his work is focused on the relationship between sleep and blood glucose control in people with Type 1 Diabetes and on the development of a... Show morehis work is focused on the relationship between sleep and blood glucose control in people with Type 1 Diabetes and on the development of a sleep module incorporating new variables and rules for use in automated insulin delivery and advisory systems. Through this research, sleep effects were identified, quantified and incorporated into a multivariable artificial pancreas system (mvAP) that is currently being developed. The mvAP uses different physiological signals acquired through non-invasive wearable sensors along with a continuous glucose monitor (CGM) to detect the state of the user to predict future blood glucose values to aid in insulin dosing decisions. The overall objective of the research was to develop and add a module to further improve the successful mvAP by incorporating sleep related information while retaining the functionality and safety of the system and improving the effectiveness in maintaining good glycemic control. Two types of sleep effects were studied: effects of sleep characteristics and stages in real-time (during sleep) and effects of sleep on glucose metabolism the next day. It was found that poor sleep quality was related to higher glycemic variability overnight in adults with Type 1 Diabetes. However, in adults without diabetes, there were no consistent relationships found between sleep stages and changes in blood glucose levels overnight. For adults with Type 1 Diabetes, it was determined that Sleep Quality, Total Sleep Time, Wake After Sleep Onset (WASO), Number of Awakenings >5 minutes, and amount of Deep sleep could be used in conjunction with insulin on board and the amount of time that has passed since the user has woken up to predict how much more insulin may be needed at the first meal of the day. This Insulin Multiplier Algorithm was tested and validated in replay simulations. Finally, in order to incorporate these relationships into the mvAP, a sleep stage detection algorithm was developed using the Empatica E4 wristband. Show less
