The effects of frontal injury on “on-line” self-monitoring during verbal learning by adults with diffuse brain injury

Abstract

The purpose of this study was to investigate to the effects of documented frontal injury that coexist with diffuse brain injury (BI), on self-monitoring, operationalised by recall predictions made during verbal learning. The data from 35 adults with diffuse BI who participated in two earlier studies, were divided into groups based on frontal injury (Kennedy, Carney, & Peters, 2003; Kennedy & Yorkston, 2000). Participants studied unrelated noun-pairs, made item-by-item immediate and delayed recall predictions, and took delayed recall tests. Outcomes included relative predictive accuracy, the proportion recalled, and confidence. With months post-onset as a covariate, those with frontal injury were less accurate at predicting recall than those without frontal injury although the pattern was the same; delayed predictions were more accurate than immediate predictions. Conversely, those without frontal injury recalled less than those with frontal injury. Consistent with a phenomenon called “discounting”, those with lower recall (BI without frontal injury) were more confident than those with higher recall (BI with frontal injury). These results help to clarify variability in earlier studies of “on-line” predictions and suggest the need to explicitly train clients with memory impairment to base strategy decisions on delayed self-feedback rather than on self-feedback generated immediately after learning.

Acknowledgements

The authors wish to thank Edward Carney and Suzanne Peters for their contribution to the original work. A portion of this paper was presented at Clinical Aphasiology Conference, Orcas Island, WA, June 2003.

Notes

This study was funded in part by a “Grant-in-Aid” (#17848) awarded to the first author by the Office of the Vice President for Research and the Graduate School, University of Minnesota, Minneapolis, MN.

Kennedy and Yorkston (2000) used the Logical memory subtest from the WMS-R to measure verbal memory, whereas Kennedy et al. (2003) used the Auditory Immediate subtests (Logical memory and the Paired-associates subtests) to provide the percentile ranking from the WMS-III. Even though there are nearly equal numbers of participants from each study in each group, caution should be used when interpreting these scores to indicate differences in memory.

Kennedy and Yorkston (2000) had TBI participants study each pair for 13 s, whereas Kennedy et al. (2003) had TBI participants study each pair for 9 s. Previous studies have found no apparent threat to the predictive accuracy with varying study times (e.g., Dunlosky & Hertzog, 1997; Nelson & Dunlosky, 1991).