Neuroimaging is increasingly used to aid diagnosis in dementia. The traditional view that imaging is important solely as means of excluding treatable causes of dementia is maintained by many guidelines. These conditions however, account for a tiny proportion (<1%) of all causes of dementia. Over the past few years it has been recognised that a more accurate diagnosis and prognosis is important for patients and their families. The different pathological processes that produce cerebral dysfunction at a cellular level also produce macroscopic effects that can be detected in vivo with imaging. Clinically useful measures that distinguish between neurodegenerative disorders at an early stage are still awaited. The most likely future use of structural imaging will be the identification of patients at risk for Alzheimer's disease or with preclinical Alzheimer's disease. For magnetic resonance imaging (MRI) this will mean focusing on those areas that are affected earliest in the disease; ie, entorhinal cortex and hippocampus, using high resolution structural MRI or sophisticated brain mapping techniques. Imaging research is also likely to focus on measuring progression and detecting therapeutic effect. As such, MRI is already become an indispensable tool in clinical trials in dementia.