Background: Functional MRI (fMRI) has shown promise as a tool to characterize altered brain function in Alzheimer disease (AD) and for use in proof of concept clinical trials. FMRI studies of subjects with AD have demonstrated altered hippocampal and neocortical activation while encoding novel stimuli compared to older controls. However, the relationship between fMRI activation and performance on standardized clinical trial memory measures has not been fully investigated.

Objective: To determine whether patterns of activation during an associative-memory fMRI paradigm correlate with performance on memory measures used in AD clinical trials.

Methods: Twenty-nine subjects with AD underwent neuropsychological testing, including the AD Assessment Scale (ADAS-Cog), and an associative-encoding fMRI paradigm. Scores were entered as regressors in SPM2 analyses of the differential fMRI activation to novel-vs-repeated (NvR) stimuli. To account for cerebral atrophy, native-space structure-function analyses were performed with subjects' high-resolution structural images.

Results: Performance on the ADAS-Cog verbal memory component, and the ADAS-Cog total score, correlated with NvR activation in left superior temporal (p = 0.0003; r = -0.51) and left prefrontal (p = 0.00001; r = -0.63) cortices. In a subgroup with more extensive neuropsychological testing (n = 14), performance on the Free and Cued Selective Reminding Test was correlated with activation in these same regions. fMRI activation remained correlated with performance even when accounting for atrophy.

Conclusions: The relationship between functional MRI (fMRI) activation and standardized memory measures supports the potential use of fMRI to investigate regional mechanisms of treatment response in clinical trials of novel therapies for Alzheimer disease. .