Abstract
Confabulation can be of two types: provoked or spontaneous. The former is the more common and can occur on challenge to an amnesic patient's memory. Spontaneous confabulation involves an unprovoked outpouring of unbelievable autobiographical claims. The purpose of the present paper is to synthesize the current literature on confabulation for the use of treating clinicians. There is a focus on the spontaneous form, which is less common, but more memorable when encountered. In this paper the history, phenomenology, incidence, anatomical underpinnings and theoretical mechanisms of spontaneous confabulations will be reviewed, and then the paper will conclude by addressing prognostic and treatment issues. A systematic literature review of electronic databases was conducted to identify the key articles, reviews and books that have shaped the understanding of spontaneous confabulation.
The evolution of the concept of confabulation over the past 150 years has been characterized by debate on its definition, aetiology, and relationship to other closely related phenomena. It first makes its appearance during the heyday of German psychiatric nosology. Kahlbaum [1] and later Wernicke [2] developed the term ‘presbyophrenia’ for a group of late life conditions, the clinical features of which included confabulations, disorientation, hyperactivity, and affective disturbance. Critics of the term believed that there was little to distinguish presbyophrenia from Korsakoff's psychosis, hypomania, or confusional states, and the term later fell into abeyance [3].
In 1886 Kraepelin suggested the term ‘paramnesia’ for those phenomena, such as confabulations and delusional memories, which attempted to describe types of false narratives [4]. The paramnesias were subdivided into three groups, of which the first, ‘simple pseudoreminiscences’, are equivalent to modern day concepts of delusion, delusional memories, and confabulations. Kraepelin thought that these pseudoreminiscences or pseudomemories were present in both normal people with well-developed fantasies and in patients for whom the boundaries between reality and fantasy had become blurred. Fantastic confabulation was particularly prominent in general paralysis of the insane, while less colourful confabulations were also observed in patients with dementia, paranoid and affective psychoses, and the intellectually impaired. Korsakoff noted the presence of pseudoreminiscences in his classic case descriptions of ‘polyneuritic psychosis,’ some of which were also of a fantastic quality [5]. Wernicke renamed Korsakoff's pseudoreminiscences as confabulations and was mainly responsible for the reduction of Kraepelin's original category into a split between delusions and confabulations, with the latter being linked to memory deficits [4].
At the turn of the century the confabulation phenomenon came under closer scrutiny. Bonhoeffer divided confabulations into two types: (i) confabulations of embarrassment, also known as momentary confabulations, in which patients try to cover up a memory gap; and (ii) spontaneous or fantastic confabulations, which he thought developed from delirious, dream-like states [6]. Berlyne confirmed Bonhoeffer's classification, and proposed that fantastic confabulators may be ‘sensitive’ personalities who are seeking to fulfil their wishes [6]. The memory-gap-filling hypothesis meanwhile gained increasing credence and found its way into standard psychiatric textbooks, but Whitlock eloquently rejected this notion by pointing out that confabulators are usually unaware of their amnesia and therefore cannot be accused of purposefully and consciously attempting to hide it [7].
Other writers also speculated on the aetiology of the phenomenon: the idea that confabulations consisted of actual experiences taken out of temporal order [6]; or because of faulty editing [8]; or as a defence against catastrophic reaction [9]. In 1935 Lafora proposed that confabulation was a type of disinhibition caused by frontal arteriosclerotic changes [3]. The idea that confabulation occurs as a result of faulty executive function makes intuitive sense and the last 30 years has seen an upsurge in research on the role of frontal lobe pathology [10–14].
Berrios, however, argued that frontal pathology does not explain all cases of confabulation and urged a broader understanding of the phenomenon and its relationship to conceptually similar phenomena, such as delusions, lying, and fable-telling [3]. He proposed the idea of confabulation as a disorder of ‘narrative function’ and believed that epidemiological research into its distribution in normal populations was of more heuristic value than descriptions of anecdotal cases.
