Defence Linda Ong: "Unobtrusive Detection of Poor Sitting and Eating Behaviour in a Healthy Population"

Promotors: 1^st promotor: Prof. M. Cao, 2^nd promotor: Prof. G.J. Verkerke, co-promotor: dr Elisabeth Wilhelm

Abstract: Musculoskeletal disorders and metabolic syndrome are two common health issues in the general population. Especially office workers are often affected due to their sedentary lifestyle. Poor sitting, eating habits, and lack of exercise are among the main risk factors that trigger the development of these two health issues. People are often unaware of the impact these behaviours have on their health because the development of these health problems into chronic diseases takes years. To support prevention, we develop unobtrusive technology for the detection of the main risk factors for metabolic syndrome and musculoskeletal disorders. These include poor sitting and eating behaviour. To ensure that the suggested technology can be easily used by end-users, we tested them in semi-controlled lab studies and uncontrolled studies. We demonstrated that machine learning-based algorithms that use pressure mat data can predict sitting positions that are related to back pain. Furthermore, we demonstrate that continuous glucose measurements have a correlation with carbohydrates and other macronutrients consumed in healthy users. This insight can be used to further develop machine learning prediction algorithms that allow the user to keep track of their food consumption effortlessly. In addition to nutritional content, portion size is an important aspect of diet monitoring. Therefore, we also investigated which wearables can be used to monitor the amount of food, drink, and combined food and drink intake. We found that the combination of PPG and IMU provides promising results. This combination of sensors is especially interesting because they are commonly integrated in smartwatches and do not interfere with the privacy of the user. The algorithm developed in this thesis opens new ways of monitoring food intake and sitting behaviour in the context of prevention. By having these technologies integrated into the consumer device, they have a potential to promote adherence to a healthy lifestyle, which could ultimately contribute to a reduction in the loss of quality-adjusted life years.

Dissertation