A multi-centre proof of concept (PoC) prospective cohort study to externally validate the PD risk assessment model and evaluate user acceptance of the mAI-Health and mAI-Insights (for PD screening) apps by persons with suspected PD and (non-)expert HCPs. Objectives The primary objective: to evaluate the superiority of the PD risk assessment model (and associated tool) developed in terms of PD risk estimation compared to the MDS research criteria for prodromal PD, considered as the standard-of-care comparator. The secondary objectives of the study: to evaluate the reproducibility and generalisability of the PD risk assessment model; to evaluate changes in a multimodal battery of diagnostic biomarkers; to validate the top-ranking Single Nucleotide Polymorphisms identified to be related to PD risk during polygenic risk score analysis of multiple biobanks for determining the genotypic input to the predictive model. This study is a non-interventional observational study. The present study aims to validate a PD risk prediction model that uses as input a person’s phenotypic, clinical, and genetic information, as well as digital biomarker measurements in daily living to estimate the risk of PD.

One of the objectives of the AI-PROGNOSIS project is to develop a system for objective tracking of key Parkinson’s disease (PD) risk and progression markers in users’ daily living. To this end, novel digital biomarkers (dBMs) will be employed, derived from data generated by smartphone and smartwatch sensors. In the dBM-DEV study, new dBMs will be developed and validated, while existing ones will be verified. Objectives The primary objective: to develop and validate dBMs for REM sleep behaviour disorder (RBD). The secondary objectives: to validate the dBMs that will be developed to track daytime somnolence; to verify existing dBMs for tracking bradykinesia, rigidity, rest tremor, dyskinesias, postural stability, and cognitive performance; to evaluate the performance of the dBMs correcting for age and sex.

A multi-centreproof of concept (PoC), interventional, randomised, controlled open-label parallel-group study to externally validate the PD progression prediction model and the medication response prediction model, produce preliminary utility evidence regarding the former, and evaluate the usability of the mAI-Care and mAI-Insights (for PD care) apps by persons with PD (PwP) and expert HCPs. Objectives The primary objective: to examine the adequacy of the PD progression prediction model by comparing the disease progression rate predicted by the model at baseline and over the course of 12 months against the actual disease progression empirically observed over the course of 12 months relative to baseline. As secondary objectives, the adequacy of the predicted disease aggravation will be further corroborated through the following observations at the 12-month visit: occurrence of motor complications in patients diagnosed in the last 5 years, or occurrence of troublesome motor complications in all patients; occurrence of at least two of the following disease severity milestones: cognitive dysfunction, hallucinations, freezing, dysphagia; increase of the Levodopa Equivalent Dose. Further secondary objectives include: validation of non-genetic factors found to be associated with PD progression during the development of the predictive model; validation of the top-ranking Single Nucleotide Polymorphisms identified to be associated with PD progression during analysis of multiple biobanks for determining the genotypic input to the predictive model; evaluation of the mAI-Care and mAI-Insights tools by end-users through the System Usability Scale score, scores of user satisfaction questionnaires and metrics of user engagement with the tools. The study primarily aims to validate a PD progression model that uses as input a person’s phenotypic, clinical, and genetic information, as well as digital biomarker measurements in daily living to predict PD progression. The study also aims at creating a proof of concept regarding the validity of a medication response prediction model.