The 'European Prospective Investigation into Cancer and Nutrition' cohort (EPIC) is a prospective study including ~ 520,000 participants recruited across Europe (1991–2000) with in-depth baseline data on nutritional, lifestyle, medical, and anthropometric variables, and baseline blood samples. Here we introduce EPIC4ND, a case-cohort study within EPIC designed to identify biomarkers predicting a future onset of dementia, Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS). EPIC4ND comprises 6415 initially non-diseased participants (aged 35–80 years, mean age at baseline: 54 ± 9, 64% women) including 1899 incident cases with up to 30 years of follow-up and data on at least one omics domain available from pre-disease blood samples. EPIC4ND includes 4604 subcohort members (4441 non-cases and 163 incident cases) and 1811 additional incident cases ascertained from the broader EPIC cohort. Among the incident cases, there are 1190 dementia cases (818 AD), 610 PD cases, and 199 ALS cases. Additionally, 72 prevalent PD cases and 118 incident Parkinsonism cases are available for comparison. Molecular data generated encompass proteomics, genome-wide DNA methylation, and SNP genotyping with 4127 EPIC4ND participants (including 1635 incident cases) having data on all three domains. Smaller studies include data on metals, metabolites, and environmental chemicals, while ongoing efforts focus on ultrasensitive targeted biomarker measurements and small RNA sequencing. Genome-wide association studies and analyses of epidemiological risk factors validate the dataset by confirming many known risk factors. Leveraging these extensive pre-disease multi-layered omics data offers a unique opportunity to identify biomarker signatures predicting neurodegenerative diseases and to explore their interplay with epidemiological risk factors.