1. Introduction
Wrist-worn devices within the cover of wearable technology attract the attention of individuals from all age groups and continue to gain popularity all over the world (Thompson, 2019). The global market for these devices in the United States is $13.865 billion and in the European Union, $12.744 billion. The annual growth rate (CAGR) for the next five years will be 12.8% in the United States and 14.2% in the European Union (Grand View Research, 2025). These devices are called wearable smart watches and mostly used for step count measurement in daily life. However, in recent years, it has also seen that they have come to the forefront due to their features such as measuring quality sleep duration, energy consumption, heart rate (HR) and continuous determination of oxygen saturation (Nissen et al., 2022; Hajj-Boutros et al., 2023). In the 2025 Fitness Trends reports, wearable technology and data-driven coaching are dominating the market, providing real-time data on activity, sleep, and recovery, enabling personalized training and goal setting. This data is increasingly integrated with fitness apps, smartwatches, and platforms, offering personalized training recommendations and tracking the trainer’s progress. They are also used to detect cardiac arrhythmias but are not validated as medical devices (Newsome et al., 2024). HR measurements are commonly monitored parameter in exercise/training planning and are important for knowing the physiological adaptations. However, the data obtained from these technological products are expected to show accuracy and consistency, and more or less than expected data such as HR may cause misguidance in training planning, adjusting training intensity, interruption or termination of training (Takacs et al., 2014; Kayabaş, Cuğ & Budak, 2022; Faqar Uz Zaman et al., 2025). In addition, accurate measurement and monitoring of devices play an important role in accessing accurate data in medical longitudinal studies, evaluating biophysiological parameters and making exercise prescription recommendations (Nissen et al., 2022; Prieto-Avalos et al., 2022; Hajj-Boutros et al., 2023).
Heart rate measurement is affected by different systems (Ruiz-Malagón et al. (2023) reported that the Photoplethysmography system was less sensitive in determining the measurement of HR. Photoplethysmography heart rate is less accurate, affected by skin pigmentation (Cabanas et al., 2022), altitude, movement and exercise, hemoglobin levels, sensor location, nail polish color, cardiac arrhythmia, low perfusion, hypoxemia, postoperative hypoxemia, or carboxyhemoglobin (León-Valladares et al., 2024). There have been previous studies conducted on activities with high upper limb movement, such as tennis and, therefore, participants’ HR was assessed with the Polar Precision Prime system on the Polar Ignite smartwatch. At the end of the study, it was reported to be a valid tool for monitoring HR during training (Ruiz-Malagón et al. 2023). However, few studies analyse the method of determining HR. Determining heart rate with a heart rate monitor is not the same as determining it with a pulse oximeter; they are different methodologies (Scalise & Cosoli, 2018). Chest strap pulse meters measure the volume changes generated by the heartbeat. But pulse oximeters obtain the HR measurement through the analysis of the photoplethysmogram, which is nothing more than the wave generated by the pulse in each heartbeat (Kim, Kim & Suh, 2022). This analysis has been seen to present many interferences in its reading, as occurs in high HR, already described by Iyriboz et al. (1991), dark skin, due to interferences of certain wavelengths with melanin, as occurs with carboxyhemoglobin (Seok et al., 2021).
Many manufacturers such as Apple, Samsung, Garmin and Fitbit continue to release new models (Hajj-Boutros et al., 2023). Although each brand or model has similar characteristics, there is no consistent regulation on the use or validity of the devices, since they are not medical devices (Benedetto et al., 2018; Darrow, Avorn & Kesselheim, 2021; Montes et al., 2020). In terms of validity and reliability, it is clear that most of the studies are based on the number of steps, the number of validity and reliability measurements related to HR is less and the results are inconsistent (Jo et al., 2016; Bai et al., 2018).
