Parkinson's disease involves widespread neurodegeneration that extends far beyond the basal ganglia, giving rise to a diverse range of non-motor symptoms that frequently emerge before motor onset. These include autonomic dysfunction, cognitive decline, neuropsychiatric disturbances, sleep-related disorders, and sensory deficits. Here, we synthesize current evidence on the anatomical, neurochemical, and network-level mechanisms that drive these symptoms, and we examine how they shape disease progression and clinical heterogeneity. We highlight the limitations of dopamine-centric models and advocate for a framework that treats non-motor symptoms as the disorder's primary, mechanistically distinct features. We also discuss how emerging technologies-such as multi-omic profiling, artificial intelligence, and network neuroscience-enable earlier identification, stratification of non-motor phenotypes, and the development of precision-based therapeutic strategies. Recognizing non-motor symptoms as central to Parkinson's disease redefines how the disorder should be diagnosed, studied, and treated.