Published online by Cambridge University Press: 17 February 2010
Clinicians are increasingly faced with the problem of interpreting subtle, early cognitive symptoms. Enhanced awareness of Alzheimer's disease (AD) and available treatments has led to a growing demand for early assessment. Although it is known that a proportion of individuals with mild cognitive impairment will progress to dementia in following years, our ability to identify these individuals and predict individual cognitive trajectories is limited. The emergence of disease-modifying treatments would make these problems more acute. In this review, the potential role of magnetic resonance imaging (MRI) in aiding the clinician in early diagnosis of AD will be considered. The changes in grey matter structure that accompany ‘normal’ ageing will be described briefly, before moving on to studies that have attempted to distinguish the onset of disease from this background of structural change. Volumetric methods range from measurements of single key structures, such as the hippocampus, to methods based on computational neuroanatomy, which evaluate subtle structural alterations across the whole brain simultaneously. Computational methods are rapidly evolving and already perform as well as radiologists in distinguishing AD from normal ageing at an individual level. This article aims to provide a practical knowledge of how and why these methods work, point out the main advantages and disadvantages and sketch out outstanding issues and possible future directions.