To err is human, to monitor divine: Environmental adaptations reduce everyday errors but do not improve monitoring

Abstract

The current study aimed to address error monitoring impairments in dementia using an intervention for execution deficits. Thirty-eight participants completed the Naturalistic Action Test (NAT) under two conditions: Standard and User-Centered. The Standard NAT followed the manual procedures; in the User-Centered NAT, objects were arranged sequentially, and distractor items were separated from target objects. While participants committed fewer errors in the User-Centered condition, there was no difference in the proportion of errors detected. However, the neuropsychological processes associated with monitoring differed across conditions. The results have implications for a neuropsychological model of error monitoring in dementia.

Keywords:

• Naturalistic action
• Activities of daily living
• Instrumental activities of daily living
• Functional abilities
• Errors of action
• Error monitoring
• Error detection

Acknowledgments

A portion of this paper was presented at the 2009 International Neuropsychological Society conference. This research was supported in part by funds from an American Psychological Association, Division 40 Early Career Award, awarded to Brianne Bettcher.

Notes

¹Stepwise regressions are not considered ideal when a large corpus of research literature is available to support one's hypothesis; however, given the limited empirical work conducted on error monitoring and dementia, this form of regression was employed to isolate predictors.

²This analysis also was performed with raw error detection scores, controlling for total errors across conditions (i.e., a repeated measures general linear model, with total errors as a covariate). Comparable, nonsignificant monitoring results were found (p = .6) using this approach.

³Considering that individuals with mild dementia were included in the study, participants with a MMSE > 23 were examined separately to determine whether their total errors rates were large enough to adequately assess detection and correction rates. All participants with a MMSE > 23 generated errors on both tasks (Standard: M = 15.9, SD = 11.6; UC: M = 9.0, SD = 5.9), suggesting that ceiling effects did not detrimentally impact the analyses.