Wrist-worn sensors have better compliance for activity monitoring compared to hip, waist, ankle or chest positions. However, wrist-worn activity monitoring is challenging due to the wide degree of freedom for the hand movements, as well as similarity of hand movements in different activities such as varying intensities of cycling. To strengthen the ability of wrist-worn sensors in detecting human activities more accurately, motion signals can be complemented by physiological signals such as optical heart rate (HR) based on photoplethysmography. In this paper, an activity monitoring framework using an optical HR sensor and a triaxial wrist-worn accelerometer is presented. We investigated a range of daily life activities including sitting, standing, household activities and stationary cycling with two intensities. A random forest (RF) classifier was exploited to detect these activities based on the wrist motions and optical HR. The highest overall accuracy of 89.6 ± 3.9% was achieved with a forest of a size of 64 trees and 13-s signal segments with 90% overlap. Removing the HR-derived features decreased the classification accuracy of high-intensity cycling by almost 7%, but did not affect the classification accuracies of other activities. A feature reduction utilizing the feature importance scores of RF was also carried out and resulted in a shrunken feature set of only 21 features. The overall accuracy of the classification utilizing the shrunken feature set was 89.4 ± 4.2%, which is almost equivalent to the above-mentioned peak overall accuracy.

Atrial fibrillation (AFib) is the most common sustained heart arrhythmia and is characterized by irregular and excessively frequent ventricular contractions. Early diagnosis of AFib is a key step in the prevention of stroke and heart failure. In this study, we present a comprehensive time-frequency pattern analysis approach for automated detection of AFib from smartphone-derived seismocardiography (SCG) and gyrocardiography (GCG) signals. We sought to assess the diagnostic performance of a smartphone mechanocardiogram (MCG) by considering joint SCG-GCG recordings from 435 subjects including 190 AFib and 245 sinus rhythm (SR) cases. A fully automated AFib detection algorithm consisting of various signal processing and multidisciplinary feature engineering techniques was developed and evaluated through a large set of cross-validation (CV) data including 300 (AFib=150) cardiac patients. The trained model was further tested on an unseen set of recordings including 135 (AFib=40) subjects considered as cross-database (CD). The experimental results showed accuracy, sensitivity, and specificity of approximately 97%, 99%, and 95% for the CV study and up to 95%, 93%, and 97% for the CD test, respectively. The F1 scores were 97% and 96% for the CV and CD, respectively. A positive predictive value of approximately 95% and 92% was obtained respectively for the validation and test sets suggesting high reproducibility and repeatability for mobile AFib detection. Moreover, the kappa coefficient of the method was 0.94 indicating a near-perfect agreement in rhythm classification between the smartphone algorithm and visual interpretation of telemetry recordings. The results support the feasibility of self-monitoring via easy-to-use and accessible MCGs.

Timely diagnosis of cardiovascular diseases (CVD) is crucial to prevent morbidity and mortality. Atrial fibrillation (AFib) and heart failure (HF) are two prevalent cardiac disorders that are associated with a high risk of morbidity and mortality, especially if they are concurrently present. Current approaches fail to screen many at-risk individuals who would benefit from preventive treatment; while others receive unnecessary interventions. An effective approach to the detection of CVDs is mechanocardiography (MCG) by which translational and rotational precordial chest movements are monitored. In this study, we collected MCG data from a study sample of 300 hospitalized cardiac patients using multidimensional built-in inertial sensors of a smartphone. Our main objective was to detect concurrent AFib and acute decompensated HF (ADHF) using smartphone MCG (or sMCG). To this end, we adopted a supervised machine learning classification using multi-label and hierarchical classification. Logistic regression, random forest, and extreme gradient boosting were used as candidate classifiers. The results of the analysis showed the area under the receiver operating characteristic curve values of 0.98 and 0.85 for AFib and ADHF, respectively. The highest percentages of positive and negative predictive values for AFib were 91.9 and 100; while for ADHF, they were 56.9 and 88.4 for the multi-label classification and 69.9 and 68.8 for the hierarchical classification, respectively. We conclude that using a single sMCG measurement, AFib can be detected accurately whereas ADHF can be detected with moderate certainty.

Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Clinical motion characterization in patients with Parkinson’s disease (PD) is challenging: symptom progression, suitability of medication, and level of independence in the home environment can vary across time and from patient to patient. Appointments at the neurological outpatient clinic provide a limited understanding of the overall situation. In order to follow up the variations in the state of PD and heterogenity of symptoms, longer-term measurements performed outside of the clinic could help to optimize and personalize therapies. Several wearable sensors have been used and algorithms have been developed to estimate the severity of symptoms in PD; however, longitudinal recordings, even for shorter duration of few days are rare. Home recordings have the potential benefit of providing a more thorough and objective follow-up about the disease, while providing more information about the possible need to change medications or consider invasive secondary treatments. Objective: The primary objective of this study is to collect a dataset for developing methods for detecting PD-related symptoms that are visible in walking patterns at home. The movement data are collected continuously and remotely at home during the normal lives of patients with Parkinson’s disease, as well as controls. The secondary objective is to use the dataset to study whether the registered medication intakes can be identified from the collected movement data by looking for and analyzing short term changes in the walking patterns. Methods: This observational case-control study protocol measures the activity using three different devices: (1) a smartphone with a built-in accelerometer, gyroscope, and phone orientation sensor, (2) a Movesense® smart sensor to measure movement data from the wrist, and (3) a Forciot® smart insole to measure the temporal and spatial distributions of the forces applied on the feet. The measurements are first collected during the appointment at the clinic through background interviews, a comprehensive Unified Parkinson’s Disease Rating Scale (UPDRS) test, and a 20-step walking test conducted by a trained clinical physiotherapist. After which the subjects wear the smartphone at home for three consecutive days. Wrist and insole sensors are not used in the home recordings. Results: Data collection began in March 2018. The subject recruitment and data collection will continue in spring 2019. The intended sample was 150 subjects. In 2018, we collected a sample of 103 subjects, 66 diagnosed with PD and 37 control subjects. Conclusions: This study aims to produce an extensive movement sensor dataset recorded from PD patients in various phases of the disease, as well as a group of control subjects for effective and impactful comparison studies. The study also aims to develop data analysis methods to monitor PD symptoms and the effects of medication intake during normal life and outside of the clinic. Further applications of these methods may include tools for healthcare professionals to better monitor PD remotely and applying them also to other movement disorders. Clinical Trial: This study is registered under ClinicalTrials.gov (NCT03366558). https://clinicaltrials.gov/ct2/show/NCT03366558?term=k%C3%A4veli&rank=1. Accessed: 2018-08-29. Archived at: http://www.webcitation.org/721n3uIqR

In principle, connected sensors allow effortless long-term self-monitoring of health and wellness that can help maintain health and quality of life. However, data collected in the “wild” may be noisy and contain outliers, e.g., due to uncontrolled sources or data from different persons using the same device. The removal of the “outliers” is therefore critical for accurate interpretation of the data. In this paper we study the detection and elimination of outliers in self-weighing time series data obtained from connected weight scales. We examined three techniques: (1) a method based on autoregressive integrated moving average (ARIMA) time series modelling, (2) median absolute deviation (MAD) scale estimate, and (3) a method based on Rosner statistics. We applied these methods to both a data set with real outliers and a clean data set corrupted with simulated outliers. The results suggest that the simple MAD algorithm and ARIMA performed well with both test sets while the Rosner statistics was significantly less effective. In addition, the ARIMA approach appeared to be significantly less sensitive to long periods of missing data than MAD and Rosner statistics.