A sleep stage estimation algorithm based on cardiorespiratory signals derived from a suprasternal pressure sensor.

Automatic estimation of sleep structure is an important aspect in moving sleep monitoring from clinical laboratories to people's homes. However, the transition to more portable systems should not happen at the expense of important physiological signals, such as respiration. Here, we propose the use of cardiorespiratory signals obtained by a suprasternal pressure (SSP) sensor to estimate sleep stages. The sensor is already used for diagnosis of sleep-disordered breathing (SDB) conditions, but besides respiratory effort it can detect cardiac vibrations transmitted through the trachea. We collected the SSP sensor signal in 100 adults (57 male) undergoing clinical polysomnography for suspected sleep disorders, including sleep apnea syndrome, insomnia, and movement disorders. Here, we separate respiratory effort and cardiac activity related signals, then input these into a neural network trained to estimate sleep stages. Using the original mixed signal the results show a moderate agreement with manual scoring, with a Cohen's kappa of 0.53 in Wake/N1-N2/N3/rapid eye movement sleep discrimination and 0.62 in Wake/Sleep. We demonstrate that decoupling the two signals and using the cardiac signal to estimate the instantaneous heart rate improves the process considerably, reaching an agreement of 0.63 and 0.71. Our proposed method achieves high accuracy, specificity, and sensitivity across different sleep staging tasks. We also compare the total sleep time calculated with our method against manual scoring, with an average error of -1.83 min but a relatively large confidence interval of ±55 min. Compact systems that employ the SSP sensor information-rich signal may enable new ways of clinical assessments, such as night-to-night variability in obstructive sleep apnea and other sleep disorders.

Methodological Shortcomings of Wrist-Worn Heart Rate Monitors Validations.

Wearable sensor technology could have an important role for clinical research and in delivering health care. Accordingly, such technology should undergo rigorous evaluation prior to market launch, and its performance should be supported by evidence-based marketing claims. Many studies have been published attempting to validate wrist-worn photoplethysmography (PPG)-based heart rate monitoring devices, but their contrasting results question the utility of this technology. The reason why many validations did not provide conclusive evidence of the validity of wrist-worn PPG-based heart rate monitoring devices is mostly methodological. The validation strategy should consider the nature of data provided by both the investigational and reference devices. There should be uniformity in the statistical approach to the analyses employed in these validation studies. The investigators should test the technology in the population of interest and in a setting appropriate for intended use. Device industries and the scientific community require robust standards for the validation of new wearable sensor technology.

Proof-of-concept model for instantaneous heart rate-drift correction during low and high exercise exertion.

Introduction: This study aimed to model below and above anaerobic threshold exercise-induced heart rate (HR) drift, so that the corrected HR could better represent V˙O2 kinetics during and after the exercise itself. Methods: Fifteen healthy subjects (age: 28 ± 5 years; V˙O2Max: 50 ± 8 mL/kg/min; 5 females) underwent a maximal and a 30-min submaximal (80% of the anaerobic threshold) running exercises. A five-stage computational (i.e., delay block, new training impulse-calculation block, Sigmoid correction block, increase block, and decrease block) model was built to account for instantaneous HR, fitness, and age and to onset, increase, and decrease according to the exercise intensity and duration. Results: The area under the curve (AUC) of the hysteresis function, which described the differences in the maximal and submaximal exercise-induced V˙O2 and HR kinetics, was significantly reduced for both maximal (26%) and submaximal (77%) exercises and consequent recoveries. Discussion: In conclusion, this model allowed HR drift instantaneous correction, which could be exploited in the future for more accurate V˙O2 estimations.

Quantitative validation of the suprasternal pressure signal to assess respiratory effort during sleep.

