Population-based disease registers identify and characterize all cases of disease, including those that might otherwise be neglected. Prospective population-based registers in neurodegeneration are necessary to provide comprehensive data on the whole phenotypic spectrum and can guide planning of health services. With the exception of the rare disease amyotrophic lateral sclerosis, few complete population-based registers exist for neurodegenerative conditions. Incomplete ascertainment, limitations and uncertainty in diagnostic categorization, and failure to recognize sources of bias reduce the accuracy and usefulness of many registers. Common biases include population stratification, the use of prevalent rather than incident cases in earlier years, changes in disease understanding and diagnostic criteria, and changing demographics over time. Future registers are at risk of funding shortfalls and changes to privacy legislation. Notwithstanding, as heterogeneities of clinical phenotype and disease pathogenesis are increasingly recognized in the neurodegenerations, well-designed longitudinal population-based disease registers will be an essential requirement to complete clinical understanding of neurodegenerative diseases.