Heart Rate Variability in Professional and Semiprofessional Soccer: A Scoping Review.

In modern soccer, fitness and fatigue monitoring tools tend to be focused on noninvasive, time-efficient and player-friendly measures. Heart rate variability (HRV) has been suggested as an effective method for monitoring training response and readiness to perform. However, there is still a lack of consensus on HRV monitoring when it comes to soccer. Thus, this scoping review aims to map existing evidence on HRV in professional and semiprofessional soccer settings, and to identify knowledge gaps to inform future research directions. A search of databases (PubMed, Scopus, Web of Science, Google Scholar) according to the PRISMA-ScR statement was employed. Studies were screened for eligibility on inclusion criteria: (1) HRV was among the topics discussed in the article; (2) adult professional or semiprofessional soccer players were involved in the study; (3) both male and female participants; (4) no geographical area exclusion; (5) articles published in English; and (6) article full text available. The search of the selected databases revealed 8456 records. The titles and abstracts of all articles were retrieved for screening of eligibility, leaving 30 articles for further consideration. Following screening against set criteria, a total of 25 studies were included in this review, the sample size of which ranged from 6 to 124 participants. The participants in the included studies were professional and semiprofessional soccer players, interviewed clubs staff, and practitioners. Along with other monitoring strategies, morning vagally mediated HRV analysis via (ultra)short-term orthostatic measurements may be an efficient way to assess training adaptations and readiness to perform in professional and semiprofessional soccer players. Further research is required to make definitive recommendations.

What Is behind Changes in Resting Heart Rate and Heart Rate Variability? A Large-Scale Analysis of Longitudinal Measurements Acquired in Free-Living.

The aim of this study was to investigate the relationship between heart rate and heart rate variability (HRV) with respect to individual characteristics and acute stressors. In particular, the relationship between heart rate, HRV, age, sex, body mass index (BMI), and physical activity level was analyzed cross-sectionally in a large sample of 28,175 individuals. Additionally, the change in heart rate and HRV in response to common acute stressors such as training of different intensities, alcohol intake, the menstrual cycle, and sickness was analyzed longitudinally. Acute stressors were analyzed over a period of 5 years for a total of 9 million measurements (320±374 measurements per person). HRV at the population level reduced with age (p < 0.05, r = -0.35, effect size = moderate) and was weakly associated with physical activity level (p < 0.05, r = 0.21, effect size = small) and not associated with sex (p = 0.35, d = 0.02, effect size = negligible). Heart rate was moderately associated with physical activity level (p < 0.05, r = 0.30, effect size = moderate) and sex (p < 0.05, d = 0.63, effect size = moderate) but not with age (p = 0.35, r = -0.01). Similar relationships between BMI, resting heart rate (p < 0.05, r = 0.19, effect size = small), and HRV (p < 0.05, r = -0.10, effect size = small) are shown. In response to acute stressors, we report a 4.6% change in HRV (p < 0.05, d = 0.36, effect size = small) and a 1.3% change in heart rate (p < 0.05, d = 0.38, effect size = small) in response to training, a 6% increase in heart rate (p < 0.05, d = 0.97, effect size = large) and a 12% reduction in HRV (p < 0.05, d = 0.55, effect size = moderate) after high alcohol intake, a 1.6% change in heart rate (p < 0.05, d = 1.41, effect size = large) and a 3.2% change in HRV (p < 0.05, d = 0.80, effect size = large) between the follicular and luteal phases of the menstrual cycle, and a 6% increase in heart rate (p < 0.05, d = 0.97, effect size = large) and 10% reduction in HRV (p < 0.05, d = 0.47, effect size = moderate) during sickness. Acute stressors analysis revealed how HRV is a more sensitive but not specific marker of stress. In conclusion, a short resting heart rate and HRV measurement upon waking using a smartphone app can effectively be used in free-living to quantify individual stress responses across a large range of individuals and stressors.

The Promise of Sleep: A Multi-Sensor Approach for Accurate Sleep Stage Detection Using the Oura Ring.