The validity and reliability of these devices is a prerequisite for future research and many studies of different brands/models were conducted in this field (Fuller et al., 2020; Karaca ve diğ., 2021; Nissen et al., 2022). For example, the Fitbit Charge HR series is one of the most frequently validated devices (Lee et al., 2016; Gorny et al., 2017). However, the lack of studies was seen on measurement tools such as Polar Ignite and Garmin (Wahl et al., 2017; Montes et al., 2020; Kastelic et al., 2021). Models such as the Polar Ignite and Garmin Vivoactive 4S are reported to use updated firmware versions that provide more advanced sensor architectures and different data-processing capabilities compared to previous generations (Budig ve diğ., 2021; Kastelic ve diğ., 2021; Nissen ve diğ., 2022). However, a gap has been identified in the literature regarding which device provides more accurate measurements in terms of sensor and software performance. Therefore, it is important to conduct appropriate validity studies by independent researchers. Accordingly, the aim of this study is to investigate the HR suitability of Polar Ignite and Garmin Vivoactive 4S devices during aerobic and resistance exercises. This study will be one of the first studies to validate Polar Ignite and Garmin Vivoactive 4S as far as we know. It was hypothesized that the different model of wearable devices may reveal the changeable scores based on the type of exercises.
2. Methods
2.1 Participants
A total of 29 participants (14 women and 15 men) (age: 22.03 ± 1.57 years, body weight: 68.52 ± 10.59 kg, height: 170.54 ± 8.18 cm, body fat percentage: 20.69 ± 9.36%) took part in this study voluntarily and without financial compensation. Inclusion criteria were: (1) aged between 18–26 years old (2) being recreationally active (3) having fair skin (4) having resistance exercises experience. Exclusion criteria were: (1) any chronic disease such as cardiovascular disease, diabetes, musculoskeletal injury (2) being outside the age range of 18–26, (3) having a tattoo where the devices are placed (4) sedentary person. A priori sample size estimation was performed using G*Power 3.1.9.7. Based on a repeated-measures ANOVA design, assuming a medium effect size (f = 0.25), an alpha level of 0.05, and a statistical power (1 – β) of 0.80, the minimum required sample size was calculated as 28 participants. Before the study, participants were informed related to the study procedures, and they provided a written informed consent. The study was conducted according to the principles stated in the Declaration of Helsinki, and it was approved by the Hitit University Non-Interventional Ethics Committee (Decision No:2023-04).
2.2 Study Design
This study is a validation study conducted in a cross-sectional design. All participants visited the laboratory 4 times among the study with separate days. Body weight, height and body composition measurements were obtained in the first visit. 10 repetition maximum for bench press (BP), shoulder press (SP), standing EZ-bar biceps curl (BC), leg extension (LE), and deadlift (DL) was determined in the second visit. All resting and walking and running HR measurements were obtained in the third visit. HR measurements were obtained during the resistance exercises at the fourth visit. Participants equipped the Polar H10 chest strap, Polar Ignite, and Garmin Vivoactive 4S during resting, walking, running and resistance exercise. Half of the participants equipped Polar Ignite on their right wrist and Garmin Vivoactive 4S on their left wrist. The other half of the participants equipped Garmin Vivoactive 4S on their right wrist and a Polar Ignite on their left wrist. Both devices were placed just above the wrist bone. Before the measurements, participants’ gender, height, body weight, date of birth, and wrist information on which they wore the device were entered into the devices. Polar H10 chest strap was used for criterion measurements of heart rate. The data obtained from Polar H10 transmitted to the Polar Beat mobile application. At the end of each measurement HR data was recorded from Polar Beat mobile application. The Polar Ignite and Garmin Vivoactive 4S were monitored simultaneously with the criterion device during all measurements.
2.3 Running and Resistance Exercises
HR measurements were obtained during the resting, walking, running and resistance exercises. The mean HR (HRmean) was recorded in resting and all exercises.
The resting heart rate (HRrest) was measured participants rested 10 min after arrived at the laboratory. Resting heart rate was measured for 10 min. After HRrest measurements participant performed walking and running on the treadmill at 5 km/h, 8 km/h, and 10 km/h respectively. Exercise at each speed lasted 5 minutes. 5-minutes rest was given to the participants between speed changes.
