Differential alteration of neuronal network with FDG-PET in Alzheimer's disease and dementia with Lewy bodies

Objectives FDG-PET is useful for the diagnosis of major neurodegenerative disorders, such as Alzheimer's disease (AD) and dementia with Lewy bodies (DLB) based on the hypometabolic pattern in comparison with normal controls. We hypothesize that not only regional hypometabolic distribution, but also alteration of regional connectivity pattern such as default mode network may give us additional pathophyisiological and diagnostic information. The purpose of this study was to examine whether the regional connectivity is different among AD, DLB, and control using seed-based analysis of FDG-PET.

Methods FDG-PET and 3D T1-weighted MR images were obtained from 45 patients with AD, 18 with DLB, and 142 with cognitively normal (CN) control. PET images for standardized uptake value ratio (SUVR) against cerebellar cortex were generated using the statistical parametric mapping 8 (SPM8) with original scripts implemented in Matlab software. We conducted a seed-based correlation analysis based on the activity of posterior cingulate cortex (PCC) or occipital cortex (OC) with SPM8. Then, we added the regions of interest based analysis focusing the following specific areas, base of pons (BP), corpus geniculatum laterale, gyrus rectus (GR), hippocampus, lateral prefrontal cortex (LPF), medial prefrontal cortex (MPF), nucleus amygdaloideum (NA), nucleus dentanus (ND), and substantia nigra, with linear discriminant function analysis to describe those differences among AD, DLB and CN.

Results The correlations of PCC with GR, MPF, LPF, NA, or ND; OC with BP, GR, hippocampus, MPF, LPF, NA, or ND, were significantly different among AD, DLB, and CN.

Conclusions Neuronal network obtained from FDG-PET is suggested to be differently altered among AD, DLB, and CN. That information may be useful for the pathophysiological understanding of the disorders and applicable for the differential diagnosis.