P2–326: Longitudinal change in FMRI activation correlates with change in clinical trial measures in Alzheimer's disease

Functional MRI (fMRI) holds promise as a biomarker in Alzheimer's disease (AD) clinical trials, but the relationship between fMRI activation patterns and standardized clinical trial memory measures remains largely unexplored. Using baseline assessments in an investigator–initiated clinical trial, we previously reported that better performance on verbal memory measures correlates with greater fMRI activation in left superior temporal and inferior prefrontal regions in AD patients. To assess longitudinally whether changes on clinical trial memory measures track with alterations in memory–related fMRI activations in mild AD. An interim blinded analysis in 11 subjects with mild Probable AD (MMSE 17–25, age 54–81, all on stable donepezil) participating in a randomized, double–blind, placebo–controlled pharmacological fMRI study with an NMDA–antagonist (memantine) of 24–week duration (including a 12–week phase in which all subjects received memantine). Subjects underwent 4 fMRI sessions using a Novel vs. Repeated (NvR) face–name encoding paradigm over 24 weeks. Change scores in performance from baseline to week 24 on the Free–and–Cued–Selective–Reminding–Test (FCSRT–cued), Visual Paired–Associate learning (VPA) task, and Mini Mental State Exam (MMSE) were entered as regressors into SPM2 whole–brain analyses for change in NvR contrast maps from baseline to final fMRI (week 24). Improvement in behavioral performance on FCSRT–cued (from 26.5 to 33.9 words; p=0.02) over 24 weeks correlated with increased NvR activation in left superior temporal (R=0.80; p=0.0011 uncorrected), left precuneal (R=0.92; p<0.0001), right inferior frontal (R=0.92; p<0.0001) and right posterior cingulate (R=0.85; p<0.001) regions. Improved performance on VPA (p<0.05) over time correlated with greater activation in right (R=0.83; p<0.001) and left (R=0.87; p<0.001) parahippocampal gyrus. No significant change was observed in MMSE (from 21.5 to 22.9, p=0.28), and no brain regions showed change in activation that correlated with MMSE. In preliminary analyses of a blinded AD clinical trial, change in performance on standardized memory measures correlated with change in fMRI activation. Furthermore, cross–modal associative memory–related activations demonstrated regionally specific changes in medial temporal lobe and neocortical areas associated with domain specific memory performance. These results support the validity of fMRI as a potentially sensitive biomarker for assessing signals of efficacy in early–stage AD clinical trials.