Performance comparison between wrist and chest actigraphy in combination with heart rate variability for sleep classification.

The concurrent usage of actigraphy and heart rate variability (HRV) for sleep efficiency quantification is still matter of investigation. This study compared chest (CACT) and wrist (WACT) actigraphy (actigraphs positioned on chest and wrist, respectively) in combination with HRV for automatic sleep vs wake classification. Accelerometer and ECG signals were collected during polysomnographic studies (PSGs) including 18 individuals (25-53 years old) with no previous history of sleep disorders. Then, an experienced neurologist performed sleep staging on PSG data. Eleven features from HRV and accelerometry were extracted from series of different lengths. A support vector machine (SVM) was used to automatically distinguish sleep and wake. We found 7 min as the optimal signal length for classification, while maximizing specificity (wake detection). CACT and WACT provided similar accuracies (78% chest vs 77% wrist), larger than what yielded by HRV alone (66%). The addition of HRV to CACT reduced slightly the accuracy, while improving specificity (from 33% to 51%, p < 0.05). On the contrary, the concurrent usage of HRV and WACT did not provide statistically significant improvements over WACT. Then, a subset of features (3 from HRV + 1 from actigraphy) was selected by reducing redundancy using a strategy based on Spearman's correlation and area under the ROC curve. The usage of the reduced set of features and SVM classifier gave only slightly reduced classification performances, which did not differ from the full sets of features. The study opens interesting possibilities in the design of wearable devices for long-term monitoring of sleep at home.

Parametric estimation of sample entropy for physical activity recognition.

Insufficient amount of physical activity, and hence storage of calories may lead depression, obesity, cardiovascular diseases, and diabetes. The amount of consumed calorie depends on the type of activity. The recognition of physical activity is very important to estimate the amount of calories spent by a subject every day. There are some research works already published in the literature for activity recognition through accelerometers (body worn sensors). The accuracy of any recognition system depends on the robustness of selected features and classifiers. The typical features reported for most physical activities recognitions are autoregressive coefficients (ARcoeffs), signal magnitude area (SMA), tilt angle (TA), and standard deviation (STD). In this study, we have studied the feasibility of using single value of sample entropy estimated parametrically (SETH) of an AR model instead of ARcoeffs. After feasibility study, we also compared the recognition accuracies between two popular classifiers i.e. artificial neural network (ANN) and support vector machines (SVM). The recognition accuracies using linear structure (where all types of activities are classified using a single classifier) and hierarchical structure (where activities are first divided into static and dynamic events, and then activities of each event are classified in the second stage). The study showed that the use of SETH provides similar recognition accuracy (69.82%) as provided by ARcoeffs (67.67%) using ANN. The linear structure of SVM performs better (average accuracy of SVM: 98.22%) than linear ANN (average accuracy with ANN: 94.78%). The use of hierarchical structure of ANN increases the average recognition accuracy of static activities to about 100%. However, no significant changes are observed using hierarchical SVM than the linear one.