mutations have been reported to show prominent striatal involvement. We report a case of early onset AD associated with a PSEN1 mutation, an atypical dementia profile, spastic paraparesis, and a novel topography of PiB retention. Methods: Chart review, genetic testing, serial neurological and neuropsychological examinations, and structural and functional neu- roimaging. Results: A 40-year-old right-handed woman presented with a three year history of progressive left leg and arm weakness accompanied by speech and memory difficulties. A standard neurological evaluation was unrevealing. Neuropsychological testing showed primary deficits in work- ing memory, complex attention, naming, and auditory and visual memory. Genetic testing revealed a PSEN1 mutation (L381V), although there was no family history of AD and both parents tested negative for the mutation. Clinically, she was markedly apathetic and child-like in demeanor; neuro- logical exam showed hyper-reflexia and spasticity of the lower extremities, left worse than right. An MRI brain scan showed no focal atrophy or masses. Fluorodeoxyglucose PET imaging revealed temporal and parietal hypometabolism consistent with AD. However, PiB retention was most prominent in the cerebellum, with only modest neocortical involvement and minimal involvement of the striatum. Conclusions: These findings extend the phenotypic spectrum and amyloid imaging correlates of PSEN1 mutation carriers with early onset AD and spastic paraparesis. This patient exhibited a dementia profile more reminiscent of a cerebellar “cognitive- affective syndrome” than typical AD, which appears more likely to be associated with marked cerebellar amyloid deposition, a novel finding using PiB, than structural abnormalities revealed on MRI or functional changes on glucose metabolic PET imaging. P1-257 VISUAL RATING OF THE HIPPOCAMPUS IN NONDEMENTED ELDERS: CAN IT DIFFERENTIATE BETWEEN HIPPOCAMPAL VOLUME AND OTHER INDICES OF BRAIN ATROPHY? Arnoud Knoops 1 , Yolanda van der Graaf 1 , Lotte Gerritsen 1 , Willem P. Th. M. Mali 2 , Mirjam I. Geerlings 1 , 1 University Medical Center Utrecht, Julius Center for Health and Primary Care, Clinical Epidemiology, Utrecht, Netherlands; 2 University Medical Center Utrecht, Department of Radiology, Utrecht, Netherlands. Contact e-mail: a.j.g.knoops@umcutrecht.nl Background: The hippocampus is considered to be one of the first brain structures affected in Alzheimer’s disease. Visual rating of hippocampal atrophy is often used to differentiate between normal aging and Alzhei- mer’s disease. However, is not clear if visual rating can differentiate between disproportionate loss of hippocampal tissue relative to other brain tissue. We therefore investigated whether visual rating of hippocampal atrophy was related to hippocampal volumes, and if visual rating was related to global and subcortical brain atrophy. Methods: Within the SMART-MR study, a prospective cohort study among patients with man- ifest arterial disease, medial temporal lobe atrophy was qualitatively rated using the 5-point MTA scale in 95 (mean age 6210 years) participants without dementia. Hippocampal volume was assessed by manual segmen- tation. Both measures were performed on coronal images of a 3-dimen- sional TFE T1-weighted MR image with isotropic voxels. Automated brain segmentation was used to quantify volumes of brain tissue and cerebro- spinal fluid. Total brain volume and ventricular volume were divided by intracranial volume to obtain brain parenchymal fraction (BPF) and ven- tricular fraction (VF). Lower BPF is an indicator of more global atrophy and higher VF is an indicator of more subcortical atrophy. Results: Mean crude hippocampal volume was 6.20.8 ml, and 82% percent of the patients had a MTA score below 2. Using ANOVA, with increasing MTA scores crude hippocampal volumes were smaller (mean volumes 3.0 vs. 2.9 vs. 2.7 vs. 2.4 ml for MTA scores 0, 1, 2 and 3, respectively; test for linear trend, p0.05) as were hippocampal volumes normalized for intracranial volume. However, hippocampal volumes normalized for total brain size were not smaller with increasing MTA scores (mean volumes 3.0 vs. 2.9 vs. 2.8 vs. 2.7 ml; test for linear trend, p=0.33). Moreover, with increasing MTA scores BPF was smaller (79.6 vs. 78.3 vs. 76.0 vs. 75.4%) and VF was larger (1.9 vs. 2.1 vs. 3.1 vs. 3.5%) (tests for linear trend, p0.001). Conclusions: Visual rating of hippocampal atrophy is related to hippocam- pal volumes. However, in this non demented population, visual rating does not differentiate between disproportionate loss of hippocampal tissue rel- ative to other brain tissue. P1-258 PINPOINTING CHANGE IN ALZHEIMER’S DISEASE: LONGITUDINAL FDG-PET ANALYSES FROM THE ALZHEIMER’S DISEASE NEUROIMAGING INITIATIVE Susan M. Landau 1 , Cindee Madison 1 , Derek Wu 1 , Connie Cheung 1 , Norman Foster 2 , Eric Reiman 3 , Robert Koeppe 4 , Michael Weiner 5 , William J. Jagust 1 , 1 UC Berkeley, Berkeley, CA, USA; 2 University of Utah, Salt Lake City, UT, USA; 3 Banner Good Samaritan Medical Ctr, Phoenix, AZ, USA; 4 Univ of Michigan, Ann Arbor, MI, USA; 5 SF Veteran Affairs Med Ctr, San Francisco, CA, USA. Contact e-mail: slandau@berkeley.edu Background: Decreased [18F]Fluorodeoxyglucose metabolism (FDG- PET) in specific brain regions is a hallmark of Alzheimer’s disease (AD) and Mild Cognitive Impairment (MCI). However, relatively small samples and varying methodologies in existing studies have made it difficult to precisely identify critical hypometabolic regions and characterize longitu- dinal change. The Alzheimer’s Disease Neuroimaging Initiative (ADNI) is a large multisite study designed to refine methodology for effectively measuring longitudinal decline in AD and MCI. Here, we developed study-independent regions of interest and used them to identify areas that show maximal annual metabolic decline in AD. Methods: Frequently reported hypometabolic regions in existing FDG-PET studies of AD and MCI were used to generate regions of interest (Meta-ROIs). AD patients (N = 28) underwent FDG-PET imaging and clinical evaluation at three time points (baseline, 6 months, and 12 months). We used an automated, group-specific, template-based approach to extract FDG-PET data from grey matter within Meta-ROIs and corresponding anatomically-based re- gions of interest (MNI ROIs). We calculated annual percent change for both sets of ROIs and determined the sample sizes needed to detect T291 Poster Presentations P1