Consumer-grade sleep trackers represent a promising tool for large scale studies and health management. However, the potential and limitations of these devices remain less well quantified. Addressing this issue, we aim at providing a comprehensive analysis of the impact of accelerometer, autonomic nervous system (ANS)-mediated peripheral signals, and circadian features for sleep stage detection on a large dataset. Four hundred and forty nights from 106 individuals, for a total of 3444 h of combined polysomnography (PSG) and physiological data from a wearable ring, were acquired. Features were extracted to investigate the relative impact of different data streams on 2-stage (sleep and wake) and 4-stage classification accuracy (light NREM sleep, deep NREM sleep, REM sleep, and wake). Machine learning models were evaluated using a 5-fold cross-validation and a standardized framework for sleep stage classification assessment. Accuracy for 2-stage detection (sleep, wake) was 94% for a simple accelerometer-based model and 96% for a full model that included ANS-derived and circadian features. Accuracy for 4-stage detection was 57% for the accelerometer-based model and 79% when including ANS-derived and circadian features. Combining the compact form factor of a finger ring, multidimensional biometric sensory streams, and machine learning, high accuracy wake-sleep detection and sleep staging can be accomplished.

Heart Rate Variability is a Moderating Factor in the Workload-Injury Relationship of Competitive CrossFit™ Athletes.

Heart rate variability (HRV) is a popular tool for monitoring training adaptation and readiness in athletes, but it also has the potential to indicate early signs of somatic tissue overload prior to the onset of pain or fully developed injury. Therefore, the aim of this study was to investigate the interaction between HRV, workloads, and risk of overuse problems in competitive CrossFit™ athletes. Daily resting HRV and workloads (duration × session-RPE) were recorded in six competitive CrossFit™ athletes across a 16 week period. The Oslo Sports Trauma Research Center Overuse Injury Questionnaire was distributed weekly by e-mail. Acute-to-chronic workload ratios (ACWR) and the rolling 7-day average of the natural logarithm of the square root of the mean sum of the squared differences between R-R intervals (Ln rMSSDweek) were parsed into tertiles (low, moderate/normal, and high) based on within-individual z-scores. The interaction between Ln rMSSDweek and ACWR on overuse injury risk in the subsequent week was assessed using a generalized linear mixed-effects model and magnitude-based inferences. The risk of overuse problems was substantially increased when a 'low' Ln rMSSDweek was seen in combination with a 'high' ACWR (relative risk [RR]: 2.61, 90% CI: 1.38 - 4.93). In contrast, high ACWRs were well-tolerated when Ln rMSSDweek remained 'normal' or was 'high'. Monitoring HRV trends alongside workloads may provide useful information on an athlete's emerging global pattern to loading. HRV monitoring may therefore be used by practitioners to adjust and individualise training load prescriptions, in order to minimise overuse injury risk.

Personalized cardiorespiratory fitness and energy expenditure estimation using hierarchical Bayesian models.

Accurate estimation of energy expenditure (EE) and cardiorespiratory fitness (CRF) is a key element in determining the causal relation between aspects of human behavior related to physical activity and health. In this paper we estimate CRF without requiring laboratory protocols and personalize energy expenditure (EE) estimation models that rely on heart rate data, using CRF. CRF influences the relation between heart rate and EE. Thus, EE estimation based on heart rate typically requires individual calibration. Our modeling technique relies on a hierarchical approach using Bayesian modeling for both CRF and EE estimation models. By including CRF level in a hierarchical Bayesian model, we avoid the need for individual calibration or explicit heart rate normalization since CRF accounts for the different relation between heart rate and EE in different individuals. Our method first estimates CRF level from heart rate during low intensity activities of daily living, showing that CRF can be determined without specific protocols. Reference VO2max and EE were collected on a sample of 32 participants with varying CRF level. CRF estimation error could be reduced up to 27.0% compared to other models. Secondly, we show that including CRF as a group level predictor in a hierarchical model for EE estimation accounts for the relation between CRF, heart rate and EE. Thus, reducing EE estimation error by 18.2% on average. Our results provide evidence that hierarchical modeling is a promising technique for generalized CRF estimation from activities of daily living and personalized EE estimation.

