The aim of this doctoral thesis was to study neuropsychological function in relation to structural and functional brain changes in Alzheimer´s disease (AD). In the first study relations between hippocampal volume, neuropsychological function and limbic-hypothalamic-pituitary-adrenal (LHPA) axis disturbances in AD were investigated with magnetic resonance imaging (MRI). Reduced hippocampal CA1 volume and suppressed cortisol levels in combination, best predicted the variation in neuropsychological performance. The conclusion was that reduced hippocampal volume and LHPA axis disturbances are associated to level of cognitive function in AD. The second study focused on whether patients with early AD showed an altered regional cerebral blood flow (rCBF) pattern compared to control persons, correlation between performance on memory tests and rCBF in sub-lobar volumes of the brain were investigated. The rCBF was measured with single photon emission computed tomography (SPECT). AD-patients showed a significantly lower rCBF in temporoparietal regions including left hippocampus compared to controls. The diagnostic sensitivity and specificity for AD was high in temporoparietal regions. AD-patients had significantly lower performance on semantic and, in particular, episodic memory-tests compared to the controls, and their performance on several episodic tests correlated with rCBF in parietal and temporal regions including left hippocampus, which suggest that abnormalities in the rCBF pattern underlie impaired episodic memory functioning in AD. The conclusion was that an observer-independent analyzing method for SPECT with sub-lobar volumes VOI´s is promising in the diagnosis of AD. In a third study possible differences in memory-related functional brain activation between persons with high versus low risk for AD were examined with functional magnetic resonance imaging (fMRI). The high-risk individuals performed worse than low-risk individuals on tests of episodic memory. Patterns of brain activity during episodic encoding and retrieval showed significant group differences. During both encoding and retrieval, the low-risk persons showed increased activity relative to a baseline condition in prefrontal and hippocampal brain regions that previously have been implicated in episodic memory. By contrast, the high-risk persons did not significantly activate any prefrontal regions, but instead showed increased activity in visual occipito-temporal regions. The conclusion was that patterns of prefrontal brain activity related to episodic memory differed between persons with high versus low risk for AD, and lowered prefrontal activity may predict subsequent disease. In a final study SPECT was used to map patterns of rCBF in an activated state (an episodic encoding task) and in a rest condition in persons with mild AD and in healthy elderly control persons. A reduction of rCBF in temporoparietal regions that was more pronounced in mild AD in the activated encoding task was observed. The conclusion was that there are rCBF differences between mild AD patients and healthy controls in temporoparietal regions, and the temporoparietal reduction is more pronounced during activation than during rest which might be important in the early diagnosis of AD. Taken together, these findings show that level of neuropsychological function, notably episodic memory, can be systematically related to functional disturbances in the LHPA axis and to the function of temporoparietal and prefrontal brain regions in AD patients. These changes are detectable in patients with risk for AD and in an early phase of AD which suggests that the obtained results might be important for early diagnosis of AD.