P3-483: amyloid and tau burden improve residual approaches to cognitive reserve quantification in ad

Cognitive reserve (CR) is a construct describing resistance to brain pathology. Residual cognitive performance, after accounting for the effect of relevant variables such as brain pathology, has been employed as a measure of CR. Most prior studies, however, have not accounted for the influence of amyloid and tau burden in modeling pathology. We extended the CR framework to include PET biomarkers of AD neuropathology. 363 participants from ADNI were analyzed, including 129 cognitively normal, 84 with significant memory concerns, 74 with early MCI, 44 with late MCI, and 32 with AD. We assessed the relationship between groups of related variables comparable to those previously delineated by Reed et al (2010) and episodic memory using multiple linear regression (MLR). We report the relationship between memory and sex, race, ethnicity, and education as MEM-D (demographic); memory and cortical volume, hippocampal volume, and white matter hyperintensities as MEM-B (brain MRI); and, memory and amyloid (mean cortical [18F]Florbetapir SUVR), tau (mean lateral temporal lobe [18F]Flortaucipir SUVR), and MEM-B variables, as MEM-B+. Combinations of these models are denoted as MEM-DB and MEM-DB+. MLR was used to assess the relation of CR, defined as unstandardized residuals from the above models, to functional outcomes including FAQ and CDR sum-of-boxes (CDRsb). Including amyloid and tau significantly improved the fit for models MEM-B+ (R2=.334, p