BP, SP, BC, LE, and DL were used in resistance exercise, respectively. All participants performed each exercise with 3 sets of 10 repetitions. Between the sets 1 min rest was given. Also, 2-minute rest was given between exercises. All equipment was paused during rest, so heart rate data was recorded only during exercise.
2.4 Statistical Analysis
All statistical analysis was performed by using SPSS 25 package (SPSS, Chicago, IL). Also, graphics were created in SPSS. The normal distribution of data was checked using the Shapiro-Wilk test. Polar H10 measurement was used as the criterion measurement. Mean absolute percentage error (MAPE) was calculated using a Microsoft Excel spreadsheet to compare both devices to the criterion measure. The following formula was used to calculate MAPE: the absolute value of [(HRdevice–HRcriterion)/(HRcriterion)] × 100. Significance level was set at 5% for MAPE. Bland-Altman analyses were performed to assess agreement between devices and criterion measurement. Bland–Altman analyses were conducted with the respective 95% limits of agreement. In addition, Intraclass Correlation Coefficient (ICC) analyses based on a two-way random-effects model with absolute agreement were conducted to assess agreement between the devices and the criterion measurement. In the process of obtaining data, all data were analyzed without missing data.
3. Results
The heart rate and mean absolute percentage error during resting, walking, and 2 different running speeds are given in Table 1. The lowest MAPE was observed on the Polar Ignite (2.12–2.74%) compared to the Garmin Vivoactive 4S (2.96–13.69%) for all activities. MAPE was <5% for all activities on the Polar Ignite, whereas on the Garmin Vivoactive 4S, it was <5% at rest only.
Table 1
The HRmean, MAPE, and ICC during resting, walking, and running at different speeds.
| TREADMILL SPEED | POLAR H10 | POLAR IGNITE | MAPE/ICC (IGNITE) | GARMIN | MAPE/ICC (GARMIN) |
|---|---|---|---|---|---|
| Resting | 74.34 ± 9.89 | 74.28 ± 10.12 | 0.74 ± 1.32 ICC: 0.995 | 74.76 ± 9.84 | 1.35 ± 1.68 ICC: 0.991 |
| 5 km/h | 109.24 ± 14.63 | 109.52 ± 13.56 | 2.12 ± 3.74 ICC: 0.957 | 93.52 ± 12.53 | 13.69 ± 13.74 ICC: 0.105 |
| 8 km/h | 142.97 ± 16.37 | 140.93 ± 15.68 | 2.25 ± 2.50 ICC: 0.948 | 122.00 ± 16.41 | 13.60 ± 16.30 ICC: –0.212 |
| 10 km/h | 156.72 ± 16.41 | 153.48 ± 17.88 | 2.74 ± 3.98 ICC: 0.913 | 141.07 ± 14.59 | 9.48 ± 11.52 ICC: 0.083 |
The mean heart rate and mean absolute percentage error during various resistance exercises are given in Table 2. MAPE was >5% for all resistance exercises on both devices. The Polar Ignite had a lower MAPE than Garmin Vivoactive 4S at SP and LE exercises. The Garmin Vivoactive 4S had a lower MAPE than Polar Ignite at BP, BC, and DL exercises. The highest MAPE was observed at BC exercises at both devices. Bland-Altman graphs are given in Figures 1 and 2.
Table 2
The HRmean, MAPE, and ICC during different resistance exercises.
