Advancing Digital Health: Early Detection of Cognitive and Physical Decline in Ageing Populations

Age-related declines in cognitive and motor function often emerge subtly yet progressively, impacting mobility, daily activities, and overall quality of life. Early identification of these changes is essential for timely intervention and the prevention of more severe outcomes such as falls, behavioural changes, loss of independence, and accelerated cognitive impairment. With the global population aged over 65 projected to double by 2050, there is an urgent need for computational approaches that can predict, monitor, and mitigate age-related decline. This project addresses this challenge by leveraging artificial intelligence to develop data-driven digital health solutions to support independent and healthy ageing.

The project focuses on the application of computational intelligence to model and predict health deterioration using multimodal physiological data. It will investigate how measurable indicators such as gait dynamics, grip strength, cardiovascular response, and body composition can be transformed into predictive features for early detection of decline. The aim is to move beyond episodic clinical assessments towards continuous, AI-IoT-based monitoring systems capable of identifying subtle, pre-clinical signals and enabling proactive, personalised intervention.

The research is fundamentally computational, addressing key challenges in modelling heterogeneous, high-dimensional, and longitudinal health data. It will explore how temporal and spatial dependencies in physiological signals can be captured, and how predictive models can remain both accurate and interpretable in real-world clinical settings. To achieve this, the candidate will design, develop, and validate hybrid, adaptive deep learning models integrating multimodal sensing data. These models will evolve from interpretable baseline methods to advanced architectures for time-series and multimodal learning, incorporating hybrid deep learning approaches with techniques to handle uncertainty and missing data.

The candidate is expected to work with large-scale cohort datasets from UK Biobank, ELSA, and NACC, complemented by controlled experimental data, enabling both methodological development and real-world validation. The research will emphasise multimodal data fusion, temporal modelling of ageing trajectories, and the development of reproducible, scalable machine learning pipelines using modern data science tools and high-performance computing environments.

The expected deliverables include a suite of validated AI models capable of early risk detection, novel biomarkers tailored to the needs of ageing populations, high-impact academic publications, and prototype decision-support tools for clinical and community use. These outputs will be supported by reproducibility protocols, open-access dissemination, and engagement with rehabilitation experts to guarantee real-world applicability.

This PhD project is ideally suited to candidates with a strong background in Computer Science, Data Science, Artificial Intelligence, or a related field, particularly those with experience in machine learning, big data analytics, and an interest in applying advanced computational methods to real-world healthcare challenges.

1. Ashraf, M., Anowar, F., Setu, J.H., Chowdhury, A.I., Ahmed, E., Islam, A. and Al-Mamun, A., 2023. A survey on dimensionality reduction techniques for time-series data. IEEE Access, 11, pp.42909-42923.

2. Vlaić, M., Mikulić, I., Delač, G., Šilić, M. and Vladimir, K., 2025, June. A Review of Time Series Dimensionality Reduction Methods. In 2025 MIPRO 48th ICT and Electronics Convention (pp. 149-154). IEEE.

3. Tanwar, R., Phukan, O.C., Singh, G., Pal, P.K. and Tiwari, S., 2024. Attention based hybrid deep learning model for wearable based stress recognition. Engineering Applications of Artificial Intelligence, 127, p.107391.

4. Chen, P., Dong, W., Wang, J., Lu, X., Kaymak, U. and Huang, Z., 2020. Interpretable clinical prediction via attention-based neural network. BMC Medical Informatics and Decision Making, 20(Suppl 3), p.131.

5. Petroșanu, D.M., Pîrjan, A. and Tăbușcă, A., 2023. Tracing the influence of large language models across the most impactful scientific works. Electronics, 12(24), p.4957.