Validation of the Polar Heart Rate Monitor for Assessing Heart Rate During Physical and Mental Stress

Abstract

The Polar Vantage XL heart rate monitor provides an ambulatory, inexpensive method of continuously measuring heart rate. To examine the validity of the Polar monitor for measuring heart rate during resting periods and while engaging in two stressful tasks, 30 students participated in a 1-hour laboratory session. Heart rates were measured simultaneously using the Polar monitor and electrocardiography (ECG) during a hand grip exercise and a mental arithmetic task, each preceded by a 4-min resting period. Within-subject correlations between the two devices were significant (mean r = 0.98, P < .001). All correlations, except for three participants, exceeded r = 0.90. Between-task correlation analyses revealed high correlations (i. e., rs ≥ 0.98) between the Polar monitor and ECG. The Polar monitor obtained readings that were slightly, though significantly higher than readings obtained using ECG. The correspondence between observed mean heart rates from the Polar monitor and ECG suggest that the Polar monitor provides a valid measure of heart rate during stationary laboratory tasks.

Full text

JeffreyL. Goodie et al.: Validation of the Polar Mon itorJOP 14 (3), 2000, Hogrefe & Huber Publishers
Validation of the Polar Heart Rate
Monitor for Assessing Heart Rate
During Physical and Mental Stress
Jeffrey L. Goodie, Kevin T. Larkin, and Scott Schauss
West Virginia University
Accepted for publication: January 22, 2000
Keywords: Polar monitor, heart rate, ambulatory, laboratory stress
Journal of Psychophysiology 14 (2000) 159–164 © 2000 Federation of European Psychophysiology Societies
Abstract The Polar Vantage XL heart rate monitor provides an ambulatory, inexpensive method of continuously measuring heart
rate. To examine the validity of the Polar monitor for measuring heart rate during resting periods and while engaging in two stressful
tasks, 30 students participated in a 1-hour laboratory session. Heart rates were measured simultaneously using the Polar monitor
and electrocardiography (ECG) during a hand grip exercise and a mental arithmetic task, each preceded by a 4-min resting period.
Within-subject correlations between the two devices were significant (mean r = 0.98, P < .001). All correlations, except for three
participants, exceeded r = 0.90. Between-task correlation analyses revealed high correlations (i. e., rs ≥ 0.98) between the Polar
monitor and ECG. The Polar monitor obtained readings that were slightly, though significantly higher than readings obtained using
ECG. The correspondence between observed mean heart rates from the Polar monitor and ECG suggest that the Polar monitor
provides a valid measure of heart rate during stationary laboratory tasks.
Continuous recording of heart rate (HR) allows research-
ers to observe moment-to-moment HR changes as indi-
viduals engage in a variety of activities (e. g., stressful
tasks). In laboratory settings, researchers have tradition-
ally used electrocardiography (ECG) to obtain these con-
tinuous HR measures. In studies of this type (e. g.,
Ravaja, Raeikkoenen, Lyytinen, Lehtimaeki, & Kelti-
kangas-Jaervinen, 1997; Voegele, 1998; Voegele, Jarvis,
& Cheeseman, 1997), measurement of ECG typically
employs equipment that is adequate for the laboratory
(e. g., computer-digitized ECG) but insufficient for mea-
surement of HR in ambulatory or non-laboratory set-
tings.
The Polar Vantage XL (Polar monitor) is an ambula-
tory HR monitor, marketed for monitoring HR during
bouts of exercise. It employs a transmitter worn around
the chest that transmits HR information to a receiver
worn on the wrist, which stores data until it is download-
ed to a computer. The Polar monitor may provide a rela-
tively inexpensive continuous measure of HR in ambu-
latory settings that cannot be obtained via standard lab-
oratory instrumentation, but there is a paucity of research
documenting the validity of the Polar monitor. Research-
ers have used the Polar monitor to obtain measures of HR
change during physical activities (Loftin, Anderson, Lyt-
ton, Pittman, & Warren, 1996) and have assessed the
validity of the Polar HR monitor during an exercise task
(Bar Or, Bar Or, Waters, Hirji, & Russell, 1996; Godsen,
Carroll, & Stone, 1991; Wajciechowski, Gayle, An-
drews, & Dintiman, 1991). Bar Or et al. reported that the
Polar monitor obtained equivalent HR measures to ECG
during rest and a bike exercise task with children, but
obtained significantly higher readings than ECG during
a recovery period. Godsen, Carroll, and Stone found that
the HR values obtained by the Polar monitor and ECG
were within 6 beats per min 95% of the time during
treadmill running, rowing, arm-leg cycle ergometry, and
weight training. Wajciechowski et al. found high corre-
lations (r = .99) between the Polar monitor and ECG
readings during aerobic dance, jogging, and walking.
