AUTHOR=Alhaddad Ahmad Yaser , Aly Hussein , Gad Hoda , Al-Ali Abdulaziz , Sadasivuni Kishor Kumar , Cabibihan John-John , Malik Rayaz A. 

TITLE=Sense and Learn: Recent Advances in Wearable Sensing and Machine Learning for Blood Glucose Monitoring and Trend-Detection

JOURNAL=Frontiers in Bioengineering and Biotechnology

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/bioengineering-and-biotechnology/articles/10.3389/fbioe.2022.876672

DOI=10.3389/fbioe.2022.876672

ISSN=2296-4185

ABSTRACT=Diabetes mellitus is a global epidemic affecting an estimated 537 million adults with an expected increase to 784 million by 2045.  Diagnosed by elevated blood glucose it leads to long-term complications including blindness, dialysis and amputation as well as myocardial infarction and stroke. In our attempts to normalize blood glucose and prevent or slow down the development of long-term complications, there is a 3-fold increase in the risk of developing hypoglycemia which can result in cardiac arrythmias and death as well as falls with fractures and road traffic accidents due to disturbance in cognition. The demand for non-invasive blood glucose and physiological monitoring has increased considerably amongst people with type 1 diabetes and insulin treated type 2 diabetes. We provide a timely update on novel technologies for non-invasive blood glucose and physiological monitoring over the past five years which include the continuous electrocardiogram, electromagnetic, bioimpedance, photoplethysmography, and acceleration measures as well as bodily fluid glucose sensors to monitor glucose detection. We also review the potential of machine learning algorithms to predict blood glucose trends from physiological monitoring, especially to predict the development of hypoglycemia. Convolutional and recurrent neural networks, support vector machines, and decision trees are examples of the machine learning algorithms used to forecast and predict blood glucose levels and trends. We also address the key limitations and challenges of these studies and provide recommendations for future work in this area.