Voxel-based morphometry to detect brain atrophy in progressive mild cognitive impairment

Abstract

Recent research has shown an increased rate of conversion to dementia in subjects with mild cognitive impairment (MCI) compared to controls. However, there are no specific methods to predict who will later develop dementia. In the present study, 22 controls and 56 MCI subjects were followed on average for 37 months (max. 60 months) and studied with magnetic resonance imaging (MRI) at baseline to assess changes in brain structure associated to later progression to dementia. Voxel-based morphometry (VBM) was used to investigate gray matter atrophy. During the follow-up, 13 subjects progressed to dementia. At baseline, no differences were detected in age or education between the control and MCI subjects, but they differed by several neuropsychological tests. The stable and progressive MCI subjects differed only by CDR sum of boxes scores and delayed verbal recall, which were also significant predictors of conversion to dementia. At the baseline imaging, the MCI subjects showed reduced gray matter density in medial temporal, temporoparietal as well as in frontal cortical areas compared to controls. Interestingly, the progressive MCI subjects showed atrophy in the left temporoparietal and posterior cingulate cortices and in the precuneus bilaterally, and a trend for hippocampal atrophy when compared to the stable MCI subjects. We conclude that widespread cortical atrophy is present already two and a half years before a clinical diagnosis of dementia can be set.