In a significant and original contribution to the field, Hirstein proposed an ‘epistemic’ understanding of the ‘riddle of confabulation’, in contrast to the foregoing ‘mnemonic’ and ‘linguistic’ concepts [15]. According to Hirstein, the confabulator suffers from an absence of doubt, a ‘pathological certainty’, when forming and expressing beliefs. This is in contrast to the obsessive–compulsive disorder patient, who suffers from pathological uncertainty. The latter checks too much, the former too little. The confabulator has also lost the normal mechanisms that inhibit inappropriate responses, such as autonomic arousal, and is, in this way, a close cousin of the sociopath. Unable to read the minds of others, the confabulator, furthermore, is unaware of the incredulity of listeners as they recount their stories. Although Hirstein's contribution focuses mainly on provoked (non-fantastic) confabulators, his philosophical rigor and use of the clinical and experimental literature provide rich insights into the nature of confabulations in general and their relationship to sociopathy, lying, self-deception, delusions of misidentification, anosognosia, and split-brain patients.
Phenomenology
Confabulation has been defined as ‘a falsification of memory occurring in clear consciousness in association with an organically derived dementia’ and as noted has been broadly divided into provoked and spontaneous types [6]. Bonhoeffer uses the term ‘momentary’ in describing provoked confabulation [6], illustrating the varying nature of these false memories. For the purposes of research and standardization, some have adopted a narrower definition of provoked confabulation as occurring in the context of questioning, often on specific items in neuropsychological tests [12], [16]. Kopelman argued that provoked confabulation occurs in healthy subjects as a normal response to faulty memory [17]. Kopelman saw spontaneous confabulators as a different group in which amnesia is accompanied by other deficits (frontal dysfunction). There remains, however, some debate as to whether these two types of confabulation differ categorically or by degree, with spontaneous confabulation the more severe [12], [18]. This debate continues despite described qualitative differences in their clinical presentations.
In provoked confabulation the patient confabulates in an inconsistent fashion when triggered and the erroneous memory communicated is generally plausible. In the spontaneous or fantastic type, the confabulations are proffered and elaborated without provocation. The content of spontaneous confabulations is said to show less variability than provoked confabulations and is generally grandiose [6]. The fantastic quality of spontaneous confabulations has been described in case studies, such as Berlyne's ‘air hostess’, and ‘ladies’ eleven cricketer’ [6]. In these descriptions there is an implausibility but consistency of content.
Some researchers, however, emphasize other characteristics as necessary for a definition of spontaneous confabulation. Schnider et al., for example, proposed action, consequent and according to the confabulation, as a requirement [12]. Thus, a patient who asked to use a telephone to arrange the afternoon's deer hunt falls within the definition, whereas others, who took no active measures to put their grandiose plans into action, are excluded. Schnider et al. included confabulators whose content is less than grandiose or elaborative. Furthermore, it is unclear from their case descriptions whether the confabulations are consistent in content. Ptak et al. described a patient as a spontaneous confabulator who has mundane and seemingly inconsistent confabulations [19]. Overall, these cases seem to represent a broad heterogeneity in the conceptualization of spontaneous confabulations.
Another important objection to the spontaneous–provoked classification is that the way in which a confabulation is elicited by a clinician may have no bearing on its actual nature. This is a worthy objection but it does not detract from the fact that spontaneous confabulations are usually fantastic and consistent, and it is this fantastic quality that makes it both unique and striking as a phenomenon.
In an effort to elucidate and differentiate spontaneous confabulations from related phenomenological entities, a brief exploration of the issues of belief and intent will follow.
Confabulation and delusion
Spontaneous confabulations that are consistently held may not differ greatly from delusions or delusional memories. Erroneously held beliefs and incorrect inferences are integral components of delusions. Kopelman described delusional memory as consisting of ‘either a true memory that gives rise to a deluded interpretation, or a false memory arising in the context of psychosis’ [20]. They are considered different from delusions in that they are primarily concerned with memory processes.
Dilemmas are raised in several case studies, including that of a spontaneous confabulator responsive to the dopamine antagonist antipsychotic, risperidone, presented by Pihan et al. [21]. The confabulations in that case included delusional misidentifications as well as rather grandiose and consistent narrative occurring while disorientated. Box et al. described a patient suffering from consistent delusional misidentification in the context of clouded consciousness as a spontaneous confabulator [22]. These cases may represent deliria excluded by Berlyne's strict interpretation of spontaneous confabulation requiring the presence of clear consciousness [6]. There are features, however, that led the authors to call their patients spontaneous confabulators rather than delusional. There remains some confusion in the understanding, interactions and overlap of these phenomena. If an incorrect inference is drawn due to the presence of faulty memory and consistently believed, then perhaps delusions and confabulations are differentiated clinically only in the presence of amnesia. This view is supported by Kopelman, who offers that, ‘often only the context distinguishes the clinical features of delusional memory from those of confabulation, i.e. psychosis rather than organic amnesia and/or frontal lobe disease’ [20]. This leaves an unsatisfactory situation in which psychiatric phenomena are distinguished not by any intrinsic difference but on the basis of the clinical syndrome in which they present.
