Mr A is a 73-year-old resident of a nursing home, where the irate aides describe him as “a liar and a troublemaker.” Mr A had been a successful and well-liked businessman until 2 years ago, when he sustained a stroke secondary to a ruptured aneurysm of the anterior communicating artery. Prominent memory deficits ensued. Mr A also became irritable, verbally abusive, and “hard to manage.” Staff also complained that Mr A “made up stories about the staff, just to get us in trouble . . . like, that we were sleeping with him.” There was no evidence supporting any such claims, and Mr A’s “stories” were regarded by the staff as deliberate mischief on his part.
Confabulation was first described by Korsakoff in amnesic patients in 1889,1 yet its nature and etiology remain a source of some controversy. Indeed, in the literature since Korsakoff’s time, even the definition of “confabulation” has been inconsistent,2,3 reflecting the diverse theories concerning the origins of this syndrome. Thus, confabulation has been variously described as a falsification of memory in association with an organically derived amnesia,4,5 an extreme form of lying or deception,6 and “honest lying.”7(p133) In more recent texts, confabulation is usually defined as statements and/or actions that are “unintentionally incongruous to the patient’s history, background, and present situation,”8(p112) or as erroneous memories—either false in themselves or derived from true memories inappropriately retrieved or interpreted.2,9
Here we broadly define confabulation as the production of false or erroneous memories without the intent to deceive. As our opening vignette illustrates, the confabulating patient’s intentions may sometimes be misread as malicious, when in reality his confabulations may be better understood as products of one or more neuropsychological deficits. The false memories of confabulation may range from inaccurate or distorted recollections of past events, to incongruous intrusions during memory tasks, to fictitious and bizarre narratives.10
In this article, we place confabulation in a historical perspective, review current theories of confabulation, and discuss clinico-anatomical syndromes often seen by psychiatrists in which confabulation frequently occurs. Finally, we summarize the implications of recognizing and understanding confabulation in psychiatric practice.
TYPES OF CONFABULATIONS: HISTORICAL OVERVIEW
In 1901, Bonhoeffer distinguished the confabulation of embarrassment (later termed “momentary confabulation”) from spontaneous confabulation.4 Confabulation of embarrassment referred to fabricated memories that appeared to compensate for memory loss—in effect, the patient attempts to cover up an exposed “gap in memory.” In contrast, Bonhoeffer described spontaneous confabulation as exceeding the need to cover a memory deficit and often consisting of “fantastic” (grossly implausible) content.4
Berlyne,4 in 1972, also delineated 2 distinct forms of confabulation. Like Bonhoeffer, he referred to the first as “momentary” confabulation and the second as “fantastic” or “productive” confabulation. Momentary confabulations, for Berlyne, consisted of autobiographical content and were rooted in true memory. Furthermore, these confabulations occurred only in response to questioning. In contrast, fantastic (or productive) confabulations consisted of grandiose content and occurred without provocation.
Kopelman5 later revised this terminology, using the terms “provoked” (rather than “momentary”) and “spontaneous” (rather than “fantastic”) confabulation. Kopelman noted that provoked confabulations commonly occurred among amnesic patients given memory tests and resembled errors produced by healthy persons on tests of memory following prolonged retention intervals. Indeed, provoked confabulations have since been noted in various experimental studies to occur in healthy people. Accordingly, many authors consider provoked confabulations to be a normal strategy to compensate for memory deficits rather than a pathological process.5,11-13 Kopelman, for example, notes that provoked confabulations under experimental conditions—defined as intrusion errors or distortions made in response to a challenge to memory—reflect the “reconstructive” nature of normal memory retrieval. When a memory trace is particularly weak (for example, after long retention intervals), reconstruction becomes distorted or frankly erroneous.11
Provoked confabulations may reflect such normal compensatory mechanisms. However, several recent studies have shown that provoked confabulations may also reflect neuropathological conditions, such as Wernicke-Korsakoff encephalopathy14,15 and Alzheimer disease (AD).16,17
Many authors continue to use Kopelman’s “spontaneous” versus “provoked” dichotomy. Some argue, however, that these are not distinct types of confabulation. Rather, spontaneous confabulation may simply represent a more severe form of memory falsification.12,18,19
CLASSICAL CONFABULATION HYPOTHESES AND THEIR CRITICS
Confabulation and memory impairment
Confabulation is classically and historically associated with memory loss. As noted, one of the earliest hypotheses proposed that confabulations occurred as a compensatory mechanism for memory loss (ie, the patient produces confabulations to fill in memory gaps and avoid embarrassment).4 Yet numerous authors have challenged this view, noting that patients with memory deficits do not always demonstrate confabulations,4,20 that confabulation usually resolves during the chronic stage of Korsakoff syndrome despite ongoing memory deficits,3 and that the severity of amnesia does not correlate with the tendency to confabulate.21 Furthermore, confabulation has been observed in the complete absence of memory deficits.22-24
Frontal lobe lesions
Because confabulation was frequently seen in patients with frontal lobe damage it led to the belief that frontal lesions were the cause of confabulation.19,20,25,26 However, several experiments suggest that frontal lobe pathology may be neither necessary nor sufficient for confabulation. On the one hand, for example, there are reports of patients who confabulate but who show no signs of frontal executive dysfunction or frontal lobe structural pathology.18,24 On the other hand, controlled studies have found that common executive function deficits reflecting frontal lobe impairment do not distinguish spontaneously confabulating amnesiacs from nonconfabulating amnesiacs—suggesting that frontal dysfunction is not sufficient to produce confabulation.13,27,28 Moreover, in a study of patients with AD who had provoked confabulations, the tendency to confabulate did not correlate with performance on frontal/executive tasks.29
Partly as a consequence of these uncertainties, the essential role of frontal lobe pathology in confabulation has been challenged by several investigators.30-33 This skepticism, in turn, led to the “dual-lesion” hypothesis, which states that confabulations arise from the concomitant presence of frontal lobe pathology and an organic amnesia.19,25,34,35
MORE RECENT HYPOTHESES OF CONFABULATION
More recent views of confabulation focus on 3 central problems:
• Deficits in reality monitoring
• Dysfunction of strategic retrieval processes
• Temporal confusion
Reality monitoring (or source monitoring) refers to the neural mechanisms by which memories are “checked” to ensure that they correspond to actual (vs imagined) events.36 For example, one might think, “Did I really see Jim at the office party last year, or did I just dream that?” According to the reality- or source-monitoring deficit hypothesis, dysfunction or loss of these “fact-check” mechanisms results in confabulations. However, source-monitoring deficits may be seen in nonconfabulating patients, which suggests that such deficits may be necessary but not sufficient to produce confabulation.28,33,37 Even more troubling for the source-monitoring hypothesis was the demonstration by Dalla Barba and colleagues29 that the degree of source-monitoring deficits in a group of patients with AD who had provoked confabulations did not correlate with the tendency to confabulate.