Applications of deep learning methods in digital biomarker research using noninvasive sensing data

Abstract

Introduction: Noninvasive digital biomarkers are critical elements in digital healthcare in terms of not only the ease of measurement but also their use of raw data. In recent years, deep learning methods have been put to use to analyze these diverse heterogeneous data; these methods include representation learning for feature extraction and supervised learning for the prediction of these biomarkers. Methods: We introduce clinical cases of digital biomarkers and various deep-learning methods applied according to each data type. In addition, deep learning methods for the integrated analysis of multidimensional heterogeneous data are introduced, and the utility of these data as an integrated digital biomarker is presented. The current status of digital biomarker research is examined by surveying research cases applied to various types of data as well as modeling methods. Results: We present a future research direction for using data from heterogeneous sources together by introducing deep learning methods for dimensionality reduction and mode integration from multimodal digital biomarker studies covering related domains. The integration of multimodality has led to advances in research through the improvement of performance and complementarity between modes. Discussion: The integrative digital biomarker will be more useful for research on diseases that require data from multiple sources to be treated together. Since delicate signals from patients are not missed and the interaction effects between signals are also considered, it will be helpful for immediate detection and more accurate prediction of symptoms.

Artificial Intelligence and Machine Learning—Powerful Yet Underutilized Tools and Algorithms in Physical Activity and Sedentary Behavior Research

“…15 These promising advances in prediction accuracy, coupled with the proliferation of wearable devices used to measure high-quality PA and sedentary behavior data, 7,8 further inspire the utilization of AI and ML approaches. The digital, noninvasive signals obtained from wearable activity monitors encapsulate traits and phenotypes 16 capable of predicting numerous health indicators and outcomes. These movement behavior traits and phenotypes potentially enable the early identification of future health risks and diseases, providing an unprecedented opportunity to initiate timely interventions to facilitate disease prevention.…”

Machine learning in physical activity, sedentary, and sleep behavior research

“…Newer research is leveraging the power of ML to automatically identify specific features that might exhibit correlations with physical activity, sedentary behavior, and sleep behaviors [69,70]. This may eventually lead to more harmonized analysis of wearable data through automation and increased objectivity achieved by machine intelligence.…”

“…Such innovative approach is exemplified in a recent study that employs an unsupervised methodology to independently uncover dimensions of accelerometry data that are closely linked with both sedentary behavior Farrahi and Rostami Journal of Activity, Sedentary and Sleep Behaviors (2024) 3:5 and physical activity [70]. While the precise relationship between these machine-learned variables and various health outcomes, as well as their practical applications, are areas that require further exploration, it is likely that such variables could potentially be a better predictor of health and diseases [69]. An early demonstration of the effectiveness of ML for learning directly from wearable data is highlighted in another recent study where numerous features extracted from accelerometer data were input into ML algorithms, leading to the remarkable prediction of Parkinson's disease onset years before clinical diagnosis [71].…”

IoT Integration in Telemedicine: Investigating the Role of Internet of Things Devices in Facilitating Remote Patient Monitoring and Data Transmission

“…Take, for instance, IoT-integrated wearable devices monitoring patients with cardiac arrhythmias. These devices continuously record electrocardiogram (ECG) data and employ machine learning algorithms to detect irregular heartbeats, promptly notifying patients and cardiologists (Jeong et al, 2022). This timely intervention, such as medication adjustments or catheter ablation, is crucial for preventing lifethreatening events and improving patients' quality of life (Frodi et al, 2021).…”