Parkinson’s disease (PD) is the most common neurodegenerative movement disorder, but current diagnostic criteria identify individuals late, impeding the development of neuroprotective therapies: There is the critical yet unmet need to identify the disease onset timely and to define the inception of pathology. Advancing recognition of early motor deterioration and coupling it with biological evidence of pathological α-synuclein will overcome present barriers and tackle the questions when and where PD starts. Isolated REM sleep behavior disorder, confidently detectable by polysomnography, signals an early α-synucleinopathy. Overseeing a large cohort, I will integrate multimodal motor data on speech, facial expression, hand movements, gait, and long-term actigraphy within a machine learning framework, permitting a tablet-based application for highly mobile uses. This earliest PD motor-fingerprint, common to PD in general, will enable prospective large-scale screening from an unselected population. Motor phenotype detection will be paired with biological confirmation of pathological α-synuclein species in the cerebrospinal fluid using precise seed aggregation assays (SAA), capturing imminent PD earlier than ever before. This universal identification of the earliest phase of PD will allow to explore the initial spread of α-synuclein pathology, elucidating variations of pathology initiation: I will employ α-synuclein SAA on biosamples from the olfactory epithelium (brain-first) and the peripheral autonomic nervous system (body-first) to decode fundamental subtypes. Re-Start PD pairs general PD characteristics - its primal motor phenotype and its biological α-synuclein signatures - outlining the earliest phase of PD prospectively and being readily transferable for broad application to kickstart timely therapeutic innovations. Complementary biobanking, brain-imaging, and follow-up of this unique cohort will foster further biomarker development and subtype determination.