State of the science and recommendations for using wearable technology in sleep and circadian research.

Wearable sleep-tracking technology is of growing use in the sleep and circadian fields, including for applications across other disciplines, inclusive of a variety of disease states. Patients increasingly present sleep data derived from their wearable devices to their providers and the ever-increasing availability of commercial devices and new-generation research/clinical tools has led to the wide adoption of wearables in research, which has become even more relevant given the discontinuation of the Philips Respironics Actiwatch. Standards for evaluating the performance of wearable sleep-tracking devices have been introduced and the available evidence suggests that consumer-grade devices exceed the performance of traditional actigraphy in assessing sleep as defined by polysomnogram. However, clear limitations exist, for example, the misclassification of wakefulness during the sleep period, problems with sleep tracking outside of the main sleep bout or nighttime period, artifacts, and unclear translation of performance to individuals with certain characteristics or comorbidities. This is of particular relevance when person-specific factors (like skin color or obesity) negatively impact sensor performance with the potential downstream impact of augmenting already existing healthcare disparities. However, wearable sleep-tracking technology holds great promise for our field, given features distinct from traditional actigraphy such as measurement of autonomic parameters, estimation of circadian features, and the potential to integrate other self-reported, objective, and passively recorded health indicators. Scientists face numerous decision points and barriers when incorporating traditional actigraphy, consumer-grade multi-sensor devices, or contemporary research/clinical-grade sleep trackers into their research. Considerations include wearable device capabilities and performance, target population and goals of the study, wearable device outputs and availability of raw and aggregate data, and data extraction, processing, and analysis. Given the difficulties in the implementation and utilization of wearable sleep-tracking technology in real-world research and clinical settings, the following State of the Science review requested by the Sleep Research Society aims to address the following questions. What data can wearable sleep-tracking devices provide? How accurate are these data? What should be taken into account when incorporating wearable sleep-tracking devices into research? These outstanding questions and surrounding considerations motivated this work, outlining practical recommendations for using wearable technology in sleep and circadian research.

Heart rate variability-guided training in professional runners: Effects on performance and vagal modulation.

PURPOSE: To analyze the training structure following a heart rate variability (HRV) -guided training or traditional training protocol, determining their effects on the cardiovascular performance of professional endurance runners, and describing the vagal modulation interaction. METHODS: This was an 8-week cluster-randomized controlled trial. Twelve professional endurance runners were randomly assigned to an HRV-guided training group (HRV-G; n = 6) or a traditional training group (TRAD-G; n = 6). The training methodology followed by the HRV-G was determined by their daily HRV scores. Training intensities were recorded daily. HRV4Training was used to register the rMSSD every morning and during a 60-second period. Cardiovascular outcomes were obtained through an incremental treadmill test. The primary outcome was the maximal oxygen uptake (VO2max). RESULTS: Total training volume was significantly higher in TRAD-G, but moderate intensity training was significantly higher in HRV-G (X ± SDDif=1.98 ± 0.1%; P = 0.006; d = 1.22) and low intensity training in TRAD-G (X ± SDDif=2.03 ± 0.74%; P = 0.004; d = 1.36). The maximal velocity increased significantly in HRV-G (P = 0.027, d = 0.66), while the respiratory exchange ratio increased in TRAD-G (P = 0.017, d = 1). There was a small effect on the LnRMSSD increment (P = 0.365, d = 0.4) in HRV-G. There were statistical inter-group differences in the ∆maximal heart rate when ∆LnrMSSD was considered as a covariable (F = 7.58; P = 0.025; d = 0.487). There were significant and indirect correlations of LnRMSSDTEST with VO2max (r =-0.656, P = 0.02), ∆LnrMSSD with ∆VO2max (r = -0.606, P = 0.037), and ∆LnrMSSDCV with ∆VENT (r = -0.790, P = 0.002). CONCLUSIONS: higher HRV scores suggest better cardiovascular adaptations due to higher training intensities, favoring HRV as a measure to optimize individualized training in professional runners.

