Cheng, Yu-HsiangYu-HsiangChengYang, Cheng-ChangCheng-ChangYangHua, Mau-SunMau-SunHuaChen, Ta-FuTa-FuChenChang, Yu-LingYu-LingChangChiu, Ming-JangMing-JangChiu2026-03-132026-03-132026-02-05https://www.scopus.com/pages/publications/105029613015https://scholars.lib.ntu.edu.tw/handle/123456789/736317Objective This cross-sectional study aimed to explore the relationship between olfactory memory function and the early detection of dementia of Alzheimer’s type (DAT), minimizing the confounding effects of previous experiences and verbal memory by employing unfamiliar stimuli. Additionally, the study developed and evaluated the Olfactory Recognition Test with Unfamiliar Stimuli (ORT-US) to assess its discriminative power for identifying individuals with DAT. Methods Participants were recruited and classified to healthy control (HC, n=30), subjective cognitive decline (SCD, n=35), mild cognitive impairment (MCI, n=18), and very mild DAT (n=7) groups. ORT-US scores were compared across groups using analysis of variance, while discriminative power was examined via receiver operator characteristic (ROC) analysis. Results Significant differences in ORT-US scores were found only between the scores of HC and very mild DAT groups. The false-alarm scores (i.e., misrecognizing non-target odors as the targets) were significantly different, while those of the hit score (i.e., correctly recognize the target odors) were insignificant. No significant interaction between participant age and total scores, supporting their irrelevancy. Unsatisfactory area under the curve was found, indicating that odor recognition with unfamiliar stimuli cannot effectively detect individuals with DAT. Conclusions While olfactory memory function presents relationships with the early detection of DAT, odor recognition using unfamiliar stimuli reveals limited efficacy as a clinical marker. Further studies are needed to validate these findings.Dementia of Alzheimer’s typeMild cognitive impairmentOdor recognition testOdor-recognition functionSubjective cognitive declinejournal article10.1093/arclin/acag006