Most experts consider that memantine has a symptomatic treatment, but clinical trials have not yet provided compelling evidence to support a disease-modifying effect. We investigate the effects of memantine on clinical ratings; fluorodeoxyglucose positron emission tomography (FDG-PET) measurements, which can monitor disease-modifying effect; and cerebrospinal fluid (CSF) assays in patients with moderate to severe probable Alzheimer disease (AD) dementia. Twenty-two patients completed a 24-week, double-blind, placebo-controlled, randomized clinical trial of memantine, titrated up to 10 mg twice per day using the Severe Impairment Battery, AD Assessment Scale-Cognitive subscale, Mini-Mental State Examination, FDG-PET measurements of the regional cerebral metabolic rate for glucose (CMRgl), and CSF amyloid β (Aβ) and tau assays. An automated brain mapping algorithm and predefined regions of interest were each used to analyze treatment-related regional CMRgl effects. In comparison with the placebo group, the memantine treatment group had significantly less cognitive decline on the Severe Impairment Battery and significantly less CMRgl declines in regions preferentially affected by AD. There were no significant treatment effects on CSF Aβ1-42, CSF Aβ1-40, total tau, or phosphor-tau levels or ratios. This relatively small and brief randomized clinical trial suggests an association between memantine’s clinical benefit and its effects on FDG-PET measurements in AD-affected brain regions. Larger and longer studies are needed to confirm these findings, extend them to earlier clinical and preclinical stages of AD, and help determine the extent to which FDG-PET should be qualified for use as a reasonably likely surrogate end point in the evaluation of putative AD-modifying treatments.