Intent
Whether the patient is conscious or otherwise of covering up their deficiency to the interviewer is not a moot point. Differentiating confabulation from lying is difficult, because one would need to establish that the subject had no conscious intent to deceive. With respect to subjects with severe memory deficits, it was suggested by Zangwill that provoked confabulation is an unconscious mechanism preventing catastrophic reactions [9]. In subjects with severe amnesia, however, determining intent would be near impossible. Spontaneous confabulations similarly might represent an amnesic's attempt to provide grandiose and positive narrative to replace damaged or absent memory, resembling some forms of lying. Berrios developed an interesting premise that spontaneous confabulators represent an extreme of the personality prone to story or myth-telling [3]. Fotopoulou et al. further explored the fantastic and wishful content of spontaneous confabulation [23]. They contend that motivational factors, such as pre-existing personality traits or personal experiences, are either required for fantastic confabulation to occur, or at least substantially affect the content of the confabulations.
Pseudologia fantastica describes subjects who construct fantastic, usually grandiose histories, often compulsively and without obvious gain. It is considered under the rubric of factitious disorders, with the motivation, perhaps unconscious, of gaining the interest of the listener. If one considers the utterances of the pathological liar compulsive or involuntary and also considers the role of motivational influences as just discussed, the distinction between spontaneous confabulating and lying becomes less clear [3].
Kopelman also considers other explanations for false memories, including false confessions and recovered memory of childhood sexual abuse [20]. In both circumstances it is considered that the subject has a belief in the veracity of their statements, thus resembling confabulations and delusional memories.
Aetiology and incidence
Spontaneous confabulation has been reported in a number of different conditions. So disparate are the aetiologies that Berrios was drawn to call them a ‘motley group’ [3]. Spontaneous confabulations are almost always associated with amnesia [13], and some authors claim that amnesia is an essential component [24]. At present it is unclear why some subjects with amnesic disorders confabulate and others do not. As noted, it has long been thought that frontal lobe dysfunction was required for the evolution of spontaneous confabulation.
As far as we are aware, no epidemiological studies have been conducted to assess the incidence of confabulation in the normal population. A number of studies and anecdotal reports, described here, have reported the presence of confabulation in specific brain-injured populations. Provoked and spontaneous confabulations have not been uniformly differentiated between authors, making the task of investigating the incidence and aetiologies of spontaneous confabulation difficult. In a meta-analysis Johnson et al. have classified patients into mild or severe, with the mild category including patients with confabulations designated by authors as ‘mild, moderate, momentary, typical, provoked, realistic, or low’, and the severe category including confabulations described as ‘fantastic, wild, bizarre, spontaneous, free, severe, profound, marked, atypical, persistent, extended, or high’ [25].
Alcohol-related brain damage
In their landmark series of 245 patients with Wernicke–Korsakoff syndrome, Victor et al. found confabulation to be ‘neither consistently present…nor essential for the diagnosis’ [26]. Nevertheless, they found that it was prominent in the early confusional stage of the syndrome (unfortunately, no statistics are provided), where it was thought to be caused by misperceptions of the environment, or in the convalescent phase where fragments of past events were mixed up in time (‘temporal dislocations’). The confabulation was usually provoked, the spontaneous sort being uncommon. Interestingly, confabulation was rarely present in the chronic stable stages of the illness. Berlyne, using Bonhoeffer's term ‘momentary’ confabulation for cases of provoked confabulation, found ‘momentary’ confabulations present in six of his seven Korsakoff patients [6]. As in the Victor et al. study, fantastic confabulation was much less common, present in only one of the patients. Berlyne also found confabulation to be a common phenomenon in dementia ‘of various causes’, present in 23 of 62 cases. Of these, 15 were momentary and eight fantastic. Johnson et al. reported 12 of 75 (16%) of confabulating patients with Korsakoff's to have the severe form as outlined here [25].
