Continuous glucose monitoring; learning from wearables to prevent disease
In this project we will derive features of metabolic health (e.g. glycemic variability) from frequently sampled time series of glucose measured by continuous glucose monitoring devices.
Blood glucose levels in the body are maintained within a narrow range through a combination of complex and interconnected mechanisms including the secretion of insulin and the disposal of glucose to peripheral tissues. Deteriorations in these mechanisms may lead to impaired glucose homeostasis culminating in larger and more frequent glucose excursions. In addition, lifestyle factors such as meals and exercise also modulate the variability of the glucose levels when studied under every-day living conditions.
Continuous glucose monitoring devices allow the collection of large amounts of frequently sampled time series of glucose in individuals. This rich data can provide an accurate measure of glycemic variability and allow studying various lifestyle conditions. However, deriving meaningful features that are representative of health state and/or certain external factors is a challenging task.