PhD defence
Machine learning methods to quantify the physical behaviour of older adults and a study of healthy ageing
- S. Paraschiakos
- Date
- Wednesday 8 April 2026
- Time
- Location
-
Academy Building
Rapenburg 73
2311 GJ Leiden
Supervisor(s)
- Prof.dr. E.P. Slagboom
- dr. A.J. Knobbe
Summary
This research investigates how wearable sensors and artificial intelligence can be used to better understand physical behaviour in older adults and its impact on healthy ageing. It had two main goals: first, to develop improved methods for measuring daily activities and energy expenditure using wearable devices; and second, to examine how changes in physical behaviour relate to health outcomes.
Using data from the Growing Old Together Validation (GOTOV) study and the 13-week Growing Old Together (GOTO) lifestyle intervention, advanced machine learning and deep learning models were developed and compared. Traditional methods were evaluated alongside recurrent neural networks, particularly GRU models. The deep learning approaches proved more accurate and efficient in recognising daily activities and estimating energy expenditure under real-life conditions. A key innovation was the combination of raw sensor data with personal characteristics such as age and body composition, enabling more personalised, scalable and realistic monitoring.
The findings show that even modest increases in daily physical activity are associated with measurable improvements in immuno-metabolic health markers in older adults. Clear differences between men and women were observed: men tended to perform shorter, higher-intensity activities and showed stronger metabolic improvements, while women demonstrated more consistent low-intensity movement with less pronounced biomarker changes. These results underline that responses to lifestyle interventions vary and that personalised, sex-specific recommendations are important.
The research is socially relevant because it demonstrates that wearable technology and AI can move beyond simple step counting toward meaningful health monitoring. By enabling accurate, real-world assessment of physical behaviour, it supports preventive healthcare strategies and lays the foundation for personalised health feedback systems that help older adults maintain independence and quality of life.
PhD dissertations
Approximately one week after the defence, PhD dissertations by Leiden PhD students are available digitally through the Leiden Repository, that offers free access to these PhD dissertations. Please note that in some cases a dissertation may be under embargo temporarily and access to its full-text version will only be granted later.
Press enquiries (journalists only)
General information
Beadle's Office
pedel@bb.leidenuniv.nl
+31 71 527 7211