Objective.Intra-esophageal pressure (Pes) measurement is the recommended gold standard to quantify respiratory effort during sleep, but used to limited extent in clinical practice due to multiple practical drawbacks. Respiratory inductance plethysmography belts (RIP) in conjunction with oronasal airflow are the accepted substitute in polysomnographic systems (PSG) thanks to a better usability, although they are partial views on tidal volume and flow rather than true respiratory effort and are often used without calibration. In their place, the pressure variations measured non-invasively at the suprasternal notch (SSP) may provide a better measure of effort. However, this type of sensor has been validated only for respiratory events in the context of obstructive sleep apnea syndrome (OSA). We aim to provide an extensive verification of the suprasternal pressure signal against RIP belts and Pes, covering both normal breathing and respiratory events.Approach.We simultaneously acquired suprasternal (207) and esophageal pressure (20) signals along with RIP belts during a clinical PSG of 207 participants. In each signal, we detected breaths with a custom algorithm, and evaluated the SSP in terms of detection quality, breathing rate estimation, and similarity of breathing patterns against RIP and Pes. Additionally, we examined how the SSP signal may diverge from RIP and Pes in presence of respiratory events scored by a sleep technician.Main results.The SSP signal proved to be a reliable substitute for both esophageal pressure (Pes) and respiratory inductance plethysmography (RIP) in terms of breath detection, with sensitivity and positive predictive value exceeding 75%, and low error in breathing rate estimation. The SSP was also consistent with Pes (correlation of 0.72, similarity 80.8%) in patterns of increasing pressure amplitude that are common in OSA.Significance.This work provides a quantitative analysis of suprasternal pressure sensors for respiratory effort measurements.

Extraction of cardiac-related signals from a suprasternal pressure sensor during sleep.

Objective.The accurate detection of respiratory effort during polysomnography is a critical element in the diagnosis of sleep-disordered breathing conditions such as sleep apnea. Unfortunately, the sensors currently used to estimate respiratory effort are either indirect and ignore upper airway dynamics or are too obtrusive for patients. One promising alternative is the suprasternal notch pressure (SSP) sensor: a small element placed on the skin in the notch above the sternum within an airtight capsule that detects pressure swings in the trachea. Besides providing information on respiratory effort, the sensor is sensitive to small cardiac oscillations caused by pressure perturbations in the carotid arteries or the trachea. While current clinical research considers these as redundant noise, they may contain physiologically relevant information.Approach.We propose a method to separate the signal generated by cardiac activity from the one caused by breathing activity. Using only information available from the SSP sensor, we estimate the heart rate and track its variations, then use a set of tuned filters to process the original signal in the frequency domain and reconstruct the cardiac signal. We also include an overview of the technical and physiological factors that may affect the quality of heart rate estimation. The output of our method is then used as a reference to remove the cardiac signal from the original SSP pressure signal, to also optimize the assessment of respiratory activity. We provide a qualitative comparison against methods based on filters with fixed frequency cutoffs.Main results.In comparison with electrocardiography (ECG)-derived heart rate, we achieve an agreement error of 0.06 ± 5.09 bpm, with minimal bias drift across the measurement range, and only 6.36% of the estimates larger than 10 bpm.Significance.Together with qualitative improvements in the characterization of respiratory effort, this opens the development of novel portable clinical devices for the detection and assessment of sleep disordered breathing.

Respiratory activity extracted from wrist-worn reflective photoplethysmography in a sleep-disordered population.

OBJECTIVE: Respiratory activity is an essential parameter to monitor healthy and disordered sleep, and unobtrusive measurement methods have important clinical applications in diagnostics of sleep-related breathing disorders. We propose a respiratory activity surrogate extracted from wrist-worn reflective photoplethysmography validated on a heterogeneous dataset of 389 sleep recordings. APPROACH: The surrogate was extracted by interpolating the amplitude of the PPG pulses after evaluation of pulse morphological quality. Subsequent multistep post-processing was applied to remove parts of the surrogate with low quality and high motion levels. In addition to standard respiration rate performance, we evaluated the similarity between surrogate respiratory activity and reference inductance plethysmography of the thorax, using Spearman's correlations and spectral coherence, and assessed the influence of PPG signal quality, motion levels, sleep stages and obstructive sleep apnea. MAIN RESULTS: Prior to post-processing, the surrogate already had a strong similarity with the reference (correlation = 0.54, coherence = 0.81), and reached respiration rate estimation performance in line with the literature (estimation error = 0.76± 2.11 breaths/min). Detrimental effects of low PPG quality, high motion levels and sleep-dependent physiological phenomena were significantly mitigated by the proposed post-processing steps (correlation = 0.62, coherence = 0.88, estimation error = 0.53± 1.85 breaths/min). SIGNIFICANCE: Wrist-worn PPG can be used to extract respiratory activity, thus allowing respiration monitoring in real-world sleep medicine applications using (consumer) wearable devices.

Wearable monitoring of sleep-disordered breathing: estimation of the apnea-hypopnea index using wrist-worn reflective photoplethysmography.

