AI-Derived Temporal and Behavioural Biomarkers of Chronic Disease Progression

Our research group is interested in joining Horizon Europe consortia developing next-generation predictive biomarkers for chronic non-communicable diseases using artificial intelligence.

We focus on a complementary class of biomarkers that remains substantially underexplored: temporal and behavioural biomarkers. Rather than relying exclusively on molecular measurements, we investigate how longitudinal changes in biological rhythms, sleep, lifestyle, behaviour and routine clinical parameters can serve as early predictors of disease progression, treatment response and multimorbidity.

Our concept is based on the premise that chronic diseases evolve over time rather than through isolated biological events. Repeated observations of sleep characteristics, circadian organisation, physical activity, dietary behaviour, treatment adherence, patient-reported outcomes, cardiovascular risk factors and routine clinical measurements generate dynamic health trajectories that may contain predictive information not captured by conventional biomarkers.

Using artificial intelligence, these multidimensional longitudinal data can be integrated into composite digital biomarkers capable of identifying patients at increased risk of disease progression before irreversible clinical deterioration becomes apparent. Such biomarkers may support earlier intervention, personalised prevention, optimisation of treatment strategies and improved patient stratification.

Our particular interest lies in chronic diseases affecting the cardiovascular, renal, metabolic and musculoskeletal systems, where multimorbidity is common and disease trajectories are highly heterogeneous. We believe that temporal behavioural phenotyping represents an important missing component of precision medicine and should complement molecular, imaging and laboratory biomarkers rather than replace them.

Our team contributes expertise in internal medicine, cardiovascular medicine, chronobiology, sleep health, lifestyle medicine, behavioural epidemiology and longitudinal health assessment. We have experience in multidimensional questionnaire development, clinical phenotyping, ambulatory blood pressure monitoring, assessment of circadian and behavioural characteristics and population-based research. These competencies provide an excellent foundation for developing clinically meaningful AI-ready datasets suitable for predictive modelling.

We seek collaboration with partners specialising in artificial intelligence, machine learning, digital biomarkers, wearable technologies, clinical informatics, systems medicine, biostatistics, biomarker validation and implementation science. We are particularly interested in developing explainable AI models that transform temporal clinical and behavioural data into robust predictive biomarkers applicable in routine healthcare.

Our role within a consortium could include clinical data acquisition, longitudinal patient phenotyping, behavioural and circadian assessment, development of temporal health indicators, validation of digital biomarkers in real-world populations and clinical interpretation of AI-generated prediction models.