Correlation between mood and heart rate variability indices during daily life

doi:10.4236/health.2011.39094

Paper Menu >>

Journal Menu >>

Vol.3, No.9, 553-556 (2011)

doi:10.4236/health.2011.39094

Health

Correlation between mood and heart rate variability

indices during daily life

Kohzoh Yoshino1*, Katsunori Matsuoka2

1Health Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Osaka, Japan;

*Corresponding Author: [email protected]

2Life Science and Biotechnology, National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, Japan.

Received 1 June 2011; revised 20 July 2011; accepted 17 August 2011.

ABSTRACT

We investigated the correlation between mood

and heart rate variability (HRV) indices during

daily life. The RR-interval and body acceleration

of 40 normal male subjects were recorded using

ambulatory device for 48 to 72 hours. Every

hour that the subjects were awake they regis-

tered their current mood on a Visual Analogue

Scale questionnaire. The questionnaire scales

eight of the subjects’ current moods. Those are

happiness, tension, fatigue, worry, depression,

anger, vigor, and confusion. The following four

HRV indices were calculated. Those are heart

rate, root mean square of successive differ-

ences of RR-interval sequence, the normalized

high-frequency (0.15 - 0.4 Hz) power of RR-in-

terval variability, and mean frequency in the

high-frequency band of RR-interval variability.

The calculated HRV indices data and the mood

data were normalized individually, the data with

body acceleration exceeding 30 mG were ex-

cluded from the analysis to reduce the effect of

exercise, and the differences from the first day

(∆mood and ∆HRV-index) were taken to reduce

the effect of circadian rhythm. The most three

highly correlated combinations were ∆vigor and

∆HFnu (R = –0.24, p < 0.0001), ∆vigor and

∆RMSSD (R = –0.24, p < 0.0001), and ∆vigor and

∆HR (R = 0.22, p < 0.001). Vigor exhibited the

most significant correlations with HRV indices

of eight moods.

Keywords: Mood; Autonomic Nervous System;

Ambulatory; Heart Rate; Emotion

1. INTRODUCTION

It is important for mental health management to de-

velop a method that can assess mood during daily life in

stressful modern society. Three methods have been pro-

posed from mental state assessment. The first method is

based on subjective self-evaluation. The second is based

on biochemical analysis that quantifies the concentration

of stress-related substrates in blood plasma, urine, and

saliva. The third method is based on variability analysis

of physiological signals reflecting the state of the auto-

nomic nervous system. The third method has the fol-

lowing advantages for mental state monitoring compared

with the other two. 1) It can be measured continuously

with a short sampling interval; 2) it does not interrupt

daily life activities; 3) it can compute (assess) the values

of indices automatically; 4) it is less affected by psycho-

logical bias than the subjective self-evaluation method.

Heart rate variability (HRV) is a representative

physiological signal reflecting states of the autonomic

nervous system that can be continuously measured dur-

ing daily life with a short time interval. In order to de-

velop a method that can assess mood states from HRV, it

is necessary to understand how mood and HRV during

daily life are correlated [1]. Several studies have inves-

tigated the relationship between stress or mood and heart

rate during daily life. Dobkin and Pihl [2] demonstrated

that the heart rate of normal subjects during daily life is

significantly higher when stressed than when not

stressed. They also revealed that anxiety and hostility

contributes to the heart rate elevation. Langewitz et al.

[3] found that the perceived stress level is significantly

correlated with heart rate in normal subjects. Johnston

and Anastasiades [4] reported that heart rate during daily

life is related to arousal, stress, and time pressure in very

few normal subjects. Shapiro et al. [5] reported no cor-

relation of mood with heart rate. Sloan et al. [6] showed

that the LF/HF ratio (sympathovagal balance index) is

significantly correlated with stress (irritable, tense,

pressured, and less happy) during daily life in normal

subjects. Yoshino and Matsuoka [7] found that both

daily depression and worry shift toward sympathetic

K. Yoshino et al. / Health 3 (2011) 553-55 6

554

dominance during subsequent sleep. Bacon et al. [8]

reported the significant association of high levels of

negative emotions (anger, stress, and sadness) with de-

creases in LF and HF powers, but higher levels of posi-

tive emotion with increases in LF power in patients with

coronary artery disease. Schwerdtfeger et al. [9] showed

that depression was related to higher heart rate through-

out the day. Moreover, there was a tendency toward

lower HRV in individuals with higher depression scores,

and this association was moderated by social context.

