Regional cerebral blood flow as assessed by principal component analysis and (99m)Tc-HMPAO SPET in healthy subjects at rest: normal distribution and effect of age and gender.

ABSTRACT

The increasing implementation of standardisation techniques in brain research and clinical diagnosis has highlighted the importance of reliable baseline data from normal control subjects for inter-subject analysis. In this context, knowledge of the regional cerebral blood flow (rCBF) distribution in normal ageing is a factor of the utmost importance. In the present study, rCBF was investigated in 50 healthy volunteers (25 men, 25 women), aged 31-78 years, who were examined at rest by means of single-photon emission tomography (SPET) using technetium-99m d, l-hexamethylpropylene amine oxime (HMPAO). After normalising the CBF data, 27 left and 27 right volumes of interest (VOIs) were selected and automatically outlined by standardisation software (computerised brain atlas). The heavy load of flow data thus obtained was reduced in number and grouped in factors by means of principal component analysis (PCA). PCA extracted 12 components explaining 81% of the variance and including the vast majority of cortical and subcortical regions. Analysis of variance and regression analyses were performed for rCBF, age and gender before PCA was applied and subsequently for each single extracted factor. There was a significantly higher CBF on the right side than on the left side ( P<0.001). In the overall analysis, a significant decrease was found in CBF ( P=0.05) with increasing age, and this decrease was particularly evident in the left hemisphere ( P=0.006). When gender was specifically analysed, CBF was found to decrease significantly with increasing age in females ( P=0.037) but not in males. Furthermore, a significant decrease in rCBF with increasing age was found in the brain vertex ( P=0.05), left frontotemporal cortex ( P=0.012) and temporocingulate cortex ( P=0.003). By contrast, relative rCBF in central structures increased with age ( P=0.001). The ability of standardisation software and PCA to identify functionally connected brain regions might contribute to a better understanding of the relationships between rCBF at rest, anatomically defined brain structures, ageing and gender.