Accuracy of metamemory after traumatic brain injury: predictions during verbal learning.

The primary intent of this study was to investigate the metamemory monitoring abilities of adult survivors of at least moderate traumatic brain injury (TBI) during a verbal-learning activity. Eighteen survivors and 18 non-injured control participants made judgment-of-learning (JOL) predictions of their recall ability immediately after studying 3 lists of noun-pairs or after a slight delay. A secondary intent of this study was to determine if verbal retrieval attempts would enhance predictive accuracy. One half of participants made retrieval attempts during the second and third list-learning task, and the other half made retrieval attempts during the third list-learning task only. Measures of the correlation between JOL predictions and recall accuracy revealed that survivors were as accurate as controls when making delayed predictions and were less accurate when making immediate predictions. This occurred regardless of retrieval attempts. Absolute measures that compared mean JOL ratings to overall recall revealed that the survivor group was well-calibrated when making delayed JOL predictions but overestimated when making immediate JOL predictions. The non-injured control group underestimated when making both types of predictions. However, within-group variability was high. These findings are compared to those from studies that investigated metamemory beliefs in which survivors' ratings were compared to family-member ratings. Clinical implications for basing executive decisions about compensatory strategies on delayed and immediate predictions of future recall are discussed. Additionally, a rationale is provided for the use of both relative and absolute measures of predictive accuracy in metamemory studies involving neurological clinical populations.