protocols. Participants were grouped by sex and APOEε4 (carrier vs. non- carrier). We used quantile regression to estimate the median amyloid PET, HVa or memory value as a function of age. We fit a single model that included age, sex, and APOEε4 status along with all two-way interac- tions. Results: Overall, memory performance worsened gradually from 30 to 70 years and more steeply above 70. HVa was relatively steady from 30 to 60 years but worsened noticeably above age 60. Median amyloid PET was low through approximately age 65 and increased beyond that age in all groups. Men performed worse than women from age 40-50 onward on mem- ory. HVa was lower in men than women beyond age 60. Therewas no sex difference in amyloid PETat any age. Within each sex, memory and HVa were not different by APOE at any age, except for worse memory among women carriers beyond their 70s. APOE ε4 increased the proportion of abnormal amyloid PET values in those above age 65, but had no impact prior to that age. Individuals with abnormal amyloid PET values first appeared around age 65 years in both APOE ε4 carriers and non-carriers. Conclusions: Advancing age and male sex had more substantial impacts on memory and HVa than APOEε4 in cognitively normal individuals across the adult lifespan (most from a population based sample). In addition, wors- ening memory and HVa preceded abnormal b -amyloidosis at the population level. We conclude that the onset of b -amyloidosis typically arises on a background of preexisting functional decline and structural degeneration that are unrelated to fibrillar b -amyloid deposition; therefore the onset of b -amyloidosis in late life likely accelerates rather than initiates memory dysfunction and brain atrophy. O3-10-04 PRECLINICAL EFFECTS OF Ab DEPOSITION ON EPISODIC MEMORY AND DISEASE PROGRESSION Victor L. Villemagne 1 , Samantha Burnham 2 , Belinda Brown 3 , Pierrick Bourgeat 4 , Kathryn A. Ellis 5 , Paul Yates 6 , Robyn Veljanovski 1 , Ralph Martins 7 , Lance Macaulay 8 , Olivier Salvado 4 , David Ames 9 , Colin Louis Masters 10 , Christopher C. Rowe 1 , 1 Austin Health, Melbourne, Australia; 2 CSIRO, Perth, Australia; 3 Edith Cowan University, Perth, WA, Australia; 4 CSIRO, Herston, Australia; 5 St. Georges Hospital, Parkville, Australia; 6 Austin Hospital, Heidelberg, VIC, Australia; 7 Edith Cowan University, Joondalup, Western Australia, Australia; 8 CSIRO, Parkville, Australia; 9 National Ageing Research Institute Inc. (NARI), Parkville, Australia; 10 Florey Institute, Parkville, Australia. Contact e-mail: pierrick. bourgeat@csiro.au Background: Predicting the rate of evolution of preclinical changes and the onset of clinical phase of AD might prove essential for the design and timing of disease-modifying therapeutic interventions. Methods: 207 participants (152 HC; 36 MCI; 19 AD) underwent periodic neuropsychological exami- nation, MRI and PiB-PET scans for a mean follow up of 4.2 (CI 3.4-4.8) years. A SUVR cb of 1.5 was used to discriminate high (Ab+) from low (Ab-) Ab burdens. Analyses were adjusted for age, gender, years of educa- tion, and ApoE status. Results: 170 (82%) of the 207 participants (79% HC, 89% MCI, 100% AD), were deemed Ab accumulators. In HC and MCI, Ab deposition and memory decline were significantly faster in Ab+ than in Ab-. Ab deposition was significantly faster in participants who transitioned to either MCI or AD, although baseline Ab burden was a better predictor of progression. Significant associations between rates of Ab deposition and rates of episodic memory (EM) decline were found in HC (R 2 ¼0.14; p¼0.009). There were no associations between Ab deposition and EM decline in MCI or AD. When Ab+HC and Ab-HC were considered sepa- rately, a significant association between rates of Ab deposition and EM decline was found only in Ab+HC (R 2 ¼0.31; p¼0.013), association that persisted after adjusting for baseline Ab-burden. There was no association between rates of Ab deposition and EM decline in Ab-HC, not even when those with positive rates of Ab accumulation were examined separately. Conclusions: Ab seems to play a threshold/trigger effect on EM at the early stages of Ab deposition. While this support the theory that Ab plays a funda- mental role in the development of AD, it suggests that other dependent or independent factors have a more direct effect on symptom progression at later stages of the disease. These results also suggest that in order to prevent memory decline, anti-Ab therapy should target these early stages of Ab deposition. O3-10-05 SPATIAL CORRELATION BETWEEN GRAY MATTER CHANGES IN CHILDHOOD, FDG METABOLISM IN YOUNG ADULTS, AND BETA- AMYLOID DEPOSITION IN THE ELDERLY Renaud La Joie, Sam Crowley, Shawn Marks, Jacob Vogel, Carter Wendelken, Silvia Bunge, William Jagust, University of California, Berkeley, Berkeley, California, United States. Contact e-mail: lajoie@ berkeley.edu Background: Brain regions that are susceptible to the deposition of beta- amyloid (Ab) share unique features, including high neural activity, func- tional connectivity, and synaptic plasticity. These characteristics may have their onset during brain development in early life. We performed this study in order to investigate relationships between brain development in early life, glucose metabolism in early adulthood, and Ab deposition in late life. Methods: The pattern of Ab was defined as the mean difference between PIB-DVR images of 10 PIB-negative and 10 PIB-positive subjects from the Berkeley Aging Cohort (age ¼ 75.4 6 7.9 years old). FDG-PET data were obtained from 11 young adults (20-30 years old) and pons- normalized SUVR images were averaged. Developmental changes in gray matter (GM) were assessed in 93 healthy children (6 to 18 years old) included in the Neurodevelopment Of Reasoning Ability study. T1- weigthted MRI scans were preprocessed using the VBM5 toolbox to deter- mine the effect of age on GM. All maps were warped to the MNI template to ensure comparability and spatial correlations were assessed both voxelwise and using regions of interest (ROIs) from the AAL atlas. Results: Develop- mental effects showed reductions in GM with advancing age that were high- ly correlated with both metabolism and Ab. All three maps (Figure, top row) showed significant spatial correlation, on a voxelwise level (GM-FDG, r ¼ 0.264; FDG-Ab,r ¼ 0.546; GM-Ab,r ¼ 0.336, all p<0.001; Figure, middle row) or with AAL ROIs (pruning-FDG, r ¼ 0.563; FDG-Ab,r ¼ 0.634; pruning-Ab,r ¼ 0.641, all p<0.001; Figure, bottom row). Conclusions: The regions that are prone to Ab already share common features in early life: they are very metabolically active in young adults and also show age-related gray matter reduction ("pruning") in childhood. While these topographical correlations do not define causal relationships, they support theories that synaptic activity influences Ab deposition. In particular, these data suggest that early life pruning may result in neural systems in associa- tion cortex that are metabolically stressed and thereby susceptible to Ab deposition. Oral Sessions: O3-10: Neuroimaging: Amyloid and the Aging Brain P229