Laukkanen and Virtanen (1998) provide a review of the
development of the Polar monitor and discuss studies
that have examined the reliability of the Polar monitor.
Absent from this literature, however, is an examination
of the ability of the Polar monitor to obtain valid HR
measures, compared to ECG during laboratory stressors
typically used in the literature to examine the relation
between stress and health.
JOP 14 (3), 2000, Hogrefe & Huber Publishers

The present study examined the ability of the Polar
monitor to obtain valid measures of HR change during
two laboratory stressors. A mental arithmetic challenge
was chosen because it is one of the most common mental
challenges used in this literature and has been widely
accepted as a task evoking a substantial HR response
(e. g., Manuck & Garland, 1980; Matthews et al., 1987).
A hand grip challenge was chosen as the second task
because it represents a combined physical-mental chal-
lenge and has also been widely used as a method to evoke
HR and blood-pressure increases among participants
(e. g., Faultisch et al., 1986; Ditto, France, & Miller,
1989). The HR measures obtained by the Polar monitor
were compared with values simultaneously obtained us-
ing ECG during these two stress challenges preceded by
two resting periods. Both within- and between-subject
analyses were conducted to determine the ability of the
Polar monitor to track HR fluctuations within each indi-
vidual participant as well as across all participants.
Method
Subjects
Participants included 30 West Virginia University stu-
dents (53% female, 80% Caucasian) between 18 to 48
years old (M = 23 years, SD = 6.5) and a mean body mass
index of 23.4 kg/m
2
(SD = 3.1). Most individuals (93%)
reported that they were nonsmokers. Individuals were
not taking any cardiovascular reactive medications (e. g.,
decongestants, sympathomimetics). Seventeen partici-
pants (57%) reported a family history of hypertension or
cardiovascular disease, and 24 participants reported that
they exercised on a regular basis. Individuals received
extra-credit in their undergraduate psychology courses
for their participation.
Measures
Electrocardiograph
Three silver silver-chloride ECG electrodes, attached bi-
laterally below the collar-bone, detected the electrocar-
diographic signals. The cardiac electrical impulses de-
tected by these electrodes were transmitted to an ECG
preamplifier and charted on a Grass Medical Instruments
(USA) Model 7 polygraph. The R-wave peaks within
each minute of the study were counted and recorded.
Polar Monitor
The Polar Vantage XL (Polar CIC, USA) HR monitor
uses a combined sensor and transmitter that is worn
across the chest with an elastic band. Electrodes embed-
ded in the sensor/transmitter detect cardiac electrical
impulses. The detection of these signals is transmitted
to the receiver that calculates the interbeat interval be-
tween these signals and converts the interbeat interval
to a HR reading, which is then displayed on the receiver.
The sensor/transmitter was positioned across the ster-
num, above the xyphoid process. HR readings were
transmitted to the receiver worn on the nondominant
wrist of the participant. The receiver was placed on the
nondominant wrist to avoid interference that might re-
sult during the hand grip task, which was performed
using the dominant hand. The Polar monitor allows the
researcher to sample HR every 5, 10, 15 or 60 seconds
(i. e., the interbeat interval at the specified epoch is re-
corded). For this study the Polar monitor was pro-
grammed to sample HR values every 5 seconds to max-
imize the correspondence with the ECG. The HR data
was stored in the receiver until it was downloaded to an
IBM-compatible computer using the Polar computer in-
terface unit and Polar Vantage XL HR Analysis Soft-
ware (Polar CIC, USA).
