Impact of Music on Human Brain Activity during Mental Stress

Roman Mou

cek

and Kl

ara Ber

ankov

Department of Computer Science and Engineering, Faculty of Applied Sciences, University of West Bohemia,

Univerzitnı 8, Plzen, Czech Republic

Keywords:

Electroencephalography, Heart Rate, Human Brain, Mental Stress, Music.

Abstract:

Because music is an integral part of our lives, every step towards a better understanding of its impact on

humans is beneﬁcial. This paper deals with the impact of several types of music on the human brain activity

during mental stress. An experiment was designed, and electroencephalography data, heart rate data and data

from questionnaires were collected, processed and analyzed. All these steps are described and a subset of the

large collection of results is presented.

1 INTRODUCTION

Music naturally accompanies our lives, evokes

a plethora of emotions, both positive and negative,

and affects our mood, physical and mental perfor-

mance. A better understanding of what type of music

and how it affects individuals can be very beneﬁcial,

for example, for the treatment of some diseases or just

for improvement of concentration and cognitive per-

formance. Some effects of music are nowadays com-

monly used in medicine, advertising, entertainment

and other areas. However, how a particular type of

music affects us is very individual and closely related

to our experience, preferences and current situation.

The aim of this paper is to contribute to the un-

derstanding of how different types of music affect

the activity of the human brain during mental stress.

The presented research focuses on the acquisition and

analysis of the electrical activity (that is considered

to be inﬂuenced by listening to music) of the human

brain by using the method and technique of electroen-

cephalography. Heart rate is collected as a source of

additional electrophysiological data.

The paper is organized as follows. The state of

the art section presents experiments and ﬁndings that

have been made in connection with the research of

the impact of music on humans. The next section

introduces the design of the experiment; the course

of the experiment, hardware and software equipment

used, participants, and acquired data are described.

The data processing section comes with the basic pro-

cessing methods and results applied to the collected

https://orcid.org/0000-0002-4665-8946

data. The data analysis section provides selected re-

sults from the data evaluation and interpretation pro-

cess. The last session concludes the obtained ﬁndings.

2 STATE OF THE ART

In this section we introduce some existing studies and

ﬁndings concerning on the impact of music on the

physical and mental state of individuals in various

situations. These studies have contributed in part to

shaping the design and course of our experiment.

The impact of music, speciﬁcally tempo, rhythm,

melodic structure, pause, individual preference, ha-

bituation, order effect of presentation, and previous

musical training on changes in the cardiovascular and

respiratory systems that could be potentially used for

stress modulation is described in (Bernardi et al.,

2006). The participants were initially allowed to rest

for 20 minutes while their cardiovascular and respi-

ratory indicators were measured. Furthermore, six

different music styles were played in random order

and two minutes without any music were randomly

inserted into the music. At the end of the experi-

ment, the participants rated each song according to

their preferences. The results obtained on 12 prac-

tising musicians and 12 age matched non-musicians

showed that music induced an arousal effect, predom-

inantly related to the tempo. This effect was also inde-

pendent of the preferences of individual participants.

Other interesting results included, for example, that

no habituation effect was seen and that musicians had

greater respiratory sensitivity to the music tempo than

750

Mou

cek, R. and Beránková, K.

Impact of Music on Human Brain Activity during Mental Stress.

DOI: 10.5220/0009175407500757

In Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2020) - Volume 5: HEALTHINF, pages 750-757

ISBN: 978-989-758-398-8; ISSN: 2184-4305

non-musicians did.

The impact of listening to music on cognitive per-

formance was investigated in (Dolegui, 2013). The

study focuses on what is the impact of different gen-

res of music, played at different volume levels, on

the cognitive abilities of college students when com-

pleting academic tasks. The results showed that the

participants performed best when they did not listen

to any music during the tasks. On the other hand,

high-intensity heavy-metal music dramatically wors-

ened their performance. Despite expectations, listen-

ing to classical music did not conﬁrm an increase in

subjects’ performance in the tasks.

The study (Lesiuk, 2005) examined the impact of

listening to music on the state of positive affectivity,

the quality of work, and time-on-task of computer in-

formation systems developers. The data was collected

from ﬁfty-six developers from four different Cana-

dian software companies in their common work en-

vironment for ﬁve weeks. The results of this study

suggested that listening to music in some work envi-

ronments provoked a positive mood change and im-

proved the quality of work done while reducing work

time.