Real-Time Estimation of Aerobic Threshold and Exercise Intensity Distribution Using Fractal Correlation Properties of Heart Rate Variability: A Single-Case Field Application in a Former Olympic Triathlete.

A non-linear heart rate variability (HRV) index based on fractal correlation properties called alpha1 of Detrended Fluctuation Analysis (DFA-alpha1), has been shown to change with endurance exercise intensity. Its unique advantage is that it provides information about current absolute exercise intensity without prior lactate or gas exchange testing. Therefore, real-time assessment of this metric during field conditions using a wearable monitoring device could directly provide a valuable exercise intensity distribution without prior laboratory testing for different applied field settings in endurance sports. Until of late no mobile based product could display DFA-alpha1 in real-time using off the shelf consumer products. Recently an app designed for iOS and Android devices, HRV Logger, was updated to assess DFA-alpha1 in real-time. This brief research report illustrates the potential merits of real-time monitoring of this metric for the purposes of aerobic threshold (AT) estimation and exercise intensity demarcation between low (zone 1) and moderate (zone 2) in a former Olympic triathlete. In a single-case feasibility study, three practically relevant scenarios were successfully evaluated in cycling, (1) estimation of a HRV threshold (HRVT) as an adequate proxy for AT using Kubios HRV software via a typical cycling stage test, (2) estimation of the HRVT during real-time monitoring using a cycling 6 min stage test, (3) a simulated 1 h training ride with enforcement of low intensity boundaries and real-time HRVT confirmation. This single-case field evaluation illustrates the potential of an easy-to-use and low cost real-time estimation of the aerobic threshold and exercise intensity distribution using fractal correlation properties of HRV. Furthermore, this approach may enhance the translation of science into endurance sports practice for future real-world settings.

Monitoring the Heart Rate Variability Responses to Training Loads in Competitive Swimmers Using a Smartphone Application and the Banister Impulse-Response Model.

PURPOSE: First, to examine whether heart rate variability (HRV) responses can be modeled effectively via the Banister impulse-response model when the session rating of perceived exertion (sRPE) alone, and in combination with subjective well-being measures, are utilized. Second, to describe seasonal HRV responses and their associations with changes in critical speed (CS) in competitive swimmers. METHODS: A total of 10 highly trained swimmers collected daily 1-minute HRV recordings, sRPE training load, and subjective well-being scores via a novel smartphone application for 15 weeks. The impulse-response model was used to describe chronic root mean square of the successive differences (rMSSD) responses to training, with sRPE and subjective well-being measures used as systems inputs. Changes in CS were obtained from a 3-minute all-out test completed in weeks 1 and 14. RESULTS: The level of agreement between predicted and actual HRV data was R2 = .66 (.25) when sRPE alone was used. Model fits improved in the range of 4% to 21% when different subjective well-being measures were combined with sRPE, representing trivial-to-moderate improvements. There were no significant differences in weekly group averages of log-transformed (Ln) rMSSD (P = .34) or HRV coefficient of variation of Ln rMSSD (P = .12); however, small-to-large changes (d = 0.21-1.46) were observed in these parameters throughout the season. Large correlations were observed between seasonal changes in HRV measures and CS (changes in averages of Ln rMSSD: r = .51, P = .13; changes in coefficient of variation of Ln rMSSD: r = -.68, P = .03). CONCLUSION: The impulse-response model and data collected via a novel smartphone application can be used to model HRV responses to swimming training and nontraining-related stressors. Large relationships between seasonal changes in measured HRV parameters and CS provide further evidence for incorporating a HRV-guided training approach.

Estimating Running Performance Combining Non-invasive Physiological Measurements and Training Patterns in Free-Living.

