Objective: We aimed to develop a measure to specifically assess the functioning of the perirhinal cortex (PRC), a brain structure affected very early in Alzheimer's disease (AD) pathology. In this novel task, participants were shown arrays of six complex figures and had to identify the "odd-one." Method: The pilot study included 50 normal controls (NCs) and 50 patients in very early stages of AD. Participants completed the task and received MRI scanning. Best differentiating items were determined and applied in a validation study including 25 NCs, 27 early-stage AD patients, and 26 patients with major depression. Logistic regression models investigated if task performance predicted group membership. Task performance was then related to whole-brain gray matter integrity. As proof of concept, cortical thickness values of four regions of interest (ROIs; e.g., medial PRC and entorhinal cortex [ERC]) were compared between the groups. The associations of task performance and cortical thickness of the ROIs were investigated using linear models. Results: Task performance showed good discriminative ability between early-stage AD patients and NCs. Whole-brain analyses revealed four significant clusters ( p < .001) with peak voxels in parahippocampal regions including PRC and ERC. ROI analyses showed distinctly reduced cortical thickness in the AD group compared to both other groups in the medial PRC and ERC ( p ≤ .001). Task performance modeled by ROI cortical thickness did not achieve significant results. Conclusion: Although further validation is needed, especially with age-matched participant groups, these findings indicate that the task detects early cognitive impairment related to AD. Neuropsychology

Background The differentiation between early Alzheimer’s disease (AD) and Major depression (MD) is a common clinical challenge. AD‐related neurofibrillary tau‐pathology starts in the perirhinal cortex (PRC) whereas little to no neurodegeneration is found in MD. Thus, testing PRC functioning could help distinguishing these pathologies. The PRC is involved, amongst others, in visual discrimination of complex objects. We assessed the ability of a novel visual oddity detection task to differentiate adults in early stages of AD from those with MD and healthy controls (HCs). Method Participants had to choose the odd figure out of six highly ambiguous, simultaneously presented Kanji‐like stimuli. The sample included 28 patients in early stages of AD (8 amnestic mild cognitive impairment, 20 Alzheimer's dementia, 14 female, age = 75.74±6.72 years, education = 13.51±3.13 years, MMSE = 26.29±1.94), 26 patients with MD (14 female, age = 61.88±10.82 years, education = 13.31±3.04 years, MMSE = 28.77±1.27), and 25 HCs (11 female, age = 68.98±10.74 years, education = 14.26±3.91 years, MMSE = 29.01±1.04). As proof of concept, only AD patients showed distinct PRC thinning adjusted for age and total intracranial volume (HC vs. AD: t = ‐3.355, p = .001; HC vs. MD: t = ‐.596, p = .553). Logistic regressions were performed using demographically adjusted z‐scores of task performance. Result Incorrect responses predicted group membership of AD vs. MD patients (odds ratio [OR] = 0.45, 95% confidence interval [CI; 0.22, 0.77], p = .011), as well as AD patients vs. HCs (OR = 0.45, 95% CI [0.22, 0.79], p = .012). Furthermore, correct responses predicted group membership of AD patients vs. HCs (OR = 0.35, 95% CI [0.16, 0.66], p = .004), but not between AD patients vs. MD (p = .11). Omitted responses did not have a distinct effect (AD vs. MD: p =.36; AD vs. HC: p = .051). Conclusion A novel oddity detection task differentiated patients in early stages of AD from MD patients and HCs. Although further validation is needed, these findings indicate that the task detects early cognitive impairment related to AD and may support differential diagnosis.