SYNTHIA at the 2nd Bonn Conference on Mathematical Life Sciences

Diego's presentation will focus on a model developed in SYNTHIA that can generate realistic multimodal longitudinal patient data, even when different types of data are collected at varying time intervals. This capability is particularly important in Alzheimer’s disease research, where patient data often come from multiple sources and follow different timelines.

A key feature of the model is its ability to simulate “what-if” scenarios. This allows researchers to explore how a patient’s disease trajectory might change under different hypothetical interventions. To make these simulations meaningful, the model integrates a causal inference framework, enabling a better understanding of cause-and-effect relationships in disease progression.

Initial results demonstrate that the generated synthetic data are consistent and realistic, supporting the validity of the approach. This work contributes to SYNTHIA’s broader goal of developing trustworthy synthetic data methodologies that can support research in Alzheimer’s disease while respecting privacy and enabling safer experimentation with sensitive health data.