This study aims to evaluate whether assessing memory using the visual recognition memory task DMS48 in amnestic mild cognitive impairment (aMCI) can contribute to the early diagnosis of Alzheimer¿s disease (AD). In an 18-month follow-up study, we assessed if longitudinal change in aMCI patients who failed on the DMS48 differs from that of patients who succeeded on this task at baseline. Method: Twenty-six controls and 33 aMCI patients underwent a complete neuropsychological assessment at baseline and at an 18-month follow-up. Patients were divided into two subgroups, according to their z score on the DMS48 (DMS48+ subgroup succeeding; DMS48¿ subgroup failing on the task). In order to detect sensitive longitudinal change over time, we calculated and compared the standardized response mean (SRM) of performance on neuropsychological tasks in the three groups. Results: We found significant differences for the mean SRM of all neuropsychological tests when comparing DMS48+ vs. controls vs. DMS48¿, which was greatest for the comparison between the DMS48¿ and the DMS48+ subgroup. Although cognitive profiles of the two patient subgroups at baseline did not differ in cognitive domains other than memory, we found a consistent decline on all neuropsychological tasks in the DMS48¿ subgroup compared with the DMS48+ subgroup, except for performance on a verbal fluency test. Conclusions: As the cognitive profile of the DMS48¿ subgroup at follow-up resembles the typical pattern of AD described in the literature, this study confirms that visual recognition memory tasks may be useful to anticipate covert cognitive decline in aMCI patients.
