Cerebral activation and deactivation during navigation in a virtual reality environment in patients with MCI. An O15-water PET activation study

Background and aims: In patients with Alzheimer's disease (AD), characteristic abnormalities of cerebral activation during cognitive performance have been previously observed. The aim of the current study was to assess the cerebral activation and deactivation patterns during navigation in a virtual reality (VR) environment in healthy volunteers and patients with mild cognitive impairment (MCI) and to correlate the cerebral activation with navigation speed. Methods: A healthy control group (n=11, right handed, 3 female, age 67+/-10 years) and a matched group of 11 MCI-patients (4 female, 71+/-9 years) underwent a navigation task in a computer-based 3D VR-labyrinth, based solely on visual clues. During active navigation 8 O15-water PET-scans were performed (370 MBq i.v.-bolus). In each subject navigation speed (time from start to destination) was measured. In addition, an FDG-PET-scan at rest condition was performed in all subjects. Statistical analysis (correlation analysis of O15 water activation with navigation speed and group comparison of O15-water and F-18 FDG PET scans) was carried out, using SPM02 software (Wellcome Inst., London, UK), p<0.05, FDR-corrected. Results: Analysis of FDG PET-data revealed a significantly lower metabolism exclusively in the posterior cingulate cortex in the MCI-group compared to controls. Controls showed a significantly better navigation performance (required time: 67±46 sec, MCI-group 139±53 sec). Analysis of O15 water PET showed a significantly different pattern of activation in MCI and controls. Stronger activations in controls were found in the left cerebellum, the occipital lobe (Brodman area [BA] 18,19) and the sensorimotor cortex (BA 3,4) and in MCI in the prefrontal cortex (BA 9), the SMA (BA 6) and the anterior cingulate cortex (BA 32). Relevant deactivation of auditory cortical regions (BA 41,42,22) ?representing cross-modal inhibition- was detected exclusively in controls. In addition different patterns of correlation between navigation speed and regional brain activation were found in the two groups. Navigation speed correlated with activation in left cerebellum, left premotor cortex (BA 6) and right thalamus in controls and in MCI with activation in left temporal cortex (BA 20), left posterior cingulate cortex (BA 31) and right cerebellum. Conclusions: Several cortical key regions have already been associated with human navigation in healthy subjects. In our study MCI showed activations differing from the activations found in controls during navigation. Also the deactivation of task-irrelevant cerebral regions observed in controls, was absent in MCI. This suggests that MCI is characterized by a progressive impairment of cross-modal cerebral deactivation-functions, resulting in a decreased ability to direct attention primarily to the relevant cognitive modality. In addition, a correlation analysis revealed a significantly different network associated with successful navigation-performance in MCI compared to controls. These results might indicate compensatory changes or functional deficits in MCI due to ongoing neurodegeneration.