Global cognitive impairment should be taken into account in SPECT-neuropsychology correlations: the example of verbal memory in very mild Alzheimer's disease.

Purpose: To examine the impact of severity of global cognitive impairment on SPECT-neuropsychology correlations, we correlated a verbal memory test with brain perfusion in patients with very mild Alzheimer's disease (AD), taking into account the Mini-Mental State Examination (MMSE) score as an index of global cognitive impairment.
Methods: Twenty-nine outpatients (mean age 78.2+/-5.5 years) affected by very mild, probable AD underwent brain SPECT with 99mTc-ethylcysteinate dimer and a word list learning test. SPM99 was used for voxel-based correlation analysis after normalisation to mean cerebellar counts (height threshold: p<0.01). In a first analysis, only age and years of education were inserted as nuisance covariates, while in a second analysis the MMSE score was inserted as well.
Results: In the first analysis, two clusters of significant correlation were found in both hemispheres, mainly including regions of the right hemisphere, such as the inferior parietal lobule, the middle temporal gyrus and the posterior cingulate. Significant correlation in the left hemisphere was observed in the lingual lobule, the parietal precuneus and the posterior cingulate. After taking into consideration the MMSE, the largest cluster of correlation was found in the left hemisphere, including the parietal gyrus angularis, the posterior cingulate and the middle temporal gyrus.
Conclusion: The wide differences observed between the correlations achieved with and without taking into account the MMSE score indicate that severity of global cognitive impairment should be considered when searching for brain perfusion-neuropsychology correlations. In the present case, this strategy resulted in correlations that more closely matched neuropsychological models of verbal memory deficit.