The experiment described in (Daly et al., 2014)

deals with the emotional change induced by listening

to music. One hundred and ten snippets of ﬁlm mu-

sic were selected for stimulation, including many dif-

ferent musical styles, offering a rich stimulus material

for exploring emotional processing. During the whole

experiment, the EEG signal was obtained from nine-

teen electrodes placed according to the international

10/20 system. The researchers found neural corre-

lates of music-induced emotion in a number of fre-

quencies over the pre-frontal cortex., e.g. between the

emotional state of the participants and the frequency

band changes in the 18 to 22 Hz range (beta activ-

ity). The study suggested that the observed changes

in EEG recordings induced by a change in a person’s

emotional state did not depend on whether emotions

were induced by listening to music or they were in-

duced by other stimuli.

There are many brain imaging studies focusing

on brain regions that are activated by familiar mu-

sic. The issue how is the brain response inﬂuenced

by the familiarity of music using electroencephalog-

raphy was addressed in (Kumagai et al., 2017). The

EEG signal was analyzed to investigate the relation-

ship between cortical response and familiarity of mu-

sic (melodies produced by piano sounds). The cross-

correlation function averaged across trials, channels,

and participants showed two pronounced peaks where

the magnitude of the cross-correlation values were

signiﬁcantly larger when listening to unfamiliar and

scrambled music. These ﬁndings suggested that the

response to unfamiliar music was stronger than that

to familiar music.

Two following articles deal with the issue of de-

coding music preference from EEG signals. A time-

windowing feature extraction approach for investi-

gation of the time-course of the discrimination be-

tween musical appraisal electroencephalogram (EEG)

responses was presented in (Hadjidimitriou and Had-

jileontiadis, 2013). The used EEG data set contained

the responses of nine subjects during music listening;

self-reported ratings of liking and familiarity were

also collected. The features were extracted from the

beta and gamma EEG bands and two classiﬁers were

used to classify feature vectors into two categories

(like, dislike) under three cases of familiarity (regard-

less of familiarity, familiar music, and unfamiliar mu-

sic).

A single-sensor EEG biomarker for the assess-

ment of spontaneous aesthetic brain responses dur-

ing music listening was introduced in (Adamos et al.,

2016). It reﬂects the listener’s fondness for music

regardless of the emotions induced by it. The par-

ticipants were not engaged in any cognitive task and

asked not to produce any kind of active response. The

EEG recordings were performed in a common envi-

ronment using a wireless EEG device. The resulting

tool was considered to facilitate the personalization of

modern music recommendation systems.

Many other studies have been carried out on how

music affects humans, and although their conclusions

do not always coincide, research has shown that mu-

sic can affect humans both physically and mentally

through emotions. Examples of such effects include

e.g. changes in respiratory rate and heart rate, brain

frequency and improved work performance. How-

ever, as these experiments have further shown, the

effects of music on humans are very individual. It

depends on the speciﬁc characteristics of the music

being played, the person’s experience of listening to

it and its musical preferences.

In view of these ﬁndings, it was assumed that

listening to different types of music while playing

a memory game (which represents mental stress) will

have an impact on a variety of electrophysiological

signals which can be captured from humans. In our

case the effect of music is considered to be trans-

lated into a change in the energy level of brain activ-

ity, heart rate, and in-game performance. At the same

time, the behavior of experiment participants (and the

nature of their data) will be inﬂuenced by their musi-

cal preferences and their overall experience with lis-

tening to music.

Impact of Music on Human Brain Activity during Mental Stress

751

3 DESIGN AND COURSE OF

EXPERIMENT

This section describes the design and course of the

experiment dealing with the inﬂuence of music on

human brain activity during mental stress. The elec-

troencephalographic (EEG) signal and heart rate data

were captured during the experimental sessions as

well as data related to the mental stress of partici-

pants.

3.1 Experiment Scenario

The participants of the experiment (subjects) played

a memory game (representing mental stress) of two

different levels of difﬁculty either listening to three

different types of music or in a quiet environment.

