Empirically-derived subtypes of mild cognitive impairment (MCI) and vascular risk factors

Shereen Haj-Hassan, Tennessee State University


Recent efforts have classified subtypes of mild cognitive impairment (MCI) using neuropsychological performance into distinct groups based on cognitive profiles (e.g., amnestic and dysexecutive). This study aimed to statistically derive MCI subtypes using an innovative analytical approach, latent class analysis (LCA), which empirically assesses the structure of heterogeneous syndromes. MCI participants from the Vanderbilt Memory and Aging Project (n=119, 73±7 years, 41% female) completed a neuropsychological protocol, including California Verbal Learning Test-II, Biber Figure Learning Test, Boston Naming Test, Animal Naming, WAIS-IV Digit Symbol Coding, DKEFS Number Sequencing, DKEFS Letter-Number Sequencing, DKEFS Tower Test, and DKEFS Color-Word Interference Test. MCI participant raw test performances were converted to z-scores. Using these z-scores, LCA was performed for the MCI sample. Three empirically-based MCI subtypes were identified, aligning with three cognitive severity classes, including early (n=33), middle (n=57), and late stage MCI (n=29) as evidenced by performance differences in the expected direction (e.g., early>late) on all neuropsychological tests (all p-values<0.001). Vascular risk factors among the three groups did not differ. Our findings suggest that MCI severity subtypes can be empirically derived. Future studies are needed to assess biomarker and risk profile differences among early, middle, and late stage MCI subtypes as well as longitudinal outcomes, such as diagnostic conversion.

Subject Area


Recommended Citation

Shereen Haj-Hassan, "Empirically-derived subtypes of mild cognitive impairment (MCI) and vascular risk factors" (2016). ETD Collection for Tennessee State University. Paper AAI10194958.