Cognitive features of amnestic Mild Cognitive Impairment using specific Cambridge Neuropsychological Test Automated Battery test scores

Abstract Background In older adults with amnestic Mild Cognitive Impairment (aMCI), the Cambridge Neuropsychological Test Automated Battery (CANTAB) probes indicated cognitive impairments most frequently in memory. Objectives This study aimed to investigate a) the cognitive features of aMCI using memory CANTAB tests and b) whether the clinical diagnosis of aMCI can be externally validated by these CANTAB measurements. Methods We tested CANTAB tests that are specific to aMCI on 65 healthy controls and 66 people with aMCI who were diagnosed using Petersen’s criteria. These tests were spatial working memory (SWM), visual pattern recognition memory (PRM), delayed matching to sample (DSM), spatial span (SSP), and rapid visual information processing (RVP). Results The key aMCI features are impairments in PRM and DSM, whilst deficits in SSP and RVP are other, albeit somewhat less important features of aMCI. Nevertheless, neural network analyses including 10 CANTAB domains specific for MCI showed that only 70.8 percent of all subjects were properly identified with a sensitivity of 77.3%, specificity of 65.4% and an area under the ROC curve of 0.760. K-means cluster analysis using the same specific CANTAB test scores discovered 2 clusters with an adequate silhouette measure of cohesion and separation including a cluster with 36 subjects showing impairments in most neurocognitive tests. Conclusion Deficits in spatial working, pattern recognition and visuospatial working memory as well as rapid visual information processing are key features of aMCI. Nevertheless, the clinical diagnosis of aMCI according to Petersen’s criteria is overinclusive because too many healthy controls are allocated to this group.