A large part of the worldwide population suffers from obstructive sleep apnea (OSA), a disorder impairing the restorative function of sleep and constituting a risk factor for several cardiovascular pathologies. The standard diagnostic metric to define OSA is the apnea-hypopnea index (AHI), typically obtained by manually annotating polysomnographic recordings. However, this clinical procedure cannot be employed for screening and for long-term monitoring of OSA due to its obtrusiveness and cost. Here, we propose an automatic unobtrusive AHI estimation method fully based on wrist-worn reflective photoplethysmography (rPPG), employing a deep learning model exploiting cardiorespiratory and sleep information extracted from the rPPG signal trained with 250 recordings. We tested our method with an independent set of 188 heterogeneously disordered clinical recordings and we found it estimates the AHI with a good agreement to the gold standard polysomnography reference (correlation = 0.61, estimation error = 3±10 events/h). The estimated AHI was shown to reliably assess OSA severity (weighted Cohen's kappa = 0.51) and screen for OSA (ROC-AUC = 0.84/0.86/0.85 for mild/moderate/severe OSA). These findings suggest that wrist-worn rPPG measurements that can be implemented in wearables such as smartwatches, have the potential to complement standard OSA diagnostic techniques by allowing unobtrusive sleep and respiratory monitoring.

Quarter-mile walk test sensitive to training-induced fitness changes.

BACKGROUND: Cardiorespiratory fitness (CRF) is an important aspect of the overall health of an individual and its monitoring must be promoted in the general population. Thus, the aim of the study was to cross-validate and improve CRF estimation based on quarter-mile Rockport Fitness Walking Test. METHODS: Thirty participants (31.4±7.99 years) were randomized in either a four-week aerobic training group (10 men and 10 women) or a control group (eight men and two women). CRF was assessed via VO2max test and estimated via quarter-mile Rockport Fitness and Ebbeling treadmill tests, before and after the training intervention. The original quarter-mile Rockport VO2max estimation was found to greatly overestimate CRF by 22 mL/kg/min. When its coefficient was updated according to our data, it largely improved (by 6.8 mL/kg/min). Furthermore, a new algorithm for predicting VO2max was designed using multi-linear regression analysis. RESULTS: The original quarter-mile Rockport Fitness Walking Test was not sensitive to CRF changes. It showed changes in VO2max, which were significantly different from the actual observed changes (-1.1±4.08 vs. 1.61±2.84, P=0.02, respectively). The Ebbeling treadmill test appeared to systematically overestimate CRF changes. Our new algorithm showed improved sensitivity for detecting CRF changes and stability. CONCLUSIONS: The original quarter-mile Rockport Fitness Walking Test equation for predicting VO2max was neither accurate nor sensitive to changes in CRF, most likely due to cardiovascular drift. Our new algorithm, based on the same brisk walking test, can provide a more accurate estimate of CRF, which is also sensitive to VO2max changes, in a broad age range (18 to 50 years).

Estimation of the apnea-hypopnea index in a heterogeneous sleep-disordered population using optimised cardiovascular features.

Obstructive sleep apnea (OSA) is a highly prevalent sleep disorder, which results in daytime symptoms, a reduced quality of life as well as long-term negative health consequences. OSA diagnosis and severity rating is typically based on the apnea-hypopnea index (AHI) retrieved from overnight poly(somno)graphy. However, polysomnography is costly, obtrusive and not suitable for long-term recordings. Here, we present a method for unobtrusive estimation of the AHI using ECG-based features to detect OSA-related events. Moreover, adding ECG-based sleep/wake scoring yields a fully automatic method for AHI-estimation. Importantly, our algorithm was developed and validated on a combination of clinical datasets, including datasets selectively including OSA-pathology but also a heterogeneous, "real-world" clinical sleep disordered population (262 participants in the validation set). The algorithm provides a good representation of the current gold standard AHI (0.72 correlation, estimation error of 0.56 ± 14.74 events/h), and can also be employed as a screening tool for a large range of OSA severities (ROC AUC ≥ 0.86, Cohen's kappa ≥ 0.53 and precision ≥70%). The method compares favourably to other OSA monitoring strategies, showing the feasibility of cardiovascular-based surrogates for sleep monitoring to evolve into clinically usable tools.

On the generalizability of ECG-based obstructive sleep apnea monitoring: merits and limitations of the Apnea-ECG database.

