Algorithms that can determine the type of physical activity (PA) and quantify the intensity can allow precision medicine approaches, such as automated insulin delivery systems that modulate insulin administration in response to PA. In this work, data from a multi-sensor wristband is used to design classifiers to distinguish among five different physical states (PS) (resting, activities of daily living, running, biking, and resistance training), and to develop models to estimate the energy expenditure (EE) of the PA for diabetes therapy. The data collected are filtered, features are extracted from the reconciled signals, and the extracted features are used by machine learning algorithms, including deep-learning techniques, to obtain accurate PS classification and EE estimation. The various machine learning techniques have different success rates ranging from 75.7% to 94.8% in classifying the five different PS. The deep neural network model with long short-term memory has 94.8% classification accuracy. We achieved 0.5 MET (Metabolic Equivalent of Task) root-mean-square error for EE estimation accuracy, relative to indirect calorimetry with randomly selected testing data (10% of collected data). We also demonstrate a 5% improvement in PS classification accuracy and a 0.34 MET decrease in the mean absolute error when using multi-sensor approach relative to using only accelerometer data.

Artificial intelligence (AI) algorithms can provide actionable insights for clinical decision-making and managing chronic diseases. The treatment and management of complex chronic diseases, such as diabetes, stands to benefit from novel AI algorithms analyzing the frequent real-time streaming data and the occasional medical diagnostics and laboratory test results reported in electronic health records (EHR). Novel algorithms are needed to develop trustworthy, responsible, reliable, and robust AI techniques that can handle the imperfect and imbalanced data of EHRs and inconsistencies or discrepancies with free-living self-reported information. The challenges and applications of AI for two problems in the healthcare domain were explored in this work. First, we introduced novel AI algorithms for EHRs designed to be fair and unbiased while accommodating privacy concerns in predicting treatments and outcomes. Then, we studied the innovative approach of using machine learning to improve automated insulin delivery systems through analyzing real-time information from wearable devices and historical data to identify informative trends and patterns in free-living data. Application examples in the treatment of diabetes demonstrate the benefits of AI tools for medical and health informatics.

Wearable devices continuously measure multiple physiological variables to inform users of health and behavior indicators. The computed health indicators must rely on informative signals obtained by processing the raw physiological variables with powerful noise- and artifacts-filtering algorithms. In this study, we aimed to elucidate the effects of signal processing techniques on the accuracy of detecting and discriminating physical activity (PA) and acute psychological stress (APS) using physiological measurements (blood volume pulse, heart rate, skin temperature, galvanic skin response, and accelerometer) collected from a wristband. Data from 207 experiments involving 24 subjects were used to develop signal processing, feature extraction, and machine learning (ML) algorithms that can detect and discriminate PA and APS when they occur individually or concurrently, classify different types of PA and APS, and estimate energy expenditure (EE). Training data were used to generate feature variables from the physiological variables and develop ML models (naïve Bayes, decision tree, k-nearest neighbor, linear discriminant, ensemble learning, and support vector machine). Results from an independent labeled testing data set demonstrate that PA was detected and classified with an accuracy of 99.3%, and APS was detected and classified with an accuracy of 92.7%, whereas the simultaneous occurrences of both PA and APS were detected and classified with an accuracy of 89.9% (relative to actual class labels), and EE was estimated with a low mean absolute error of 0.02 metabolic equivalent of task (MET).The data filtering and adaptive noise cancellation techniques used to mitigate the effects of noise and artifacts on the classification results increased the detection and discrimination accuracy by 0.7% and 3.0% for PA and APS, respectively, and by 18% for EE estimation. The results demonstrate the physiological measurements from wristband devices are susceptible to noise and artifacts, and elucidate the effects of signal processing and feature extraction on the accuracy of detection, classification, and estimation of PA and APS.

Background: Hybrid closed-loop control of glucose levels in people with type 1 diabetes mellitus (T1D) is limited by the requirements on users to manually announce physical activity (PA) and meals to the artificial pancreas system. Multivariable automated insulin delivery (mvAID) systems that can handle unannounced PAs and meals without any manual announcements by the user can improve glycemic control by modulating insulin dosing in response to the occurrence and intensity of spontaneous physical activities. Methods: An mvAID system is developed to supplement the glucose measurements with additional physiological signals from a wristband device, with the signals analyzed using artificial intelligence algorithms to automatically detect the occurrence of PA and estimate its intensity. This additional information gained from the physiological signals enables more proactive insulin dosing adjustments in response to both planned exercise and spontaneous unanticipated physical activities. Results: In silico studies of the mvAID illustrate the safety and efficacy of the system. The mvAID is translated to pilot clinical studies to assess its performance, and the clinical experiments demonstrate an increased time in range and reduced risk of hypoglycemia following unannounced PA and meals. Conclusions: The mvAID systems can increase the safety and efficacy of insulin delivery in the presence of unannounced physical activities and meals, leading to improved lives and less burden on people with T1D.

Background: Predicting carbohydrate intake and physical activity in people with diabetes is crucial for improving blood glucose concentration regulation. Patterns of individual behavior can be detected from historical free-living data to predict meal and exercise times. Data collected in free-living may have missing values and forgotten manual entries. While machine learning (ML) can capture meal and exercise times, missing values, noise, and errors in data can reduce the accuracy of ML algorithms. Methods: Two recurrent neural networks (RNNs) are developed with original and imputed data sets to assess detection accuracy of meal and exercise events. Continuous glucose monitoring (CGM) data, insulin infused from pump data, and manual meal and exercise entries from free-living data are used to predict meals, exercise, and their concurrent occurrence. They contain missing values of various lengths in time, noise, and outliers. Results: The accuracy of RNN models range from 89.9% to 95.7% for identifying the state of event (meal, exercise, both, or neither) for various users. “No meal or exercise” state is determined with 94.58% accuracy by using the best RNN (long short-term memory [LSTM] with 1D Convolution). Detection accuracy with this RNN is 98.05% for meals, 93.42% for exercise, and 55.56% for concurrent meal-exercise events. Conclusions: The meal and exercise times detected by the RNN models can be used to warn people for entering meal and exercise information to hybrid closed-loop automated insulin delivery systems. Reliable accuracy for event detection necessitates powerful ML and large data sets. The use of additional sensors and algorithms for detecting these events and their characteristics provides a more accurate alternative.

Detection and classification of acute psychological stress (APS) and physical activity (PA) in daily lives of people with chronic diseases can provide precision medicine for the treatment of chronic conditions such as diabetes. This study investigates the classification of different types of APS and PA, along with their concurrent occurrences, using the same subset of feature maps via physiological variables measured by a wristband device. Random convolutional kernel transformation is used to extract a large number of feature maps from the biosignals measured by a wristband device (blood volume pulse, galvanic skin response, skin temperature, and 3D accelerometer signals). Three different feature selection techniques (principal component analysis, partial least squares–discriminant analysis (PLS-DA), and sequential forward selection) as well as four approaches for addressing imbalanced sizes of classes (upsampling, downsampling, adaptive synthetic sampling (ADASYN), and weighted training) are evaluated for maximizing detection and classification accuracy. A long short-term memory recurrent neural network model is trained to estimate PA (sedentary state, treadmill run, stationary bike) and APS (non-stress, emotional anxiety stress, mental stress) from wristband signals. The balanced accuracy scores for various combinations of data balancing and feature selection techniques range between 96.82% and 99.99%. The combination of PLS–DA for feature selection and ADASYN for data balancing provide the best overall performance. The detection and classification of APS and PA types along with their concurrent occurrences can provide precision medicine approaches for the treatment of diabetes.