In this work, we use data acquired longitudinally, in free-living, to provide accurate estimates of running performance. In particular, we used the HRV4Training app and integrated APIs (e.g. Strava and TrainingPeaks) to acquire different sets of parameters, either via user input, morning measurements of resting physiology, or running workouts to estimate running 10 km running time. Our unique dataset comprises data on 2113 individuals, from world class triathletes to individuals just getting started with running, and it spans over 2 years. Analyzed predictors of running performance include anthropometrics, resting heart rate (HR) and heart rate variability (HRV), training physiology (heart rate during exercise), training volume, training patterns (training intensity distribution over multiple workouts, or training polarization) and previous performance. We build multiple linear regression models and highlight the relative impact of different predictors as well as trade-offs between the amount of data required for features extraction and the models accuracy in estimating running performance (10 km time). Cross-validated root mean square error (RMSE) for 10 km running time estimation was 2.6 minutes (4% mean average error, MAE, 0.87 R2), an improvement of 58% with respect to estimation models using anthropometrics data only as predictors. Finally, we provide insights on the relationship between training and performance, including further evidence of the importance of training volume and a polarized training approach to improve performance.

Towards Non-invasive Labour Detection: A Free-Living Evaluation.

In this paper we show early evidence of the feasibility of detecting labour during pregnancy, non-invasively and in free-living. In particular, we present machine learning models aiming at dealing with the challenges of unsupervised, free-living data collection, such as identifying periods of high quality data and detecting physiological changes as labour approaches. During a first phase, physiological data including electrohysterography (EHG, the electrical activity of the uterus), heart rate (HR) and gestational age (GA) were collected in laboratory conditions for model development. In particular, data were collected 1) during simulated activities of daily living, aiming at eliciting artifacts and developing diagnostic models for free-living data 2) during pregnancy, including labour, aiming at developing labour probability models from clean, supervised physiological recordings. Machine learning models using datasets 1) and 2) were deployed in free-living, longitudinally, in 142 pregnant women, between week 22 of pregnancy and delivery. A total of 1014 hours of data and an average of 7 hours per person were collected. Output of the developed models was analyzed to determine the feasibility of detecting labour non-invasively using physiological data, acquired with a single sensor placed on the abdomen. Results showed that the probability of being in labour for recordings collected during the last 24 hours of pregnancy was consistently higher than the probability during any other pregnancy week. Thus, non-invasive labour detection from physiological data seems promising.

Comparison of Heart-Rate-Variability Recording With Smartphone Photoplethysmography, Polar H7 Chest Strap, and Electrocardiography.

PURPOSE: To establish the validity of smartphone photoplethysmography (PPG) and heart-rate sensor in the measurement of heart-rate variability (HRV). METHODS: 29 healthy subjects were measured at rest during 5 min of guided breathing and normal breathing using smartphone PPG, a heart-rate chest strap, and electrocardiography (ECG). The root mean sum of the squared differences between R-R intervals (rMSSD) was determined from each device. RESULTS: Compared to ECG, the technical error of estimate (TEE) was acceptable for all conditions (average TEE CV% [90% CI] = 6.35 [5.13; 8.5]). When assessed as a standardized difference, all differences were deemed "trivial" (average standard difference [90% CI] = 0.10 [0.08; 0.13]). Both PPG- and heart-rate-sensor-derived measures had almost perfect correlations with ECG (R = 1.00 [0.99; 1.00]). CONCLUSION: Both PPG and heart-rate sensors provide an acceptable agreement for the measurement of rMSSD when compared with ECG. Smartphone PPG technology may be a preferred method of HRV data collection for athletes due to its practicality and ease of use in the field.

HRV4Training: Large-scale longitudinal training load analysis in unconstrained free-living settings using a smartphone application.