| EXERCISE | POLAR H10 | POLAR IGNITE | MAPE/ICC (IGNITE) | GARMIN | MAPE/ICC (GARMIN) |
|---|---|---|---|---|---|
| BP | 103.83 ± 15.73 | 100.07 ± 13.80 | 5.81 ± 5.79 ICC: 0.816 | 103.55 ± 12.94 | 5.29 ± 4.03 ICC: 0.878 |
| SP | 110.03 ± 18.39 | 105.72 ± 16.99 | 5.39 ± 4.74 ICC: 0.898 | 104.28 ± 15.20 | 6.49 ± 4.74 ICC: 0.861 |
| BC | 124.90 ± 14.38 | 108.86 ± 17.26 | 13.21 ± 9.05 ICC: 0.447 | 108.41 ± 12.69 | 13.03 ± 5.49 ICC: 0.491 |
| LE | 112.55 ± 16.41 | 106.34 ± 17.37 | 6.59 ± 4.99 ICC: 0.852 | 105.28 ± 14.24 | 6.61 ± 3.92 ICC: 0.846 |
| DL | 133.07 ± 16.25 | 117.76 ± 20.05 | 12.36 ± 8.61 ICC: 0.527 | 120.79 ± 15.13 | 9.29 ± 5.30 ICC: 0.665 |
[i] BP: Bench Press, SP: Shoulder Press, BC: Standing EZ-Bar Biceps Curl, LE: Leg Extension, DL: Deadlift.
Figure 1
Bland–Altman analysis comparing heart rate measurements from the Polar Ignite and Garmin Vivoactive 4S with the Polar H10 during aerobic activities at different speeds and at rest. Outer dotted lines represent upper and lower limits of agreement (95% CI). A: Resting, B: Walking at 5 km/h speed, C: Running at 8 km/h speed, D: Running at 10 km/h speed.
Figure 2
Bland–Altman analysis comparing heart rate measurements from the Polar Ignite and Garmin Vivoactive 4S with the Polar H10 during various resistance exercises. Outer dotted lines represent upper and lower limits of agreement (95% CI). A: Bench Press, B: Shoulder Press, C: Standing EZ-Bar Biceps Curl, D: Leg Extension, E: Deadlift.
4. Discussion
The main goal of the present study was to assess by using the Polar H10 as the reference measurement the accuracy of the Polar Ignite and the Garmin Vivoactive 4S for heart rate during 2 different model activities of running and resistance exercise. The results in different situations such as resting, walking, and running indicated that similar MAPE values (2.48 ± 2.16 and 2.96 ± 2.54 respectively) were seen in resting for Polar ignite and the Garmin Vivoactive 4S, but in walking and different treadmill speeds, especially Garmin Vivoactive 4S showed higher MAPE values (9.48 ± 11.52–13.69 ± 13.74) according to the Polar ignite (2.12 ± 3.74–2.74 ± 3.98) (Table 1). Based on these results, we suggest that MAPE is under 5% for all activities on the Polar Ignite, whereas on the Garmin Vivoactive 4S, it is only during the rest. These results indicate that HR measurements on the device Polar Ignite, are the acceptable error range in resting, and running activities included at 5, 8, and 10 km/h, while the Garmin Vivoactive 4S is the highest, therefore these results may be taken into account for consumers who are interested in health and training. The findings from the previous studies related to the assessment of the accuracy of the wrist-worn monitoring devices reported that the Apple Watch (Shcherbina et al., 2017) for running and cycling exercises; the Apple Watch 6 and Polar Vantage V (Hajj-Boutros et al., 2023) for running and resistance exercises; the Apple Watch 4 and Polar Vantage V (Düking et al., 2020) for running; the Apple Watch and Tomtom runner Cardio (Martín-Escudero et al., 2023) during physical activity at different levels were the most accurate for heart rate measurement. In other studies, Thiebaud et al. (2018) researched the validity of wrist-worn using the TomTom Cardio, Microsoft Band, and Fitbit Surge at different running speeds on a treadmill at 3.2, 4.8, 6.4, 8, and 9.7 km/h. They reported that the heart rate changed from 2.17 to 8.06% for the Fitbit Surge, from 1.01 to 7.49% for the TomTom Cardio, and from 1.31 to 7.37% for the Microsoft Band with increasing intensity. These results exhibit that the accuracy of the devices may be influenced by the increasing intensity. Martín-Escudero et al. (2023), compared four wrist-worn devices, including the Fitbit Charge, Apple Watch, Tomtom Runner Cardio, and Samsung G2 testing on a cycle ergometer and a treadmill. Their results indicated that as exercise intensity increased there was a higher underestimation of HR across all devices. However, it demonstrated that the Apple Watch, Tomtom Runner Cardio showed the highest validity for monitoring HR during physical activity at different levels. In addition, when examining the literature studies and our findings, it indicates that the accuracy of the wrist-devices may be influenced by the higher intensity especially in the continuing cardio exercises, such as running, cycling etc. (Shcherbina et al., 2017; Boudreaux et al., 2018; Thiebaud et al., 2018; Düking et al., 2020; Martín-Escudero et al., 2023). The accuracy of smart watches in measuring HR varies greatly depending on the type of wearable of portable device (Pasadyn et al., 2019). There are multiple studies to validate HR measurement both between different devices and between different HR measurement methods (Tedesco et al., 2019; Stone et al., 2021; Alfonso et al., 2022). These studies have also been applied to the study of the accuracy of HR measurement in different situations, both at rest and in relation to different levels of physical activity, exercise or sport practiced (Dooley, Golaszewski & Bartholomew, 2017; Boudreaux et al., 2018; Xie et al., 2018; Jacko et al., 2024).
