Central nervous system diseases are usually associated with significant modifications in brain morphometry and function. These alterations may be subtle, in particular at early stages of the disease progress, and thus not evident by visual inspection alone in magnetic resonance imaging. Group-level statistical comparisons have dominated neuroimaging studies for many years, leading to better insight into the patterns of regional vulnerability in brain neurodegenerative pathologies. However, such group-level results have no diagnostic value at the individual level. Recently, pattern recognition approaches using multivariate analyses have led to a fundamental shift in paradigm, aiming to predict the cognitive fate of each individual on the basis of MRI-based algorithms of structural parameters. We review here the state-of-the-art fundamentals of pattern recognition including feature selection, cross validation, and classification techniques and discuss limitations including interindividual variation in normal brain anatomy and neurocognitive reserve. We conclude with a special reference to future trends including multimodal pattern recognition and multicenter approaches with data sharing and cloud computing.
CITATION STYLE
Haller, S., Zanchi, D., Rodriguez, C., & Giannakopoulos, P. (2018). Brain structural imaging in Alzheimer’s disease. In Neuromethods (Vol. 137, pp. 107–117). Humana Press Inc. https://doi.org/10.1007/978-1-4939-7674-4_7
Mendeley helps you to discover research relevant for your work.