Decreased Self-Appraisal Accuracy on Cognitive Tests of Executive Functioning Is a Predictor of Decline in Mild Cognitive Impairment

Abstract

Objective: Mild cognitive impairment (MCI) in older individuals is associated with increased risk of progression to dementia. The factors predicting progression are not yet well established, yet cognitive performance, particularly for memory, is known to be important. Anosognosia, meaning lack of awareness of one's impaired function, is commonly reported in dementia and is often also a feature of MCI, but its association with risk of progression is not well understood. In particular, self-appraisal measures provide an autonomous measure of insight abilities, without the need of an informant.

Methods: The present study examined the utility of self-appraisal accuracy at baseline for predicting cognitive decline in 51 patients using an informant-free assessment method. Baseline task performance scores were compared to self-assessments of performance to yield a discrimination score (DS) for tasks tapping into memory and executive functions.

Results: Linear regression revealed that a larger DS for executive function tasks in MCI predicted functional decline, independent of age, education, and baseline memory and executive task scores.

Conclusion: These findings indicate that objective estimates of self-appraisal can be used to quantify anosognosia and increase predictive accuracy for decline in MCI.

Keywords: Alzheimer’s disease; anosognosia; cognition; dementia; disease progression; neurodegeneration; neuropsychology; prodromal symptoms.

Figure 1

Picture of the bell curve provided as a visual aid for the self-appraisal component of assessment, with labels corresponding to percentile rankings ranging from 1 to 99. Prior to the memory and executive tasks, participants were informed that they were going to do a post-task performance assessment, rating themselves on the task compared to what they would expect to be average performance. In addition, they were consistently reminded that on most tasks the majority of people would score in the average range, at 50th percentile, while fewer individuals would have much higher or much lower scores than average.

Figure 2

(1) Diagnosis breakdown for the ALLsample at Baseline and Final testing sessions (Note: Final box indicates conversions from baseline status); NC, normal controls; MIC, mild cognitive impairment; AD, Alzheimer’s disease; bvFTD, behavioral variant frontotemporal dementia. (2) Components making up the regression model (Pr, performance; SAPr, Self assessment of performance; DS, Discrimination score). (A) Memory component of the model, with DS and Pr control scores; (B) Executive component of the model, with DS and Pr control scores; (C) Impairment index corrected for the variable times between baseline and final diagnosis; (D) Other important variables in the model, including Age at baseline assessment and Education level (# of years).