However, to our knowledge, no studies have investi-

gated the relationship between HRV indices and various

moods including vigor in normal subjects during daily

life. In this study, we analyzed the correlation between

four HRV indices and eight moods including vigor in

normal male subjects during daily life.

2. METHODS

Normal healthy male subjects (N = 40, age: 20 to 39

years) participated in this experiment. None of the sub-

jects had a history of cardiovascular, respiratory, or brain

disease and none were currently using any medications.

All subjects provided their informed written consent as

approved by the Ethical Committee on Human Research

at National Institute of Advanced Industrial Science and

Technology. Subjects intermittently wore an ambulatory

device Active tracer (AC-301, GMS, Japan) in a pouch

at the waist for a total of 48 to 72 hours during daily life,

including sleep, except when taking a bath or shower.

This device records the RR-intervals of the electrocar-

diogram and body acceleration.

Every hour that the subjects were awake they regis-

tered their current mood on a 100-mm-long Visual Ana-

logue Scale (VAS) questionnaire on which the end-

points were labeled “lowest” and “highest”. The ques-

tionnaire scales eight of the subjects’ current moods:

happiness, tension, fatigue, worry, depression, anger,

vigor, and confusion. We instructed subjects not to an-

swer the questionnaire more than 15 minutes after the

predetermined time, and we strictly prohibited subjects

from recalling a past mood when answering the ques-

tionnaire.

The following four HRV indices 1) to 4) were calcu-

lated from RR-interval sequence data collected during

10 min before answering the VAS questionnaires if at

least 96% of 10-min RR-interval records were within

valid range (0.3 - 2.0 sec). The RR-intervals out of this

range were omitted from the analysis. If the mean body

acceleration in the 10-min exceeded 30 mG, the corre-

sponding RR-interval data were excluded from the

analysis in order to exclude the effect of body movement.

It has been reported that 157.5 mG is the optimum value

of threshold to discriminate activity state and rest by

body acceleration [10]. We used the value 30 mG which

is five times smaller than it in this study. The inter-indi-

vidual mean excluded ratio was 47.2% ± 12.5%, and

there was no significant difference between before and

after noon.

1) Heart rate (HR) and 2) root mean square of succes-

sive differences (RMSSD) of RR-interval sequence were

calculated for each 10-min data set. HR takes low values

and RMSSD takes high values when the balance of

autonomic nervous system activity (sympathovagal bal-

ance) shifts toward parasympathetic dominance.

The following frequency analysis was performed in

addition to the time domain analysis. The power spec-

trum was calculated by applying a fast Fourier transform

(FFT) to the RR-interval signal for each 10-min data set.

The area (i.e., power) of high-frequency components

(HF: 0.15 to 0.4 Hz) of the power spectrum was divided

by the total power from which the very-low-frequency

component (VLF: <0.04 Hz) had been subtracted in or-

der to calculate (3) the normalized HF power (HFnu),

which takes higher values when the sympathovagal bal-

ance shifts toward parasympathetic dominance. 4) Mean

frequency in HF band of the power spectrum (MFHF)

were calculated. MFHF is related to respiratory fre-

quency.

To reduce inter-individual variability, we normalized

the measured values of mood level and HRV indices by

dividing by their amplitudes (maximum-minimum in

measurement) after subtracting their minimum values for

each subject. Moreover, to reduce the effect of circadian

rhythm, the differences of the normalized values of

mood level and HRVI indices on the second or third day

from those at the same time of the first (reference) day

(denoted as ∆mood and ∆HRVI) were calculated for

each subject.

