AI-PROGNOSIS is developing predictive models to estimate the risk of developing Parkinson’s disease (PD), track its progression, and assess response to treatment.

Using multimodal data – including genetics, wearable device signals, and large-scale datasets (PPMI, AMP-PD) – the project has already harmonised core data, applied bias mitigation, and improved model robustness and fairness.

Early results show:

• Genetic variants linked to PD risk, with further validation needed

• Wearable-derived digital biomarkers that distinguish PD from healthy controls

• Promising models for predicting one-year progression and treatment side effects

  
  

These advances mark important steps towards trustworthy, AI-driven solutions that can support clinicians and improve personalised PD care.

Read the full report on Zenodo: https://zenodo.org/records/15542465