Prior to the start of the experiment, the subjects

were allowed to rest while measuring their heart rate

for ﬁve minutes. After this time period, the subject set

the volume of the music played on the headphones to

the level he/she was used to. He/she played one trail

of the memory game to get acquainted with the game

and to know how to switch to a new game.

The experiment was divided into two blocks (two

difﬁculty levels). Each block consisted of four ses-

sions that differed only in the type of music played:

• Playing the memory game in a quiet environment,

• Playing the memory game while listening to fast

music,

• Playing the memory game while listening to slow

(relaxing) music,

• Playing the memory game listening to music pre-

ferred by the participant.

Each session lasted four minutes and contained

a four minute musical composition of one of the types

of music described above. During each musical com-

position, the subject tried to reveal as many pairs of

memory cards as possible with the highest accuracy.

The subject could complete the memory game sev-

eral times during one musical composition. There was

a 15 second pause between the sessions for the subject

to prepare for a new game. In the second block, the

difﬁculty of the game increased - the playing cards

were more similar. In the ﬁrst block the pictures on

playing cards were colored and more different (differ-

ent musical instruments), while the second block they

were black and white and contained only one type of

picture (owl), which was slightly varied. Between two

blocks there was a ﬁve minute pause, during which

the subject could relax while switching the difﬁculty

of the game. After completion of the experiment, the

subject was rested for ﬁve minutes while his/her heart

rate was still measured.

Throughout the experiment, the EEG signal from

the Cz, Fz, and Pz electrodes was captured. At the

same time, heart rate, and average accuracy together

with the number of pairs successfully found in the

memory game were recorded.

3.2 Hardware Equipment

The experiment was performed in (no info for the

review process). It was equipped with all neces-

sary hardware infrastructure for EEG and heart rate

recordings. The V-Amp ampliﬁer produced by the

Brain Products company was used for EEG record-

ings. Intraaural earphones were used to isolate the

subjects from the surrounding noise and to respect the

constraints caused by wearing the EEG cap. Com-

mon ECI EEG caps (the 10-20 system deﬁning lo-

cations of scalp electrodes) were used depending on

the size of the subjects’ heads. The reference elec-

trode was placed approximately 1 cm above the nose

and the ground electrode was placed on the ear. The

Xiaomi Miband 2 smart wristband was used to mea-

sure heart rate. The wristband was connected to the

mobile application Tools & Mi Band to allow contin-

uous heart rate recording during each measurement.

Two computers were used, the ﬁrst one for visualiz-

ing, recording and storing EEG data, and the second

one for playing a memory game.

3.3 Software Tools

The BrainVision Recorder was used to visualize and

record the EEG signal. A simple and user friendly

web application Pairs One (Gorin, 2016) was chosen

for the experiment as a suitable representative of the

memory game. It allows users to customize the game

easily. The software Matlab R2015b with freely avail-

able extensions EEGLAB v13.5.4b and ERPLAB 5.0

were used for processing and analysing the captured

EEG data.

3.4 Music Compositions

The music played to the subjects during the experi-

ment was divided into three categories: slow, fast and

preferred one. The musical compositions labeled as

slow and fast music were selected from the same artist

and genre. They did not contain any lyrics, so that

the text did not distract the subject from playing the

game. The preferred musical compositions were se-

lected directly by the subjects. They were instructed

to select two of their favorite musical compositions.

HEALTHINF 2020 - 13th International Conference on Health Informatics

752

Table 1: Detailed information about the music played during the experiment.

Type Author Musical Composition (1st level) Musical composition (2nd level) Style

Fast Metallica Master of Puppets Enter Sandman metal

Slow Kevin MacLeod Sovereign Fresh Air classic

Figure 1: The frequency spectrum of subject 1 on the electrode Fz.

No restrictions were put on the subject in this selec-

tion.

Finally two sets of musical compositions were

created, each containing one musical composition

from each category. The ﬁrst set of musical compo-

sitions was played during playing the ﬁrst difﬁculty

level of the memory game while the second set of mu-

sical compositions was played during playing the sec-

ond difﬁculty level of the memory game. The detailed

information on individual musical compositions (for

the fast and slow music types, i.e. not for the music

preferred by the participants) is given in Table 1.

