The Mechanism of Confabulation Commentary by Marcel Kinsbourne

Which brain structures are impaired in patients who confabulate? By what mechanism is the confabulation generated? What determines the content of the confabulation? DeLuca concentrates on the first question, and Solms on the last, with little overlap. I hope to find a common thread between them by addressing the middle question.

DeLuca offers a comprehensive discussion of the localization of brain pathology in confabulation. He concludes that confabulation is apt to arise when there is conjoint frontal (specifically ventromedial) dysfunction and impaired memory (due to basal forebrain, diencephalic, or mesial temporal damage). This conclusion is persuasive for the bulk of confabulation, which refers to events in the past, and on which DeLuca focuses. Confabulating about the present, and even the future, which is also well documented in the more severe cases, may not require defective memory. Does orbitofrontal dysfunction suffice to explain it?

Following structural lesions, disordered frontal function and impaired memory are constant and persistent, but confabulation comes and goes. Confabulators only sometimes confabulate, and only on some topics. Confabulations may be offered spontaneously, but more often they are uttered only when they are provoked by leading questions. The suggested mechanisms for confabulation that DeLuca briefly summarizes, impaired self-monitoring and strategic retrieval (Moscovitch and Mello, 1997), may be too much of an explanation for the syndrome. They suggest that when memories are imprecise, it is normal to generate confabulatory material up to the point of utterance, and that it is self-monitoring that warns the individual not to speak but to try again, or admit ignorance. If the reasons for confabulation are so general, then why do even severely affected patients not confabulate during much or all deliberate activity, but only periodically? There must be additional factors that interact with the neurop