The progressive development of Alzheimer disease (AD)-related lesions, such as neurofibrillary tangles (NFT), amyloid deposits and synaptic loss, and the occurrence of microvascular and small macrovascular pathology within the cerebral cortex are conspicuous neuropathologic features of brain aging. Recent neuropathologic studies strongly suggested that the clinical diagnosis of dementia depends more on the severity and topography of pathological changes than on the presence of a qualitative marker. However, several methodological problems, such as selection biases, case-control design, density-based measures and masking effects, of concomitant pathologies persisted. In recent years, we performed several clinicopathologic studies using stereological counting of AD lesions. In order to define the cognitive impact of lacunes and microvascular lesions, we also analyzed pure vascular cases without substantial AD pathology. Our data revealed that total NFT numbers in the CA1 field, cortical microinfarcts and subcortical gray matter lacunes were the stronger determinants of dementia. In contrast, the contribution of periventricular and subcortical white matter demyelinations had a modest cognitive effect even in rare cases with isolated microvascular pathology. Importantly, in cases with pure AD pathology, more than 50% of Clinical Dementia Rating scale variability was not explained by NFT, amyloid deposits and neuronal loss in the hippocampal formation. In cases with microvascular pathology or lacunes, this percentage was even lower. The present review summarizes our data in this field and discusses their relevance within the theoretical framework of the functional neuropathology of brain aging and with particular reference to the current efforts to develop standardized neuropathological criteria for mixed dementia.