Abstract
The in vitro techniques of proton induced x-ray emission (PIXE), Rutherford backscattering (RBS), instrumental neutron activation analysis (INAA) and the in vivo technique of positron emission tomography (PET) have been employed to examine and compare elemental concentrations with metabolic rates within the brain. The element analysis study builds upon work previously performed at the University of Surrey whereby trace element concentrations were compared between 'normal' and Alzheimer disease (AD) brain states. Additionally to these aspects, differences between hemispheres and gender were investigated between the disease states for the frontal, occipital and parietal regions. Brain tissue samples used in analysis were obtained from the Alzheimer's Disease Brain Bank, Institute of Psychiatry, London. In the frontal lobe concentrations of phosphorus and iron were seen to have different concentrations between hemispheres with higher values in the right hemisphere 'normal' brain and lower concentrations in the right hemisphere AD brain. Concentrations were significantly lower in the right hemisphere of AD subjects although these concentrations did decrease between AD and 'normal' states. Chlorine and calcium were increased in AD subjects. The differences between hemispheres implies that the hemispheres are affected differently with the onset of dementia. Elements enter the brain by passing through the blood brain barrier (BBB) and any change in its structure will alter concentrations along with uptake of glucose and hence metabolic rates. PET using the tracer 18F-fluorodeoxyglucose (18F-FDG) was employed to preliminary investigate region, gender and age variations of metabolism within the AD brain. Eighteen main regions of interest were chosen for each hemisphere of eighteen subjects, nine male and nine female scanned at the Paul Scheixer Institute (PSI) Villigen Switzerland. Differences between hemispheres, brain regions and gender were observed within the AD subject group. Throughout disease duration metabolic rates were found to decrease significantly in the majority of lobes over 2 years in the one female subject who was followed by taking 3 PET scans at intervals over the period. The temporomedial region was one exception to this trend where metabolic rates increased over the first year and decreased thereafter; temporomedial vulnerability is much quoted in AD. To compare AD subjects with a baseline a control group of seventy two subjects of age range 22 to 82 years was examined to determine gender, age and sex differences in the 'normal' brain. Up to the age of 60 years a low steady decline in the regional cerebral metabolic rate of glucose (rCMRGlu) was observed but above the age of 60 a considerably higher rate of decrease was found. Gender differences were also discovered. Between hemispheres significances in rCMRGlu were also observed to vary with increased age with the youngest subject group having higher metabolic rates in the left hemisphere while this dominance decreased to no significance with the increasing age of groups. These results have clear implications with regards subject selection in any future group study as subjects must be age and gender matched to as high a degree as possible. The primary objective of this work was to determine whether variations in elemental concentrations between 'normal' and AD brains could be correlated with glucose metabolic rate changes within the same regions. The general change in elemental concentrations between disease states was observed to decrease more so in female brains than male. Metabolic rates also decreased at a greater rate in female subjects, so suggesting a possible relationship. Statistical techniques of analysing data are also studied with the result that different ways of reconstructing data in terms of glucose metabolic rates resulted in different absolute values and significances between brain regions. Two normalisation routines also resulted in significantly different absolute metabolic rates. Additionally the techniques of hierarchical clustering, statistical parametric mapping (SPM) and subprofile scalar modelling (SSM) were used to analyse these absolute values, each with its own interpretation of data and so each suggesting unique conclusions. The latter method suggests changes in neuronal connectivity between the control and AD brain.