With rapid aging of the US population, understanding trends over time in dementia occurrence is essential to public health planning and intervention; this understanding includes trends in neuropathologies underlying clinical dementia. To characterize trends in pathways underlying dementia by examining prevalence of postmortem neuropathologies in birth cohorts across 25 years. Two longitudinal cohorts, the Religious Orders Study and the Rush Memory and Aging Project, with autopsy data from 1997 to 2022 with up to 27 years follow-up were analyzed. Deceased individuals with complete postmortem neuropathology evaluations were included, and 177 individuals with most distant (<1905) or recent (>1930) years of birth were excluded. Four categories of year of birth: 1905-1914, 1915-1919, 1920-1924, and 1925-1930. Outcomes included pathologic diagnosis of Alzheimer disease (AD), global AD pathology, amyloid load, tau tangles, neocortical Lewy bodies, limbic-predominant age-related TDP-43 encephalopathy neuropathological change, atherosclerosis, arteriolosclerosis, gross chronic infarcts, and chronic microinfarcts. For comparison, pathologies in each birth epoch were age-standardized to age distribution of the cohorts. χ2 Tests were used for categorical outcomes, and analysis of variance was used to compare means across birth epochs. Overall, 1554 participants were examined (510 [33%] male; median [range] age at death, 90 [66-108] years). Participants were distributed fairly evenly across birth epochs (1905-1914: n = 374; 1915-1919: n = 360; 1920-1924: n = 466; 1925-1930: n = 354). Across year of birth groups, no differences were found in prevalence of pathologic AD diagnosis; age-standardized prevalence fluctuated between 62% and 68% in the birth cohorts (χ2 test: P = .76 across birth epochs). Similarly, no differences were found in mean levels of global AD pathology, although there was greater density specifically of tau tangles in later birth cohorts (eg, age-standardized mean [SD], 1.53 [1.20] years for the 1905-1914 cohort and 1.87 [1.47] years for the 1925-1930 cohort; analysis of variance test: P = .01 across birth cohorts). There were no differences over time in other neurodegenerative pathologies. In contrast, atherosclerosis and arteriosclerosis were dramatically lower over time; for example, age-standardized prevalence of moderate to severe atherosclerosis ranged from 54% among those born from 1905-1914 to 22% for 1925-1930 (χ2 test: P < .001 across birth epochs). In this study, few differences in neurodegenerative pathologies were found, but there may be worse levels of tau tangles across birth cohorts over 25 years. This indicates that any improvements over time in clinical dementia observed by cohorts are likely in part associated with improved resilience to pathology rather than reduced AD pathology. Finally, vessel pathologies were markedly lower over birth cohorts, indicating the assocation with brain health of populationwide improvements in several vascular risk factors.