A brief history of voxel-based grey matter analysis in Alzheimer's disease

Voxel-based morphometry (VBM) and cortical thickness measurement are common techniques to identify regional atrophy in neurodegenerative diseases such as Alzheimer's disease (AD). Because studies employing these methods draw conclusions regarding patterns of regional cortical degeneration, it is important to be aware of their possible limitations. To evaluate the effect of different VBM versions, we performed voxel-based analyses through successive versions - from SPM99 to SPM8 - as well as FSL-VBM on n = 20 AD patients and n = 20 controls. Reproducibility was assessed in an independent sample, again of n = 20 per group, from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Further, we tested the hypothesis that VBM can sensitively detect hippocampal atrophy, but is relatively insensitive to changes in the cortical ribbon, by contrasting VBM with FreeSurfer cortical thickness measurements. The results with both datasets confirmed that VBM preferentially identifies grey matter lesions in the mesial temporal lobe but is largely insensitive to isocortical atrophy. In contrast, FreeSurfer identified thinning of cortical ribbon association cortex more significant in post- rather than pre-Rolandic areas and with relative preservation of primary sensory-motor regions - in other words precisely as would be expected in AD. The results highlight a bias that VBM has toward detecting mesial temporal lobe atrophy. This finding has important implications for interpretation of clinical and cognitive studies in AD