Website BristolUni University of Bristol
Closing date: 1st December
The University of Bristol seeks a PhD student to develop predictive digital biomarkers for dementia from wearable sensor data and machine learning, on a competition-funded PhD with Dr Dalila O’Grady in the School of Engineering Mathematics and Technology.
- Supervisor: Dr Dalila O’Grady, School of Engineering Mathematics and Technology
- Funding: Competition Funded PhD Project
- Eligibility: Students Worldwide
- Closing date: Tuesday 1st December 2026
About the Project
Dementia is often diagnosed only once cognitive symptoms become pronounced, by which time opportunities for early intervention are lost. Yet, biological and behavioral changes occur many years before clinical diagnosis. This project aims to develop predictive digital biomarkers for dementia by leveraging longitudinal data from wearable devices such as smartwatches and fitness trackers. These devices can capture continuous measures of activity, sleep, heart rate variability, and gait – features that may reflect subtle preclinical changes in brain health.
We hypothesise that multimodal data from wearables, analysed using advanced artificial intelligence and machine learning techniques, can detect dementia risk earlier and with greater accuracy than current clinical approaches. The research will follow a mixed-method design, combining analysis of existing dementia research cohort datasets (~1,000 participants) with a prospective pilot study of 50 at-risk individuals aged 60+. Participants will be monitored for six months using research-grade wearable devices, generating high-resolution behavioural and physiological data. These data will undergo preprocessing, feature extraction, and modelling with deep learning and multimodal integration techniques such as recurrent neural networks and transformers.
Validation will involve cross-cohort testing and external evaluation on pilot data, with performance assessed using standard metrics (AUC-ROC, sensitivity, specificity). In addition, stakeholder workshops with patients, carers, clinicians, and healthcare system partners will ensure clinical relevance, ethical acceptability, and translational potential.
Anticipated outcomes include the identification of reliable, non-invasive digital biomarkers for early dementia detection and an evidence base for incorporating wearable-derived monitoring into clinical practice. The impact could be substantial: empowering individuals and families with earlier support, reducing healthcare costs through delayed disease progression, and advancing national priorities around AI-enabled, community-based health care.
To apply for this job please visit www.findaphd.com.