Before applying the correlation analysis, we con-

firmed that the inter-individual differences of ∆mood

and ∆HRVI variations were not relatively high. This was

done by comparing the inter-individual means and stan-

dard deviations of ∆mood and ∆HRVI variations. Fol-

lowing the confirmation, Pearson’s correlation coeffi-

cients between eight ∆moods and four ∆HRVIs were

calculated.

3. RESULTS

The comparison results are summarized in Table 1.

The most three highly correlated combinations were

∆vigor and ∆HFnu (R = –0.24, p < 0.0001) (Figure 1

right), ∆vigor and ∆RMSSD (R = –0.24, p < 0.0001),

and ∆vigor and ∆HR (R = 0.22, p < 0.001) (Figure 1

left). There were three combinations whose p-values

were less than 0.001 (Table 1). All of them were com-

binations with ∆vigor. This indicates that vigor has the

K. Yoshino et al. / Health 3 (2011) 553-55 6

555555

Table 1. Pearson’s correlation coefficients between four ∆HRVIs and eight ∆moods.

Δhappiness Δtension Δfatigue Δworry Δdepression Δanger Δvigor Δconfusion

ΔHR 0.11 0.05 –0.02 0.05 0.05 0.01 0.22** 0.03

ΔRMSSD –0.19* –0.07 0.08 –0.03 0.00 –0.01 –0.24*** –0.05

ΔHFnu –0.03 –0.13 –0.07 –0.05 –0.04 –0.10 –0.24*** –0.10

ΔMFHF –0.12 –0.08 –0.11 –0.06 –0.01 0.01 –0.12 –0.00

*: p < 0.01, **: p < 0.001, ***: p < 0.0001.

Figure 1. Scatter plots of Δvigor versus ΔHR (left) and Δvigor versus ΔHFnu (right).

most significant correlations with HRV indices of eight

moods.

Openly accessible at

4. DISCUSSION

The significant negative correlations between ∆vigor

and ∆HFnu (R = –0.24, p < 0.0001) and between ∆vigor

and ∆RMSSD (R = –0.24, p < 0.0001) and the signifi-

cant positive correlation between ∆vigor and ∆HR (R =

0.22, p < 0.001) suggest that vigor during daily life shifts

the autonomic nervous system balance toward sympa-

thetic dominance.

Russell proposed a circumplex model that defines

emotion space by pleasure-displeasure (valence) axis

and arousal-sleep (arousal) axis [11]. Thayer decom-

posed the arousal axis into two dimensions based on the

results from factor analysis [12,13]. One is tense arousal

and the other is energetic arousal. The tense arousal

represents a continuum from calmness to anxiety,

whereas the energetic arousal reflects a continuum from

tiredness to energy. Dickman broke down the energetic

arousal still further into two dimensions of wakefulness

and vigor [14]. The result from the presented study im-

plies that autonomic nervous system (HRV) indices co-

variate with vigor the fourth dimension axis in emotion

space during daily life.

In our previous study, we demonstrated that both daily

depression and worry shift toward sympathetic domi-

nance during subsequent sleep, whereas vigor had little

effect on the autonomic nervous system balance during

subsequent sleep [7]. The results in this study and in our

previous study imply that vigor during daily life concur-

rently has a relatively strong effect on shifting the auto-

nomic nervous system balance toward sympathetic

dominance, but the effect is not sufficiently long-lasting

to make an effect during subsequent sleep.

There are several limitations pertain to this study.

First, the effects of circadian rhythm on mood and HRVI

have not been completely removed, although we took

the differences of the values of indices on the second or

third day from those at the same time of the first (refer-

ence) day. This is because of the day-to-day variation in

the circadian rhythm of mood and autonomic nervous

activity.