Procedure
Participants were scheduled for one 1-hour laboratory
session. Each individual was informed about the purpose
and procedures of the study and signed an approved in-
formed consent form. Individuals were interviewed to
obtain demographic information and height and weight
were measured. Recording devices were attached as out-
lined above and the participant was positioned in a com-
fortable arm-chair in a laboratory chamber. Each partic-
ipant sat quietly for a 10-minute habituation period.
Heart rate data was not collected during the habituation
phase. Standardized instructions, lasting 20 seconds dur-
ing baseline periods and approximately 60 seconds dur-
ing task periods, were presented prior to the commence-
ment of each period. Task order was not counterbalanced
across participants because the primary purpose of this
study was not to examine the influence of these stressors
on HR responses.
Baseline
During the 4-minute baseline periods immediately pre-
ceding each task, participants were asked to sit quietly,
avoid moving, and keep their hands resting on the chair
and feet flat on the floor. The first baseline immediately
followed the 10-minute habituation period. HR measures
were recorded continuously throughout each 4-minute
resting period.
160 Jeffrey L. Goodie et al.: Validation of the Polar Monitor
JOP 14 (3), 2000, Hogrefe & Huber Publishers

Hand Grip Task
Prior to the instructions for the hand grip task, partici-
pants were asked to grip a hand dynamometer (Lafayette,
USA, Model 78010) with their full strength. An adjust-
able indicator on the hand dynamometer was positioned
at 30% of the individual’s full grip strength. During the
3-min task, participants were asked to maintain a grip on
the hand dynamometer at 30% of their full strength (i. e.,
maintaining the current grip strength indicator with the
indicator marked at the 30% level). HR measures were
recorded continuously during the 3-minute task.
Mental Arithmetic
During the mental-arithmetic challenge, participants
were asked to count backwards by 17 from a four-digit
number that was provided to them. The task lasted 3
minutes and a different four-digit number was provided
at the beginning of each minute. Participants were asked
to work as quickly and accurately as possible. HR mea-
sures were recorded continuously throughout the task.
Data Analyses
HR from the ECG was determined by manually counting
the number of R-wave deflections during each minute for
each baseline and task period. Polar monitor HR values
for each minute were determined by calculating the mean
of the 12 HR determinations obtained during each
minute of each baseline and task period.
To examine the validity of the Polar monitor for as-
sessing intra-individual HR fluctuations, a correlation
coefficient was calculated for each individual. These
Table 1 Participant mean HRs during rest periods and tasks for ECG and the polar monitor.
Rest 1 Hand Grip Rest 2 Mental Arithmetic
Subj. # Age Sex BMI ECG Polar ECG Polar ECG Polar ECG Polar M diff
a
r
b
1 22 female 16.6 73.8 73.9 86.7 86.2 70.5 70.3 80.3 80.1 0.18 0.99
2 21 male 22.7 78.8 79.4 96.3 96.4 77.8 79.8 85.3 86.1 –0.89 0.96
3 23 female 22.0 82.8 83.1 92.0 96.5 81.8 82.2 85.3 85.0 –1.25 0.98