3.5 Subjects

Twenty volunteers between the ages of 21 and 30 par-

ticipated in the experiment, including ten men and ten

women. Prior to the start of the measurement, all vol-

unteers got necessary information about the experi-

ment and were asked to give informed consent. In-

formed consent was obtained from all of them.

3.6 Data and Metadata

EEG data were recorded with the sampling frequency

of 1 kHz; no ﬁlters were used during data record-

ing. The resulting signal was stored into three ﬁles

of the BrainVision format but two ﬁles were impor-

tant in this case (.eeg ﬁle containing raw data, and

.vhdr ﬁle containing metadata). All recorded data to-

gether with collected metadata were stored into the

EEG/ERP portal, they are publicly available for reg-

istered users.

3.7 Questionnaire

At the end of the experiment, each subject was asked

to ﬁll in personal information and questionnaire. The

questionnaire contained twelve questions relating to

the experiment, e.g. how the participant was inﬂu-

enced by different music during the measurement,

how boring the memory game was at the end of the

experiment or how stressful the overall measurement

for the subject was.

4 DATA PROCESSING

4.1 EEG Data Processing

The changes in energy levels on the alpha and beta

frequency bands when playing the memory game un-

der different conditions (listening to different types of

music) were primarily investigated. The Fast Fourier

Impact of Music on Human Brain Activity during Mental Stress

753

Table 2: Energy levels for all subjects in alpha and beta bands - the ﬁrst block of the experiment.

Subject 1st session 2nd session 3rd session 4th session

α β α β α β α β

1 287,2 745,7 288,7 736,5 288,2 751,8 288,8 742,7

2 273,3 702,7 273,6 695,9 273,8 691,9 270,8 687,9

3 293,2 783,1 293,0 776,5 290,3 774,1 292,5 773,6

4 298,2 747,0 298,3 744,0 295,8 738,9 294,8 738,8

5 303,7 779,3 302,8 775,0 301,6 777,5 299,8 770,2

6 302,8 742,9 301,0 745,8 298,7 742,1 301,9 748,0

7 287,1 767,0 287,0 800,3 288,0 821,4 290,2 818,2

8 283,6 733,8 284,1 748,1 286,5 759,0 285,8 753,3

9 289,1 746,5 288,6 744,7 286,1 735,2 286,4 734,3

10 287,8 769,0 293,2 777,3 292,9 776,1 294,2 767,3

11 303,5 757,5 304,2 752,7 302,6 744,8 301,2 744,7

12 290,6 741,4 289,6 740,4 284,0 734,3 292,5 737,5

13 291,2 758,3 296,0 800,9 289,8 758,1 294,2 795,0

14 274,8 724,6 275,2 724,2 277,4 722,2 283,6 737,2

15 289,4 730,1 287,8 725,7 291,8 727,0 291,6 728,1

16 299,8 735,4 301,5 750,1 301,7 745,6 297,6 724,6

17 303,9 758,0 301,2 753,9 301,7 756,0 303,1 761,6

18 270,3 696,3 272,5 699,9 274,3 702,8 274,3 708,7

19 288,5 731,8 286,5 726,5 287,2 732,6 286,2 730,4

20 300,2 779,1 298,9 800,2 300,8 797,2 301,0 808,2

Transform was used to convert a discrete signal from

the time domain to the frequency domain.

For each subject, eight EEG recordings corre-

sponding to eight different measurement conditions

(two blocks, each block containing four sessions)

were obtained. An analysis of the frequency spec-

trum of the EEG signal was done for these individ-

ual recordings. To obtain the values of the energy

level on the alpha and beta frequency bands several

preprocessing methods had to be performed with the

recorded EEG data: data ﬁltering (the basic FIR ﬁl-

ter was used) in the frequency range of 0,1 Hz –

30 Hz, automated (the moving window method was

used) and manual detection and removal of artifacts,

and calculation and visualization of the EEG signal

spectrum.

Visualization of the frequency spectrum and cal-

culation of energy levels in the alpha and beta bands

were processed for all subjects and for the Cz, Fz, and

Pz electrodes. As an example, the resulting graph for

subject 1 on the electrode Fz is shown in Figure 1.

Tables 2 and 3 show energy levels for all subjects

in the alpha and beta bands on the electrode Pz for the

ﬁrst and second block of the experiment. The alpha

and beta bands were the most prominent on the Pz

electrode, therefore these values were further used in

data analysis.