Obstructive sleep apnea syndrome (OSAS) is a sleep disorder that affects a large part of the population and the development of algorithms using cardiovascular features for OSAS monitoring has been an extensively researched topic in the last two decades. Several studies regarding automatic apneic event classification using ECG derived features are based on the public Apnea-ECG database available on PhysioNet. Although this database is an excellent starting point for apnea topic investigations, in our study we show that algorithms for apneic-epochs classification that are successfully trained on this database (sensitivity < 85%, false detection rate <20%) perform poorly (sensitivity\textit<55%, false detection rate < 40%) in other databases which include patients with a broader spectrum of apneic events and sleep disorders. The reduced performance can be related to the complexity of breathing events, the increased number of non-breathing related sleep events, and the presence of non-OSAS sleep pathologies.

Ecological momentary assessment of food perceptions and eating behavior using a novel phone application in adults with or without obesity.

We developed a smart phone application to measure participants' food-reward perceptions and eating behavior in their naturalistic environment. Intensity ratings (0 - not at all to 10 - very strongly) of perceived anticipation of food (wanting) and food enjoyment at endpoint of intake (liking) were recorded as they occurred over a period of 14 days. Moreover, food craving trait, implicit and explicit attitude towards healthy food, and body composition were assessed. 53 participants provided complete data. Participants were classified by percentage of body fat; 33 participants with lower body fat (L-group) and 20 with higher body fat (H-group; ≥25% body fat for males and ≥32% for females). L-group participants reported 6.34 (2.00) food wanting events per day, whereas H-group participants recorded significantly fewer food wanting events (5.07 (1.42)); both groups resisted about the same percentage of wanting events (L-group: 29.2 (15.5)%; H-group 27.3 (12.8)%). Perceived intensity ratings were significantly different within the L-group in the order liking (7.65 (0.81)) > un-resisted wanting (leading to eating) (7.00 (1.01)) > resisted wanting (not leading to eating) (6.02 (1.72)) but not in the H-group. Liking scores (L-group: 7.65 (0.81); H-group: 7.14 (1.04)) were significantly higher in L-group than in H-group after controlling for age. Our results show that individuals with higher percentage of body fat show less food enjoyment after intake and reveal no differentiation in intensity ratings of perceived anticipatory and consummatory food reward. These results are consistent with a hypothesized reward deficiency among individuals with higher percentage of body fat.

Wrist-worn optical and chest strap heart rate comparison in a heterogeneous sample of healthy individuals and in coronary artery disease patients.

BACKGROUND: The need for unobtrusive HR (heart rate) monitoring has led to the development of a new generation of strapless HR monitors. The aim of this study was to determine whether such an unobtrusive, wrist-worn optical HR monitor (OHRM) could be equivalent and therefore a valid alternative to a traditional chest strap during a broad range of activities in a heterogeneous healthy population and coronary artery disease (CAD) patients. METHODS: One hundred ninety-nine healthy volunteers, 84 males and 115 females, including 35 overweight-obese subjects, 53 pregnant women, and 20 CAD patients were tested in the present study. Second-by-second HR measured by the OHRM was concurrently evaluated against an ECG-based chest strap monitor during a broad range of activities (i.e., walking, running, cycling, gym, household, and sedentary activities). RESULTS: Data coverage, percentage of time the OHRM provides a HR not larger than 10 bpm from the reference, went from a minimum of 92% of the time in the least periodic activity (i.e., gym), to 95% during the most intense activity (i.e., running), and to a maximum of 98% for sedentary activities. The limits of agreement of the difference between the OHRM and the chest strap HR were within the range of ±15 bpm. The OHRM showed a concordance correlation coefficient of 0.98. Overall, the mean absolute error was not larger than 3 bpm, which can be considered clinically acceptable for a number of applications. A similar performance was found for CAD (94.2% coverage, 2.4 bpm error), but the small sample size does not allow any quantitative comparison. CONCLUSION: Heart rate measured by OHRM at the wrist and ECG-based HR measured via a traditional chest strap are acceptably close in a broad range of activities in a heterogeneous, healthy population, and showed initial promising results also in CAD patients.

Sinus or not: a new beat detection algorithm based on a pulse morphology quality index to extract normal sinus rhythm beats from wrist-worn photoplethysmography recordings.

