Personalized Parkinson Project (PPP)

The Personalized Parkinson Project (PPP, Dutch name "Parkinson op Maat") is a prospective, single-centre cohort run by Radboud university medical center in Nijmegen together with Verily. It follows people in the early years of Parkinson's disease to describe how the disease progresses and to search for biomarkers that predict it. Brain MRI is one of several data streams, next to clinical tests, biospecimens and continuous smartwatch recordings. The data are released to qualified researchers after a reviewed project proposal.

Composition

The data sharing website lists 517 participants with Parkinson's disease. The design paper had planned 650. Inclusion required age 18 or older and a disease duration of at most five years. Each participant has three study-centre visits (baseline, one year, two years), followed by yearly questionnaires for up to five more years. MRI is part of the baseline and two-year visits. Two related cohorts, PPP De Novo (103 people with Parkinson's disease) and PPP PSP (50 people with progressive supranuclear palsy and 50 controls), are shared through the same procedure but are not counted here.

Acquisition

Scans are acquired on a single Siemens MAGNETOM Prisma 3 T scanner with a 32-channel head coil at the Donders Center for Cognitive Neuroscience. Published PPP analyses report T1-weighted MPRAGE, multiband diffusion imaging with b-values up to 2000 s/mm² and task fMRI (a visuomotor action-selection task and a reinforcement learning task). The study protocol also planned T2-weighted, FLAIR, quantitative susceptibility mapping and resting-state fMRI scans, and surface EMG during scanning to track tremor. Other data include motor and neuropsychological assessments in ON and OFF medication states, blood, stool and cerebrospinal fluid, 12-lead and Holter ECG, questionnaires and Verily Study Watch sensor data.

Known limitations

All imaging comes from one scanner at one site in the Netherlands. The website does not publish how many participants have MRI at each visit; one task fMRI analysis drew on 367 patients with baseline scans. Some diagnoses were revised at follow-up, so a few participants turned out not to have Parkinson's disease. Data are released at cost, through a pseudonymised PEP repository, and may not be shared beyond the approved project.