In the present study, the statistical analysis indicated that MAPE was higher than 5% according to the reference method for heart rate monitoring and absolute percentage error during all resistance activities for both devices. However, similar MAPE values for both devices were observed for all resistance exercises except the deadlift exercise (Table 2). The results revealed based on the reference value for 5% that none of the devices provided estimates of HR that were within acceptable ranges in resistance activities. Boudreaux et al. (2018) conducted a study using HR measures from wrist-worn wearable devices including Apple Watch Series 2 (AWS2), Fitbit Blaze, Fitbit Charge 2, Polar A360, Garmin Vivosmart HR, TomTom Touch at different exercise modes that consisted of a graded cycling exercise and resistance exercise. The authors reported that MAPE values were observed ranging from 1.44% to 9.97% at rest, and 5.47% to 21.20% after the last exercise in resistance exercise and revealed that more accurate results were seen at rest and lower exercise intensities than the higher intensities, similar for others researches.
5. Conclusion
In conclusion, when taking into account the reference value for 5%, the findings indicate that HR readings are typically more accurate on rest and the different running activities using the Polar Ignite, but the Garmin Vivoactive 4S was only while resting. However, the resistance exercises provided poor accuracy for both devices. The reason for this may be the arm blood flow is changeable while performing to the muscle motion. Because, in resistance exercises, these body parts usually involve movements such as flexion and extension while resistance performed, but running exercises involve more stable wrist and arm movements compared to the resistance exercise. Based on these findings obtained from the healthy, young, fair-skinned adults with normal body fat, it concludes that the Polar Ignite has a higher accuracy for measuring heart rate for just resting and running activities, whereas the Garmin Vivoactive 4S has poor accuracy for both running and resistance exercises.
Limitations
Some of limitations related to this study were given respectively: 1) The study was conducted just two wrist devices including Polar Ignite, Garmin Vivoactive 4S, and used the Polar H10 as the reference measurement. 2) It was performed two types of exercises including running on treadmill by different speeds and five different resistance exercises. 3) The study had a small sample size (n = 29), 4) HR was only used as a dependent variable. 4) In this study, the polar chest strap was used as the reference method because it is frequently used in the field for obtaining HR data. 5) The participants consisted entirely of healthy, young, fair-skinned adults with normal body fat persons.
Data Accessibility Statement
The datasets generated and/or analyzed during the current study are not publicly available due to privacy and ethical restrictions involving human participants, but are available from the corresponding author upon reasonable request.
Ethics and Consent
The study was conducted according to the principles stated in the Declaration of Helsinki, and it was approved by the Hitit University Non-Interventional Ethics Committee (Decision No:2023-04). Informed consent was obtained from all the participants involved in the study.
Author Contributions
EMA and SC contributed to the conception and design of the study, and EMA, AFS, and BG contributed to the acquisition of data. ED and SC contributed to the interpretation of data. EMA performed the data analysis. All authors participated in drafting the manuscript, and author PM revised it critically. All authors read and approved the final version of the manuscript.