4.2 Heart Rate Data Processing

Nine different heart rate values were obtained for each

subject. The ﬁrst value corresponds to the baseline

heart rate, which was deﬁned as the median of the

values measured before and after the experiment. The

other heart rate values correspond to the eight sessions

of the experiment. The heart rate calculated for each

session is the median of all the values captured during

the session.

No heart values were obtained for subject 5 in the

last two parts of the experiment, and no heart rate val-

ues at all were obtained for subject 15. In both cases

it was due to technical troubles with the measuring

device.

4.3 Memory Game Data Processing

The data related to the accuracy of subjects in the

memory game were collected for each session and

each subject. The accuracy was calculated directly

by the application Pairs One. Two values were cal-

culated for each subject. The total accuracy for each

session of the experiment was calculated as the av-

erage of the accuracy of completed memory games in

this session, and the total number of exposed cards for

each session of the experiment was calculated as the

sum of exposed cards for all (even incomplete) mem-

ory games in that session. Table 4 shows the total

accuracy for all subjects in individual sessions.

HEALTHINF 2020 - 13th International Conference on Health Informatics

754

Table 3: Energy levels for all subjects in alpha and beta bands - the second block of the experiment.

Subject 5th session 6th session 7th session 8th session

α β α β α β α β

1 284,7 737,4 286,6 755,0 286,5 742,3 286,8 752,0

2 270,6 682,9 269,1 685,2 270,8 685,3 267,8 683,7

3 294,1 778,4 291,9 782,7 292,1 779,7 294,4 787,8

4 296,3 738,5 289,8 731,9 292,0 729,4 291,5 732,1

5 302,9 775,0 299,7 768,0 300,6 767,7 300,7 763,2

6 297,7 736,0 303,5 741,2 301,3 742,1 303,5 743,1

7 283,1 761,4 284,8 786,9 284,6 811,2 284,1 811,5

8 280,3 725,1 281,2 726,8 280,9 730,2 282,1 727,7

9 286,0 731,7 287,0 734,4 285,7 733,2 290,9 740,0

10 285,4 771,6 295,1 762,1 298,6 757,2 289,6 767,6

11 303,9 750,1 302,4 758,0 301,7 758,5 301,3 761,1

12 286,6 749,0 284,4 737,6 288,1 741,5 291,1 734,3

13 294,5 802,9 292,8 823,0 300,1 869,4 296,2 862,6

14 276,8 724,2 275,4 724,8 279,1 740,4 282,5 734,4

15 292,0 725,4 291,7 724,1 287,6 723,0 291,2 724,9

16 300,6 744,1 297,8 733,8 298,6 728,4 300,3 739,3

17 301,4 748,7 299,9 754,7 300,7 751,8 304,4 755,3

18 272,4 709,4 274,2 698,9 274,0 699,7 274,6 704,0

19 284,6 724,4 285,2 726,6 287,8 741,0 287,8 727,4

20 301,4 803,8 303,3 803,4 303,1 803,3 302,3 821,5

5 DATA ANALYSIS

This section describes the control measurement which

was performed to verify the effect of the order in

which the music compositions were played and selec-

tion of results of applying statistical methods to the

collected data.

5.1 Control Experiment

In the experiment, a control measurement was per-

formed to verify that the order in which the musical

compositions were played had/had not an impact on

the measured values. One of the subjects was asked

to undergo the experiment again; the order of the mu-

sical compositions was changed. In the beta band, an

increase in energy level values was observed during

the ﬁrst three sessions and a subsequent decrease of

energy level values was visible during the fourth and

ﬁfth sessions. In the rest of the experiment, there was

no such a trend visible. No dependence on the order

of sessions was observed in the alpha band and within

the heart rate values.

5.2 Statistical Results

Finally the collected and preprocessed data were sta-

tistically evaluated. The Repeated Measures ANOVA

was used for the evaluation of the heart rate and game

performance in eight different conditions (eight ses-

sions of playing the memory game) and the classic

t-test was applied to the data to answer the interesting

question given later in this paper.

Two following H

hypotheses were set for the

evaluation of heart rate and memory game perfor-

mance in the eight session of the experiment:

• The mean values of heart rate are the same under

all conditions (i.e. without statistically signiﬁcant

differences)

• The mean values of the number of cards played

are the same under all conditions (i.e. without sta-

tistically signiﬁcant differences).