4 25 male 23.4 64.0 64.4 76.0 74.9 64.8 64.6 66.7 66.3 0.30 0.99
5 21 female 19.3 93.5 93.4 99.0 98.4 88.8 88.5 96.7 96.9 0.16 0.99
6 25 male 24.4 76.0 76.0 78.0 77.4 73.0 73.1 80.3 80.2 0.18 0.98
7 22 male 22.3 53.5 53.9 53.7 54.4 51.8 51.8 62.7 63.1 –0.39 0.98
8 25 male 24.5 71.0 72.6 77.3 79.1 70.0 70.4 84.3 86.0 –1.37 0.96
9 48 female 21.8 74.3 74.2 76.0 77.4 69.8 69.8 71.7 71.4 –0.30 0.86
10 21 female 22.1 63.8 67.3 70.3 71.6 64.5 67.2 73.0 75.6 –2.54 0.91
11 26 female 25.7 79.3 80.1 84.0 84.1 79.8 80.2 80.3 83.4 –1.14 0.99
12 19 female 23.1 73.0 73.3 89.0 87.3 73.5 73.4 84.3 84.1 0.44 0.98
13 21 female 20.8 98.0 98.4 104.0 103.8 92.3 92.7 110.0 111.9 –0.62 0.99
14 19 female 22.5 78.5 78.5 86.7 82.8 69.8 70.5 90.7 91.1 0.69 0.96
15 25 female 21.3 69.8 71.4 71.0 72.4 67.5 68.8 68.3 69.3 –1.33 0.93
16 18 male 25.1 68.8 68.7 77.7 77.8 67.5 68.0 75.7 75.6 –0.14 0.99
17 25 male 27.1 80.0 80.5 84.3 84.2 81.0 80.9 79.3 79.7 –0.18 0.95
18 19 male 31.5 73.8 73.7 81.3 81.3 75.8 75.6 79.3 79.0 0.16 0.99
19 26 female 16.8 101.0 101.1 100.0 100.7 96.5 97.6 96.7 95.1 –0.08 0.85
20 25 female 19.9 71.3 71.0 72.3 71.7 68.3 68.0 69.0 73.2 –0.79 0.75
21 42 female 27.5 93.8 93.7 95.3 95.4 92.5 92.2 92.7 92.4 0.12 0.95
22 21 female 23.2 62.8 62.5 62.0 63.5 59.3 59.3 74.0 73.8 –0.27 0.98
23 18 male 24.3 60.0 59.9 73.7 73.8 61.5 61.7 76.0 76.3 –0.15 1.00
24 20 female 22.2 85.0 84.8 102.3 102.5 75.5 76.1 92.7 92.8 –0.19 1.00
25 19 male 22.2 68.8 68.6 84.0 83.8 67.5 68.0 87.7 90.0 –0.61 0.98
26 19 male 23.7 67.8 67.8 79.3 79.3 70.0 70.0 79.7 79.5 0.04 0.99
27 26 male 25.1 65.0 65.0 66.7 65.9 62.3 61.8 69.0 69.7 0.14 0.97
28 19 male 25.1 69.5 69.7 93.0 92.5 65.5 65.4 79.0 80.2 –0.17 1.00
29 18 male 28.5 66.5 66.8 73.3 74.2 65.3 65.4 71.0 71.9 –0.57 0.98
30 21 male 25.9 71.8 71.7 74.3 74.3 71.8 71.9 78.7 78.9 –0.10 0.99
M 23 23.4 74.5 74.8 82.0 82.1 72.5 72.8 80.7 81.3 –0.36 0.98
SD 6.5 3.1 11.2 11.1 12.2 12.1 10.3 10.3 10.4 10.4 0.67 0.05
Notes. BMI = body mass index (kg/m
2
); ECG = heart rate readings obtained using electrocardiograph; Polar = heart rate readings
obtained using Polar Monitor. Age is in years. All heart rate values are in beats per minute.
a
The mean heart rate across the baseline
periods and tasks was calculated for the ECG and Polar monitor then the difference between Polar monitor and ECG means were
calculated (ECG – Polar).
b
Within-subject correlations.
Jeffrey L. Goodie et al.: Validation of the Polar Monitor 161
JOP 14 (3), 2000, Hogrefe & Huber Publishers

within-subject correlation coefficients between the ECG
and Polar monitor readings were based on 14 observa-
tions (i. e., 8 baseline, 3 hand-grip, 3 mental arithmetic)
for each participant. Validity of the Polar monitor was
also examined using between-subject correlation analy-
ses. HR measures from the ECG and Polar monitor were
averaged within each baseline and task period, yielding
four HR values for each monitor per participant (rest
period 1, hand grip, rest period 2, mental arithmetic). The
HR values from the ECG and Polar monitor were com-
pared using a 2 × 4 [
HR MEASURE (Polar, ECG) × PERIOD
(Rest Period 1, Hand Grip, Rest Period 2, Mental Arith-
metic)] repeated measures ANOVA; significant effects
were analyzed using paired sample t-tests.