OBJECTIVE: Wrist-worn photoplethysmography (PPG) can enable free-living physiological monitoring of people during diverse activities, ranging from sleep to physical exercise. In many applications, it is important to remove the pulses not related to sinus rhythm beats from the PPG signal before using it as a cardiovascular descriptor. In this manuscript, we propose an algorithm to assess the morphology of the PPG signal in order to reject non-sinus rhythm pulses, such as artefacts or pulses related to arrhythmic beats. APPROACH: The algorithm segments the PPG signal into individual pulses and dynamically evaluates their morphological likelihood of being normal sinus rhythm pulses via a template-matching approach that accounts for the physiological variability of the signal. The normal sinus rhythm likelihood of each pulse is expressed as a quality index that can be employed to reject artefacts and pulses related to arrhythmic beats. MAIN RESULTS: Thresholding the pulse quality index enables near-perfect detection of normal sinus rhythm beats by rejecting most of the non-sinus rhythm pulses (positive predictive value 98%-99%), both in healthy subjects and arrhythmic patients. The rejection of arrhythmic beats is almost complete (sensitivity to arrhythmic beats 7%-3%), while the sensitivity to sinus rhythm beats is not compromised (96%-91%). SIGNIFICANCE: The developed algorithm consistently detects normal sinus rhythm beats in a PPG signal by rejecting artefacts and, as a first of its kind, arrhythmic beats. This increases the reliability in the extraction of features which are adversely influenced by the presence of non-sinus pulses, whether related to inter-beat intervals or to pulse morphology, from wrist-worn PPG signals recorded in free-living conditions.

Proof of concept of a 45-second cardiorespiratory fitness self-test for coronary artery disease patients based on accelerometry.

Cardiorespiratory fitness (CRF) provides important diagnostic and prognostic information. It is measured directly via laboratory maximal testing or indirectly via submaximal protocols making use of predictor parameters such as submaximal [Formula: see text], heart rate, workload, and perceived exertion. We have established an innovative methodology, which can provide CRF prediction based only on body motion during a periodic movement. Thirty healthy subjects (40% females, 31.3 ± 7.8 yrs, 25.1 ± 3.2 BMI) and eighteen male coronary artery disease (CAD) (56.6 ± 7.4 yrs, 28.7 ± 4.0 BMI) patients performed a [Formula: see text] test on a cycle ergometer as well as a 45 second squatting protocol at a fixed tempo (80 bpm). A tri-axial accelerometer was used to monitor movements during the squat exercise test. Three regression models were developed to predict CRF based on subject characteristics and a new accelerometer-derived feature describing motion decay. For each model, the Pearson correlation coefficient and the root mean squared error percentage were calculated using the leave-one-subject-out cross-validation method (rcv, RMSEcv). The model built with all healthy individuals' data showed an rcv = 0.68 and an RMSEcv = 16.7%. The CRF prediction improved when only healthy individuals with normal to lower fitness (CRF<40 ml/min/kg) were included, showing an rcv = 0.91 and RMSEcv = 8.7%. Finally, our accelerometry-based CRF prediction CAD patients, the majority of whom taking ß-blockers, still showed high accuracy (rcv = 0.91; RMSEcv = 9.6%). In conclusion, motion decay and subject characteristics could be used to predict CRF in healthy people as well as in CAD patients taking ß-blockers, accurately. This method could represent a valid alternative for patients taking ß-blockers, but needs to be further validated in a larger population.

Photoplethysmography beat detection and pulse morphology quality assessment for signal reliability estimation.

Photoplethysmography (PPG) is one of the key technologies for unobtrusive physiological monitoring, with ongoing attempts to use it in several medical fields, ranging from night to night sleep analysis to continuous cardiac arrhythmia monitoring. However, the PPG signals are susceptible to be corrupted by noise and artifacts, caused, e.g., by limb or sensor movement. These artifacts affect the morphology of PPG waves and prevent the accurate detection and localization of beats and subsequent cardiovascular feature extraction. In this paper a new algorithm for beat detection and pulse quality assessment is described. The algorithm segments the PPG signal in pulses, localizes each beat and grades each segment with a quality index. The obtained index results from a comparison between each pulse and a template derived from the surrounding pulses, by mean of dynamic time warping barycenter averaging. The quality index is used to discard corrupted pulse beats. The algorithm is evaluated by comparing the detected beats with annotated PPG signals and the results are published over the same data. The described method achieves an improved sensitivity and a higher predictive value.

Energy expenditure estimation in beta-blocker-medicated cardiac patients by combining heart rate and body movement data.

