Divided attention during retrieval suppresses false recognition in confabulation.

The present research tested the hypothesis that false recognition in confabulating patients results, at least in part, from sustained processing of irrelevant information about test items. Thus, we predicted that false recognition would be reduced if cognitive resources available for such a processing are experimentally reduced. Two experiments provided evidence consistent with this proposal. In Experiment 1, confabulating patients exhibited lower false recognition in a divided attention (DA) compared to a full attention (FA) retrieval condition. The opposite trend was observed in non-confabulating patients with comparable memory and executive deficits and normal controls. In Experiment 2, in a Deese-Roediger-McDermott paradigm, confabulating patients exhibited lower false recognition to lures that were unrelated to the studied words in the DA compared to the FA retrieval condition, but unvaried levels of false recognition to critical lures. In contrast, false recognition to both unrelated and critical lures increased in non-confabulating patients and normal controls in the DA compared to the FA condition. These results are consistent with the hypothesis that an excessive processing of task-irrelevant memories at retrieval may contribute to false recognition in confabulating patients.