Second, the result of this study can be applied to only

half of the daily awaking time. This is because about

half (47.2%) of the case was excluded from the analysis,

since the mean body acceleration was over 30 mG.

In summary, we demonstrated that vigor during daily

K. Yoshino et al. / Health 3 (2011) 553-55 6

556

life has the most significant correlations with HRV indi-

ces (autonomic nervous system balance) of eight moods.

5. ACKNOWLEDGEMENTS

This work was supported by Grant-in-aid for Scientific Research

(KAKENHI) for Young Scientists (B) (No.19700557) from the Minis-

try of Education, Culture, Sports, Science, and Technology, Japan.

REFERENCES

[1] Kreibig, S.D. (2010) Autonomic nervous system activity

in emotion: A review. Biological Psychology, 84, 394-

421. doi:10.1016/j.biopsycho.2010.03.010

[2] Dobkin, P.L. and Pihl, R.O. (1992) Measurement of psy-

chological and heart rate reactivity to stress in the real

world. Psychotherapy and Psychosomatics, 58, 204-214.

doi:10.1159/000288629

[3] Langewitz, W., Ruddel, H. and Von Eiff, A.W. (1987)

Influence of perceived level of stress upon ambulatory

blood pressure, heart rate, and respiratory frequency.

Journal of Clinical Hypertension, 3, 743-748.

[4] Johnston, D.W. and Anastasiades, P. (1990) The rela-

tionship between heart rate and mood in real life. Journal

of Psychosomatic Research, 34, 21-27.

doi:10.1016/0022-3999(90)90004-N

[5] Shapiro, D., Jamner, L.D. and Goldstein, I.B. (1997)

Daily mood states and ambulatory blood pressure. Psy-

chophysiology, 34, 399-405.

doi:10.1111/j.1469-8986.1997.tb02383.x

[6] Sloan, R.P., Shapiro, P.A., Bagiella, E., Boni, S.M., Paik,

M., Bigger, J.T. Jr., Steinman, R.C. and Gorman, J.M.

(1994) Effect of mental stress throughout the day on car-

diac autonomic control. Biological Psychology, 37, 89-99.

doi:10.1016/0301-0511(94)90024-8

[7] Yoshino, K. and Matsuoka, K. (2009) Effect of mood

during daily life on autonomic nervous activity balance

during subsequent sleep. Autonomic Neuroscience, 150,

147-149. doi:10.1016/j.autneu.2009.03.013

[8] Bacon, S.L., Watkins, L.L., Babyak, M., Sherwood, A.,

Hayano, J., Hinderliter, A.L., Waugh, R. and Blumenthal,

J.A. (2004) Effects of daily stress on autonomic cardiac

control in patients with coronary artery disease. Ameri-

can Journal of Cardiology, 93, 1292-1294.

doi:10.1016/j.amjcard.2004.02.018

[9] Schwerdtfeger, A. and Friedrich-Mai, P. (2009) Social

interaction moderates the relationship between depres-

sive mood and heart rate variability: Evidence from an

ambulatory monitoring study. Health Psychology, 28,

501-509. doi:10.1037/a0014664

[10] Mathie, M.J., Coster, A.C.F., Lovell, N.H. and Celler,

B.G. (2003) Detection of daily physical activities using a

triaxial accelerometer. Medical & Biological Engineering

& Computing, 41 , 296-301. doi:10.1007/BF02348434

[11] Russell, J.A. (1979) Affective space is bipolar. Journal of

Personality and Social Psychology, 37, 345-356.

doi:10.1037/0022-3514.37.3.345

[12] Thayer, R.E. (1967) Measurement of activation through

self-report. Psychological Reports, 20, 663-678.

[13] Thayer, R.E. (1978) Factor analytic and reliability studies

on the Activation-Deactivation Adjective Checklist.

Psychological Reports, 42, 747-756.

[14] Dickman, S.J. (2002) Dimensions of arousal: Wakeful-

ness and vigor. Human Factors, 44, 429-442.

doi:10.1518/0018720024497673