BACKGROUND: Accurate assessment of energy expenditure provides an opportunity to monitor physical activity during cardiac rehabilitation. However, the available assessment methods, based on the combination of heart rate (HR) and body movement data, are not applicable for patients using beta-blocker medication. Therefore, we developed an energy expenditure prediction model for beta-blocker-medicated cardiac rehabilitation patients. METHODS: Sixteen male cardiac rehabilitation patients (age: 55.8 ± 7.3 years, weight: 93.1 ± 11.8 kg) underwent a physical activity protocol with 11 low- to moderate-intensity common daily life activities. Energy expenditure was assessed using a portable indirect calorimeter. HR and body movement data were recorded during the protocol using unobtrusive wearable devices. In addition, patients underwent a symptom-limited exercise test and resting metabolic rate assessment. Energy expenditure estimation models were developed using multivariate regression analyses based on HR and body movement data and/or patient characteristics. In addition, a HR-flex model was developed. RESULTS: The model combining HR and body movement data and patient characteristics showed the highest correlation and lowest error (r2 = 0.84, root mean squared error = 0.834 kcal/minute) with total energy expenditure. The method based on individual calibration data (HR-flex) showed lower accuracy (i2 = 0.83, root mean squared error = 0.992 kcal/minute). CONCLUSIONS: Our results show that combining HR and body movement data improves the accuracy of energy expenditure prediction models in cardiac patients, similar to methods that have been developed for healthy subjects. The proposed methodology does not require individual calibration and is based on the data that are available in clinical practice.

A 45-Second Self-Test for Cardiorespiratory Fitness: Heart Rate-Based Estimation in Healthy Individuals.

Cardio-respiratory fitness (CRF) is a widespread essential indicator in Sports Science as well as in Sports Medicine. This study aimed to develop and validate a prediction model for CRF based on a 45 second self-test, which can be conducted anywhere. Criterion validity, test re-test study was set up to accomplish our objectives. Data from 81 healthy volunteers (age: 29 ± 8 years, BMI: 24.0 ± 2.9), 18 of whom females, were used to validate this test against gold standard. Nineteen volunteers repeated this test twice in order to evaluate its repeatability. CRF estimation models were developed using heart rate (HR) features extracted from the resting, exercise, and the recovery phase. The most predictive HR feature was the intercept of the linear equation fitting the HR values during the recovery phase normalized for the height2 (r2 = 0.30). The Ruffier-Dickson Index (RDI), which was originally developed for this squat test, showed a negative significant correlation with CRF (r = -0.40), but explained only 15% of the variability in CRF. A multivariate model based on RDI and sex, age and height increased the explained variability up to 53% with a cross validation (CV) error of 0.532 L ∙ min-1 and substantial repeatability (ICC = 0.91). The best predictive multivariate model made use of the linear intercept of HR at the beginning of the recovery normalized for height2 and age2; this had an adjusted r2 = 0. 59, a CV error of 0.495 L·min-1 and substantial repeatability (ICC = 0.93). It also had a higher agreement in classifying CRF levels (κ = 0.42) than RDI-based model (κ = 0.29). In conclusion, this simple 45 s self-test can be used to estimate and classify CRF in healthy individuals with moderate accuracy and large repeatability when HR recovery features are included.

Predicting energy expenditure from photo-plethysmographic measurements of heart rate under beta blocker therapy: Data driven personalization strategies based on mixed models.

Energy expenditure have been often estimated using computational models based on heart rate (HR) and appropriate personalization strategies to account for users cardio-respiratory characteristics. However, medications like beta blockers which are prescribed to treat several cardiac conditions have a direct influence on the cardiovascular system and may impact the relationship between HR and energy expenditure during physical activity (AEE). This study proposes to estimate AEE from HR using mixed models (MIX-REG) by introducing a novel method to personalize the prediction equation. We selected as features to represent the individual random effect in the MIX-REG model those subject characteristics which minimized both estimation error (RMSE) and between-subjects error bias variability. Data from 17 patients post-myocardial infarction were collected during a laboratory protocol. AEE was measured using indirect calorimetry and HR using an innovative wrist worn activity monitor equipped with the Philips Cardio and Motion Monitoring Module (CM3-Generation-1), which is an integrated module including a photo-plethysmographic and accelerometer sensor. The presented method showed large AEE estimation accuracy (RMSE = 1.35 kcal/min) which was comparable to that of models personalized using data from laboratory calibration protocols (HR-FLEX) and was superior to multi-linear regression and MIX-REG models trained using a stepwise features selection procedure.