Results
Within-Subject Comparisons
Correlations between the Polar and ECG HR measures
ranged from r = 0.75 to r = 1.00. Three individuals dem-
onstrated correlation coefficients less than r = 0.90. On
average, ECG and Polar monitor readings were highly
correlated (M = 0.98, SD = 0.05; see Table 1). The mean
ECG and Polar HR readings for each individual are pre-
sented in Table 1.
Between-Task Comparisons
HRs from the ECG and Polar monitor were highly cor-
related during the first (r = 0.997, P < .0001) and second
rest periods (r = 0.987, P < .0001), as well as during the
mental arithmetic (r = 0.980, P < .0001) and hand-grip
tasks (r = 0.995, P < .0001). The mean HRs during each
baseline and task period are presented in Table 1.
The repeated measures ANOVA revealed a significant
main effect for the
HR MEASURE (F(1, 29) = 8.44, P < .01)
and
PERIOD (F(3, 87) = 37.59, P < .001). The HR MEASURE
× PERIOD interaction was not significant. The ECG ob-
tained significantly lower HR levels (M = 77.4 bpm, SD
= 11.63 bpm) compared to the Polar monitor (M =
77.8 bpm, SD = 11.6). The average difference between
70
72
74
76
78
80
82
84
Baseline 1 Hand Grip Baseline 2 Mental Arithmetic
Tasks
ECG
Polar
Figure 1 Mean ECG and Polar heart rate across rest and task periods.
162 Jeffrey L. Goodie et al.: Validation of the Polar Monitor
JOP 14 (3), 2000, Hogrefe & Huber Publishers

the HR monitoring methods was M = –0.4 bpm (SD =
0.7 bpm) with a range between –2.5 to 0.7 bpm for indi-
vidual participants (see Table 1). The relationship be-
tween the Polar monitor and ECG measurements are
plotted in Figure 1.
Regarding follow-up mean comparisons between the
rest periods and tasks, HR during rest period 1 was sig-
nificantly higher (M = 74.7, SD = 11.1) than the HR level
observed during rest period 2 (M = 72.7, SD = 10.3; P <
.001). HR during the hand-grip challenge (M = 82.1, SD
= 12.2) was not significantly different from the HR levels
observed during the mental arithmetic task (M = 81.0, SD
= 10.4). Both challenges yielded significantly higher HR
levels than both rest periods (Ps < .001).
Discussion
The Polar monitor yielded HR values that closely corre-
sponded to the readings obtained using ECG, suggesting
that the Polar monitor provided readings that are valid
for tracking HR changes during laboratory stressors. Re-
garding absolute HR values, the Polar monitor provided
readings that were slightly, although significantly, higher
than the readings obtained using ECG. However, be-
cause the average HR difference between the two meth-
ods was 0.4 bpm, it is unlikely that these differences rep-
resent differences that would be considered clinically
significant. The significant difference is likely related to
the different strategies of sampling employed for each
HR measurement method. Although calculation of HR
via ECG was done by counting R-wave deflections con-
tinuously for each minute, the Polar monitor sampled HR
at 5-second intervals, which were later averaged for each
minute. It is possible that the sampling method used by
the Polar monitor may introduce a systematic, albeit
small, overestimation of HR. Nevertheless, we observed
this phenomenon in this data set, and it has been reported
previously as well (Bar Or et al., 1996).
Participants demonstrated HR responses during the
baseline and task periods that were consistent with pre-
vious studies that have used similar tasks (e. g., Ditto,
France, & Miller, 1989). Three individuals demonstrated
correlations between the Polar and ECG readings that
were lower than r = .90. It is interesting to note that these
individuals were all women and were older than the
mean age (M = 21 years) of the sample. More important-
ly, these individuals demonstrated less than a 2 bpm HR
response to both tasks, compared to the sample average
of an 8 bpm HR response to the tasks. The limited HR
variability associated with these three participants likely
contributed to the lower correlations between the Polar
and ECG readings.
