., 1954. Positive Reinforcement Produced

by Electrical Stimulation of Septal Area and Other

Regions of Rat Brain. Journal of Comparative and

Physiological Psychology, 47(6), 419–427

Klüver, H., Bucy, P. C., 1937. “Psychic blindness" and

other symptoms following bilateral temporal lobectomy

in Rhesus monkeys, American Journal of Physiology,

119, 352–353

Rush, S., Driscoll, D. A., 1969. EEG electrode sensitivity—

An application of reciprocity, IEEE Transactions on

Biomedical Engineering, 16, 15-22

Ary, J. P., Klein, S. A., Fender, D. H., 1981. Location of

Sources of Evoked Scalp Potentials: Corrections for

Skull and Scalp Thicknesses, IEEE Transactions on

Biomedical Engineering, 28(6), 447-452

Salu, Y., Cohen, L. G., Rose, D., Sato, S., Kufta, C., Hallett,

M., 1990. An improved method for localizing electric

brain dipoles, IEEE Transactions on Biomedical

Engineering. 37(7), 699-705

Sidman, R., Ford, M., Ramsey, G., Schlichting, C., 1990.

Age-related features of the resting and P300 auditory

evoked responses using the dipole localization method

and cortical imaging technique, Journal of

Neuroscience Methods. 33(1), 23-32

Dan-Glauser, E. S., Scherer, K. R., 2011. The Geneva

affective picture database (GAPED): a new 730-picture

database focusing on valence and normative

significance, Behavior Research Methods, 43(2), 468-

477

Cuffin, BN., 1998, EEG dipole source localization, IEEE

Eng Med Biol Mag. 17(5), 118-122

Awada, K. A., Jackson, D. R., Williams, J. T., Wilton, D.

R., Baumann, S. B., Papanicolaou, A. C., 1997.

Computational aspects of finite element modeling in

EEG source localization, IEEE Transactions on

Biomedical Engineering, 44(8), 736-752.

Fuchs, M., Drenckhahn, R., Wischmann, H. A., Wagner, M.,

1998. An improved boundary element method for

realistic volume-conductor modelling, IEEE

Transactions on Biomedical Engineering, 45(8), 980-

997

Baillet, S., Mosher, J. C., Leahy, R. M., 2001.

Electromagnetic brain mapping, IEEE Signal

Processing Magazine, 18(6), 14-30

Pascual-Marqui, R., 1999. Review of methods for solving

the EEG inverse problem, Bioelectromagnetism, 1, 1-

Gramfort, A., Luessi, M., Larson, E., Engemann, D.,

Strohmeier, D., Brodbeck, C., Goj, R., Jas, M., Brooks,

T., Parkkonen, L., Hämäläinen, M., 2013. MEG and

EEG data analysis with MNE-Python, Frontiers in

Neuroscience, 7, 267

Comfort Evaluation from EEG Dipole Imaging

429