The results for both hypotheses (at the signiﬁ-

cance level α = 0.05) showed that there was no statis-

tically signiﬁcant difference between the mean values

of heart rate as well as there was no statistically signif-

icant difference between cards played under speciﬁed

conditions.

For the next analysis of the data two variants of

the classic t-test were used. The following research

questions created the base for the subsequent deter-

mination of the hypotheses:

• What was the impact of the subjects’ sex on the

collected data?

• What was the impact of the game difﬁculty on the

Impact of Music on Human Brain Activity during Mental Stress

755

Table 4: The overall accuracy of playing the memory game in individual sessions of the experiment.

Subject Memory Game Accuracy (%) in sessions

1 2 3 4 5 6 7 8

1 37,0 44,0 34,0 36,0 44,0 34,0 36,0 39,5

2 56,5 49,0 47,0 39,0 42,0 36,0 65,0 38,0

3 53,0 54,0 52,0 57,0 54,0 59,5 47,0 68,0

4 59,0 59,0 56,0 66,0 69,0 46,0 60,0 55,0

5 40,0 37,0 52,0 49,5 52,0 35,0 37,0 55,0

6 59,0 72,5 72,0 68,5 64,5 72,0 69,0 53,0

7 65,0 65,0 68,0 61,5 63,0 59,0 57,5 57,5

8 58,0 36,0 49,0 49,5 50,0 61,0 60,5 59,5

9 67,0 68,5 72,5 60,5 72,5 72,5 72,5 62,5

10 54,0 45,0 52,0 40,0 - 50,0 50,0 52,0

11 77,5 66,0 68,0 70,5 74,0 79,0 81,5 68,0

12 68,0 56,0 73,0 71,0 61,0 59,0 72,0 71,0

13 44,0 44,5 31,0 48,0 35,0 33,0 33,0 58,0

14 63,0 72,5 53,0 60,5 43,0 44,0 78,0 65,5

15 61,0 50,0 58,0 45,0 64,0 45,0 61,0 56,0

16 62,5 54,0 54,5 64,0 61,5 51,0 59,0 68,0

17 44,0 59,0 40,0 38,0 50,0 47,5 42,0 47,0

18 56,5 48,0 41,0 44,5 53,0 29,0 42,0 26,0

19 63,5 52,5 55,0 45,5 52,0 37,0 45,0 44,0

20 41,0 50,5 38,0 36,5 49,0 50,0 37,0 46,0

collected data?

• What was the impact of the music played to the

subjects and music preferences of the subjects on

the collected data?

• What was the impact of music on the accuracy of

subjects during playing the memory game?

• What was the impact of the subjects’ interest in

the memory game on the collected data?

• What was the impact of listening to music on the

collected data?

Because the number of statistical results calcu-

lated over this data was very large, we focused only

on some of them in this paper.

The effect of subjects’ sex on the collected data

was evaluated for the spectral values in the alpha and

beta frequency bands as well as for the heart rate val-

ues. While the differences in heart rate values were

not statistically signiﬁcant for men and women, there

was a statistically signiﬁcant difference between the

values in the beta frequency band on the Pz electrode

(t = 2,4925; p = 0,0227) for men and women. It could

indicate that women concentrated more on their game

performance throughout the experiment.

No signiﬁcant changes in the alpha and beta

bands were observed when the difﬁculty of the game

changed. While the subjects were playing the mem-

ory game in a quiet environment, with slow and pre-

ferred music, a slight drop in heart rate values was

recorded (compared to a game with lower difﬁculty)

for the game with a higher difﬁculty setting. This

ﬁnding suggests that there might be a relationship be-

tween higher difﬁculty of the game and slowing of

the heart rate. It was not conﬁrmed that changing the

difﬁculty of the memory game would bring different

values of heart rate and in alpha and beta frequency

bands for men and women.

The values in alpha and beta frequency bands dif-

fered minimally when it came to the effect of the type

of music being played, only the values of heart rate

were seen to slightly increase when playing fast mu-

sic. For subjects who preferred fast music, the values

observed in the beta band were signiﬁcantly lower in

all parts of the measurement than in the group which

preferred slow music. Despite these differences, indi-

vidual values were not found to be statistically signif-

icant within each part of the experiment.