Overall, these results are consistent with studies that
have found the Polar monitor readings to be highly cor-
related with ECG readings during exercise and physical
tasks (Godsen, Carroll, & Stone, 1991; Wajciechowski
et al., 1991). These results suggest that the Polar monitor
is appropriate for monitoring HR responses to stationary,
laboratory tasks, with both mental and physical stress
features, among healthy adults. Future studies should
address the appropriateness of the Polar monitor for as-
sessing HR changes in ambulatory settings or among
individuals who may demonstrate health problems that
may affect HR detection accuracy (e. g., obesity, heart
disease), which was not addressed in the current study.
The validity of the Polar monitor to determine HR, in
conjunction with its lightweight and portable design,
suggest that it is a useful device for obtaining continuous
measures of HR in the laboratory on healthy individuals.
Authors’ Note
The authors thank Brandon Aragona for his assistance
with running subjects and scoring data.
References
Bar Or, T., Bar Or, O., Waters, H., Hirji, A., & Russell, S. (1996).
Validity and social acceptability of the Polar Vantage XL for
measuring heart rate in preschoolers. Pediatric Exercise Sci-
ence, 8, 115–121.
Ditto, B., France, C., & Miller, S. (1989). Spouse and parent off-
spring similarities in cardiovascular response to mental arith-
metic and isometric hand-grip. Health Psychology, 8, 159–173.
Faultisch, M.E., Williamson, D.A., McKenzie, S.J., Dutchman,
E.G., Hutchinson, K.M., & Bouin, D.C. (1986). Temporal sta-
bility of psychophysiological responding: A comparative anal-
ysis of mental and physical stressors. International Journal of
Neuroscience, 30, 65–72.
Godsen, R., Carroll, T., & Stone, S. (1991). How well does Polar
Vantage XL Heart Rate Monitor estimate actual heart rate?
Medicine and Science in Sports and Exercise, 26, 1041–1046.
Laukkanen, R. T., & Virtanen, P. K. (1998). Heart rate monitors:
State of the art. Journal of Sports-Sciences, 16, S3–S7.
Loftin, M., Anderson, P., Lytton, L., Pittman, P., & Warren, B.
(1996). Heart rate response during handball singles match-play
and selected physical fitness components of experienced male
handball players. Journal of Sports Medicine and Physical Fit-
ness, 36, 95–99.
Manuck, S.B., & Garland, F.N. (1980). Stability of individual
differences in cardiovascular reactivity: A13-month follow-up.
Physiology and Behavior, 24, 621–624.
Ravaja, N., Raeikkoenen, K., Lyytinen, H., Lehtimaeki, T., &
Keltikangas-Jaervinen, L. (1997). Apolipoprotein E pheno-
types and cardiovascular responses to experimentally induced
mental stress in adolescent boys. Journal of Behavioral Medi-
cine, 20, 571–587.
Voegele, C. (1998) Serum lipid concentrations, hostility and car-
Jeffrey L. Goodie et al.: Validation of the Polar Monitor 163
JOP 14 (3), 2000, Hogrefe & Huber Publishers

diovascular reactions to mental stress. International Journal of
Psychophysiology, 28, 167–179.
Voegele, C., Jarvis, A., & Cheeseman, K. (1997). Anger suppres-
sion, reactivity, and hypertension risk: Gender makes a differ-
ence. Annals of Behavioral Medicine, 19, 61–69.
Wajciechowski, J., Gayle, R., Andrews, R., & Dintiman, G.
(1991). The accuracy of radio telemetry heart rate monitor
during exercise. Clinical Kinesiology, 45, 9–12.
Jeffrey L. Goodie
P.O. Box 6040
Department of Psychology
West Virginia University
Morgantown, WV 26506-6040
USA
Tel. +1 304 293-2001 x859
Fax +1 304 293-6606
E-mail jgoodie@wvu.edu
164 Jeffrey L. Goodie et al.: Validation of the Polar Monitor
JOP 14 (3), 2000, Hogrefe & Huber Publishers