Probabilistic brain atlases in health & disease

Brain imaging has become an important tool to define tissue changes and atrophy related to neurodegenerative and cerebrovascular processes, which determine cognitive decline and dementia. Periventricular and deep white matter hyperintensities (WMHs) on T2-W magnetic resonance imaging (MRI) have been one of the most investigated changes ascribed to cerebrovascular disease (CVD). Meta-analysis suggests increased WMHs are associated with cognitive impairment. Pathological changes associated with WMHs include arteriolosclerosis, microvascular pathology, lacunar infarcts, microinfarcts, perivascular spacing, myelin loss and axonal damage. Diffusion tensor imaging on MR has also been useful to detect changes in white matter fibre tracks associated with cognitive impairment in CVD. White matter rarefaction is common in CVD pathology but it may not be consistently differentiated from normal appearing white matter. Neuroimaging is not yet sufficiently sensitive to detect individual cell e.g. glia changes even with tracer studies with positron emission tomography. Gradient Echo (GRE) or T2* MRI relates to both lobar and subcortical locations of microhaemorrhages or hemosiderin accumulation on pathology. These changes appear increased in CVD but they are not consistently correlated with impairment. The radiological signature characteristic of cerebral amyloid angiopathy (CAA) comprises multiple lobar microbleeds on GRE, posterior dominant white matter hyperintensities on MRI FLAIR, dilated perivascular spaces in white matter areas on T1-W and multiple areas of superficial siderosis on GRE. However, this signature is not entirely specific for CAA. Incorporating the hippocampal formation, medial temporal lobe atrophy (MTLA) is widely regarded as a key biomarker of Alzheimer disease but MTLA or hippocampal atrophy has also been described in vascular dementia and post-stroke dementia. Hippocampal volumes could be reduced by as much as 15-20% in subcortical vascular dementia relative to normal ageing controls. These findings relate to pyramidal neuron atrophy within the hippocampus proper and loss of white matter in the temporal limb. These likely results from reductions in blood flow in the posterior cerebral artery territories. In summary, various radiological features associated with CVD are now increasingly recognised and correlated with cognitive function during ageing. While neuropathological examination allows more scrutiny of tissue changes as they relate to cognitive function it is apparent that current advances in neuroimaging do not permit high specificity.