Whether the subject perceived the type of mu-

sic positively, neutrally or negatively did not affect

his/her achieved accuracy in the memory game as it

was expected.

Another question examined whether the alpha,

beta, and heart rate values differ between the group of

subjects who stated in the questionnaire that the game

was boring at the end of the experiment and the group

of the subjects who enjoyed the memory game. The

collected data were compared for these two groups of

subjects and for each session of the experiment.

HEALTHINF 2020 - 13th International Conference on Health Informatics

756

In all eight sessions of the experiment, the alpha

activity was signiﬁcantly higher in the group of peo-

ple who were bored at the end of the memory game.

In contrast, the beta activity was signiﬁcantly lower

in this group. The heart rate values were higher in

all sessions of the experiment for the group that en-

joyed the memory game. These results could indicate

that the subjects who enjoyed the memory game more

were also more focused and excited throughout the

experiment. Despite these signiﬁcant differences, in-

dividual values were not found to be statistically sig-

niﬁcant within each session of the experiment.

It was observed that the subjects’ experience of

listening to music at work did not signiﬁcantly affect

brain activity or heart rate during the experiment.

6 CONCLUSIONS

The work investigated the impact of listening to dif-

ferent types of music on the activity of the human

brain during mental stress (while the subjects were

playing the memory game). Changes in the values

of heart rate and brain activity in the alpha and beta

bands were measured and analyzed, and the relation-

ship of these changes to the music played and to

the data obtained from questionnaires were examined.

During the analysis the impact of music on the num-

ber of game cards exposed and the overall accuracy

during playing the memory game was also evaluated.

Performing the control experiments, the depen-

dence of the music on the order in which it was played

was not conﬁrmed.

The differences in heart rate values and the num-

ber of cards successfully played in the memory game

were not found to be statistically signiﬁcant in the in-

dividual sessions of the experiment. Even when we

statistically evaluated other hypotheses, most of the

results were statistically insigniﬁcant.

Since most of the results were found to be statisti-

cally not signiﬁcant, the question of the construction

of the experimental scenario itself arises. Given the

results of the analysis of groups with different pref-

erences in the selection of fast and slow music and

groups with different levels of engagement, it might

be beneﬁcial to examine these differences in more de-

tail in a separate experiment and to perform measure-

ments on more subjects.

ACKNOWLEDGEMENTS

This work was supported by the University speciﬁc

research project SGS-2019-018 Processing of hetero-

geneous data and its specialized applications (project

SGS-2019-018).

REFERENCES

Adamos, D. A., Dimitriadis, S. I., and Laskaris, N. A.

(2016). Towards the bio-personalization of music rec-

ommendation systems: A single-sensor eeg biomarker

of subjective music preference. Information Sciences,

343-344:94 – 108.

Bernardi, L., Porta, C., and Sleight, P. (2006). Cardio-

vascular, cerebrovascular, and respiratory changes in-

duced by different types of music in musicians and

non-musicians: the importance of silence. Heart,

92(4):445–452.

Daly, I., Malik, A., Hwang, F., Roesch, E., Weaver, J.,

Kirke, A., Williams, D., Miranda, E., and Nasuto, S. J.

(2014). Neural correlates of emotional responses to

music: An eeg study. Neuroscience Letters, 573:52 –

57.

Dolegui, A. S. (2013). The impact of listening to music

on cognitive performance. Inquiries Journal/Student

Pulse, 5(9).

Gorin, M. (2016). Pairs one.

Hadjidimitriou, S. K. and Hadjileontiadis, L. J. (2013). Eeg-

based classiﬁcation of music appraisal responses us-

ing time-frequency analysis and familiarity ratings.

IEEE Transactions on Affective Computing, 4(2):161–

172.

Kumagai, Y., Arvaneh, M., and Tanaka, T. (2017). Familiar-

ity affects entrainment of eeg in music listening. Fron-

tiers in Human Neuroscience, 11. cited By 4.

Lesiuk, T. (2005). The effect of music listening on work

performance. Psychology of Music, 33(2):173–191.

Impact of Music on Human Brain Activity during Mental Stress

757