We describe an approach to support athletes at various fitness levels in their training load analysis using heart rate (HR) and heart rate variability (HRV). A smartphone-based application (HRV4Training) was developed that captures heart activity over one to five minutes using photoplethysmography (PPG) and derives HR and HRV features. HRV4Training integrated a guide for an early morning spot measurement protocol and a questionnaire to capture self-reported training activity. The smartphone application was made publicly available for interested users to quantify training effect. Here we analyze data acquired over a period of 3 weeks to 5 months, including 797 users, breaking down results by gender and age group. Our results suggest a strong relation between HR, HRV and self-reported training load independent of gender and age group. HRV changes due to training were larger than those of HR. We conclude that smartphone-based training monitoring is feasible and a can be used as a practical tool to support large populations outside controlled laboratory environments.

Estimating Oxygen Uptake During Nonsteady-State Activities and Transitions Using Wearable Sensors.

In this paper, we present a method to estimate oxygen uptake ( VO2) during daily life activities and transitions between them. First, we automatically locate transitions between activities and periods of nonsteady-state VO2. Subsequently, we propose and compare activity-specific linear functions to model steady-state activities and transition-specific nonlinear functions to model nonsteady-state activities and transitions. We evaluate our approach in study data from 22 participants that wore a combined accelerometer and heart rate sensor while performing a wide range of activities (clustered into lying, sedentary, dynamic/household, walking, biking and running), including many transitions between intensities, thus resulting in nonsteady-state VO2. Indirect calorimetry was used in parallel to obtain VO2 reference. VO2 estimation error during transitions between sedentary, household and walking activities could be reduced by 16% on average using the proposed approach, compared to state of the art methods.

Cardiorespiratory fitness estimation in free-living using wearable sensors.

OBJECTIVE: In this paper we propose artificial intelligence methods to estimate cardiorespiratory fitness (CRF) in free-living using wearable sensor data. METHODS: Our methods rely on a computational framework able to contextualize heart rate (HR) in free-living, and use context-specific HR as predictor of CRF without need for laboratory tests. In particular, we propose three estimation steps. Initially, we recognize activity primitives using accelerometer and location data. Using topic models, we group activity primitives and derive activities composites. We subsequently rank activity composites, and analyze the relation between ranked activity composites and CRF across individuals. Finally, HR data in specific activity primitives and composites is used as predictor in a hierarchical Bayesian regression model to estimate CRF level from the participant's habitual behavior in free-living. RESULTS: We show that by combining activity primitives and activity composites the proposed framework can adapt to the user and context, and outperforms other CRF estimation models, reducing estimation error between 10.3% and 22.6% on a study population of 46 participants. CONCLUSIONS: Our investigation showed that HR can be contextualized in free-living using activity primitives and activity composites and robust CRF estimation in free-living is feasible.

Cardiorespiratory fitness estimation using wearable sensors: Laboratory and free-living analysis of context-specific submaximal heart rates.

In this work, we propose to use pattern recognition methods to determine submaximal heart rate (HR) during specific contexts, such as walking at a certain speed, using wearable sensors in free living, and using context-specific HR to estimate cardiorespiratory fitness (CRF). CRF of 51 participants was assessed by a maximal exertion test (V̇o2 max). Participants wore a combined accelerometer and HR monitor during a laboratory-based simulation of activities of daily living and for 2 wk in free living. Anthropometrics, HR while lying down, and walking at predefined speeds in laboratory settings were used to estimate CRF. Explained variance (R(2)) was 0.64 for anthropometrics, and increased up to 0.74 for context-specific HR (0.73-0.78 when including fat-free mass). Next, we developed activity recognition and walking speed estimation algorithms to determine the same contexts (i.e., lying down and walking) in free living. Context-specific HR in free living was highly correlated with laboratory measurements (Pearson's r = 0.71-0.75). R(2) for CRF estimation was 0.65 when anthropometrics were used as predictors, and increased up to 0.77 when including free-living context-specific HR (i.e., HR while walking at 5.5 km/h). R(2) varied between 0.73 and 0.80 when including fat-free mass among the predictors. Root mean-square error was reduced from 354.7 to 281.0 ml/min by the inclusion of context-specific HR parameters (21% error reduction). We conclude that pattern recognition techniques can be used to contextualize HR in free living and estimated CRF with accuracy comparable to what can be obtained with laboratory measurements of HR response to walking.

