The challenge of describing subgroups is particularly important in vascular dementia, which, in contrast to more stereotypic processes affecting cognitive function, is better thought of as several syndromes rather than as a disease. Many current diagnostic descriptions lack a strong empiric basis. Some of the categories now in use suffer from a priori assumptions about causality and pattern associations, which themselves have not been validated. The so-called mixed dementia syndrome may have been underrepresented in our estimation of dementia subtypes, in comparison with so-called pure vascular causes. Within the vascular syndrome, whether seen in isolation or in combination with other causes of dementia, the relative contributions of white matter changes as compared with multiple cortical strokes needs to be clarified. It remains a matter of controversy as to whether prolonged or chronic intermittent cerebral ischemia is a statistically important part of the dementia. The variable relation between clinical presentation and neuroimaging localization has important consequences for understanding the pathophysiology of cognitive impairment arising from vascular causes. Recent data also suggest that we should focus away from both the Alzheimer disease model of dementia and the multi-infarct model of vascular dementia. There are important opportunities available to clinicians from many disciplines to collaborate in precise clinical descriptions of large numbers of patients to advance our understanding of the spectrum of vascular cognitive impairment.