Verbal episodic learning curves predict MCI outcomes 12 months later

Recent research indicates that Mild Cognitive Impairment (MCI) is longitudinally unstable, with a significant proportion of clinically identified cases of MCI reverting to an unimpaired status at subsequent assessment. The aim of the present study was to examine the capacity of a single time point assessment of verbal episodic learning (RAVLT) to differentiate between longitudinally stable and progressive variants of MCI from cases that display recovery 12 months later. A composite sample of 233 adults aged 60+ years were drawn from two existing (2006-2010 and 2011-2013) longitudinal clinical studies of MCI (healthy controls = 59; MCI variants = 164; AD = 10). Repeated measures ANOVA identified significant differences between the learning curves in groups defined by diagnostic outcome 12 months later: control and recovered-MCI groups displayed intact learning curves; persistent na-MCI and unstable MCI groups displayed mildly lowered learning curves; longitudinally persistent a-MCI, persistent a-MCI+, and cases progressing from MCI to AD, all displayed significant learning curve decrements; the lowest learning curves were seen in the confirmed AD referent group. The use of serial list learning episodic memory tasks to assess learning over repeated trials may enhance predictive diagnosis of outcome in suspected cases of, thereby enhancing MCI diagnostic accuracy