Memory loss is one of the earliest and most common symptoms of Alzheimer's disease (AD). Neurofibrillary tangles (NFT's) and neuritic plaques (NP's) are abnormal structures that develop in the brains of those with AD. NFT's progress in a predictable pattern beginning in areas important for memory: the entorhinal and transentorhinal areas and then the hippocampus. Research also suggests that the number of NFT's in the olfactory bulb is highly associated with the number in the entorhinal cortex (EC), an area also important for olfactory processing. Thus, olfactory impairment might be expected to reflect the number of NFT's in the EC. The present study used archival data and participants included 75 pathologically confirmed AD patients and 10 non-demented older controls. Participants were given a recognition memory task, which required participants to recognize previously, presented odors, faces, and symbols. Post-mortem brain analyses were conducted and numbers of NFT's and NP's were recorded. Braak & Braak neurofibrillary staging criteria was used to evaluate NFT distribution and Consortium to Establish a Registry of Alzheimer's Disease (CERAD) standards were used to assess NP density. We hypothesized that AD participants would display poorer recognition memory performance than controls and that both groups would commit significantly more false recognitions for odor than visual stimuli. Using ANOVA with repeated measures, these hypotheses were supported (p < .05). We further hypothesized that greater numbers of NFT's in the EC, increased Braak stage, and increased CERAD score would be significantly correlated with greater likelihood of committing false recognitions. Partial correlations suggested that greater numbers of NFT's in the EC and increased Braak stage was associated with greater likelihood of committing false recognitions for odor and facial stimuli (p < .05), but not for symbol stimuli (p > .05). CERAD score was not associated with false recognitions for our three types of stimuli (p > .05). These results suggest that odor (and face) recognition memory may predict NFT degeneration, but not NP density. Taken together with previous research, these results support the potential use of olfactory testing as a low-cost, noninvasive procedure to aid in the diagnosis and understanding of AD.