FIGURE 5.

Detection of early regional amyloid accumulation in cognitively healthy older individuals. Effect of partial volume correction (PVC) method on first detectable signal increases in voxel-wise 18F-Florbetapir-PET SUVR_WC values as determined by the sliding-window analysis with decreasing CSF Aβ42 levels in cognitively healthy controls. Clusters of >150 voxels passing a family wise error corrected level of significance of p < 0.05 are plotted on rendered cortical surfaces and a midsagittal brain section. Effects are shown for the index sample (group mean CSF Aβ level of 840 pg/ml) where first regional signal increases compared to the reference sample (group mean CSF Aβ level of 2,512 pg/ml) were detectable in standard uncorrected 18F-Florbetapir-PET data (‘no PVC’). Across PVC methods regional effects appear in similar locations but are much more widespread in PVC-3 data (last row) compared to no PVC (first row) or PVC-2 data (second row).