Detection of fetal kicks using body-worn accelerometers during pregnancy: Trade-offs between sensors number and positioning.

Monitoring fetal wellbeing is key in modern obstetrics. While fetal movement is routinely used as a proxy to fetal wellbeing, accurate, noninvasive, long-term monitoring of fetal movement is challenging. A few accelerometer-based systems have been developed in the past few years, to tackle common issues in ultrasound measurement and enable remote, self-administrated monitoring of fetal movement during pregnancy. However, many questions remain unanswered to date on the optimal setup in terms of body-worn accelerometers as well as signal processing and machine learning techniques used to detect fetal movement. In this paper, we systematically analyze the trade-offs between sensor number and positioning, the presence of reference accelerometers outside of the abdominal area and provide guidelines on dealing with class imbalance. Using a dataset of 15 measurements collected employing 6 three-axial accelerometers we show that including a reference accelerometer on the back of the participant consistently improves fetal movement detection performance regardless of the number of sensors utilized. We also show that two accelerometers plus a reference accelerometer are sufficient for optimal results.

Estimating energy expenditure using body-worn accelerometers: a comparison of methods, sensors number and positioning.

Several methods to estimate energy expenditure (EE) using body-worn sensors exist; however, quantifications of the differences in estimation error are missing. In this paper, we compare three prevalent EE estimation methods and five body locations to provide a basis for selecting among methods, sensors number, and positioning. We considered 1) counts-based estimation methods, 2) activity-specific estimation methods using METs lookup, and 3) activity-specific estimation methods using accelerometer features. The latter two estimation methods utilize subsequent activity classification and EE estimation steps. Furthermore, we analyzed accelerometer sensors number and on-body positioning to derive optimal EE estimation results during various daily activities. To evaluate our approach, we implemented a study with 15 participants that wore five accelerometer sensors while performing a wide range of sedentary, household, lifestyle, and gym activities at different intensities. Indirect calorimetry was used in parallel to obtain EE reference data. Results show that activity-specific estimation methods using accelerometer features can outperform counts-based methods by 88% and activity-specific methods using METs lookup for active clusters by 23%. No differences were found between activity-specific methods using METs lookup and using accelerometer features for sedentary clusters. For activity-specific estimation methods using accelerometer features, differences in EE estimation error between the best combinations of each number of sensors (1 to 5), analyzed with repeated measures ANOVA, were not significant. Thus, we conclude that choosing the best performing single sensor does not reduce EE estimation accuracy compared to a five sensors system and can reliably be used. However, EE estimation errors can increase up to 80% if a nonoptimal sensor location is chosen.

Personalization of Energy Expenditure Estimation in Free Living Using Topic Models.

We introduce an approach to personalize energy expenditure (EE) estimates in free living. First, we use topic models to discover activity composites from recognized activity primitives and stay regions in daily living data. Subsequently, we determine activity composites that are relevant to contextualize heart rate (HR). Activity composites were ranked and analyzed to optimize the correlation to HR normalization parameters. Finally, individual-specific HR normalization parameters were used to normalize HR. Normalized HR was then included in activity-specific regression models to estimate EE. Our HR normalization minimizes the effect of individual fitness differences from entering in EE regression models. By estimating HR normalization parameters in free living, our approach avoids dedicated individual calibration or laboratory tests. In a combined free-living and laboratory study dataset, including 34 healthy volunteers, we show that HR normalization in 14-day free-living data improves accuracy compared to no normalization and normalization based on activity primitives only ( 29.4% and 19.8 % error reduction against lab reference). Based on acceleration and HR, both recorded from a necklace, and GPS acquired from a smartphone, EE estimation error was reduced by 10.7 % in a leave-one-participant-out analysis.