Cerebrovascular lesions
Logue et al. found evidence of confabulation at some point in the postoperative period in 18 of 79 patients with disturbances of the anterior cerebral artery, which supplies frontal regions [27]. A number of studies have implicated anterior communicating artery (ACoA) disease, generally aneurysms, as a cause of fantastic confabulation [10], [28–31]. The greater the degree of frontal lobe dysfunction caused by the cerebrovascular lesion, the more severe the confabulation and the more likely it will be spontaneous. In the Johnson et al. analysis 69% of confabulating ACoA patients were classified as severe [25], making it the commonest cause of spontaneous/fantastic confabulations.
Patients with rostral brainstem infarcts have shown marked ‘extraordinary confabulation’, often associated with disturbed wakefulness [32]. Other reported cases of cerebrovascular injuries associated with spontaneous confabulations include bithalamic infarction [33], hypothalamic lesion [19] and rupture of the posterior communicating artery [34].
Traumatic brain injury
Whitty and Zangwill noted confabulation in 38 of 1931 cases of blunt head injury from World War II victims [35]. The confabulation in such cases is usually more florid and spontaneous in the early post-traumatic period, and those that persist usually have significant frontal lobe pathology [36]. Box et al. also described a case of spontaneous confabulation following head injury [22].
Other conditions
Confabulation has also been reported in the acute phases of herpes encephalitis [36], in nicotinic acid deficiency [37], in multiple sclerosis causing frontal and parietal lesion [38], and following an attempted hanging [39]. Gustafson and Hagberg have suggested that confabulation is one of the cardinal signs of hydrocephalic dementia [40]. Nedjam et al. also reported spontaneous confabulation in a patient with fronto-temporal dementia [24].
Anatomical considerations
Current evidence identifies four areas of the brain to be involved in spontaneous confabulation: (i) the orbito-medial frontal cortex and its connections, in particular, the amygdala; (ii) the cingulate gyrus; (iii) the dorsomedial nucleus of the thalamus; and (iv) the medial hypothalamus [41].
Spontaneous confabulations have been reported in patients with orbitofrontal and basal forebrain lesions [42], [43] and dorsomedial thalamic nucleus lesions [44]. Benson et al. reported improvement in orbitofrontal and cingulate gyrus perfusion coinciding with disappearance of confabulation in a patient with alcohol-induced Korsakoff amnesia [45]. Furthermore, rupture of anterior communicating artery aneurysms, already mentioned as a common cause of spontaneous confabulations, cause lesions in the orbitomedial and basal forebrain regions [46]. Memory circuits found not to be associated with spontaneous confabulations are those of the medial temporal lobe, in particular, the hippocampus [12], [46].
Why might damage to these areas cause confabulations? This question remains a topic of debate. One possible answer is that the orbitomedial frontal cortex and the cingulate gyrus, both of which form parts of the anterior limbic system, appear to have a role in the monitoring of verbal responses [45] and the suppression of irrelevant memories [41]. The other is that the orbitomedial region is also involved in the chronological ordering of memories, which is known as ‘temporal order recognition’, and Schnider et al. proposed that failures in temporal ordering are caused by interruption of the loop connecting the orbitofrontal cortex to the amygdala via the dorsomedial thalamic nucleus [12].
Theoretical mechanisms
A couple of important problems constrain confabulation research. The first is that confabulation is uncommon, and spontaneous confabulation even less so. Recruiting patients for a study, therefore, is challenging. Most of the theories mentioned here are speculative, based on single or double case reports [16], [24], [34], [47], [48], and only a few contain sufficient numbers from which something meaningful might be derived [12–14], [49]. Second, there is the problem of classification of the phenomenon under study. Most of the published papers provide insufficient detail about the type of confabulation being studied and, when it is provided, the lack of a standard confabulation battery makes comparison across studies difficult [25]. The problem here is complicated by the fact that the spontaneous/provoked and fantastic/realistic distinctions can overlap, that patients may exhibit more than one type, and that these may vary over time. The studies also, understandably, vary in their definition and assessment of executive function because there are so many of them. This again makes it difficult to draw conclusions about executive function and how it might change as confabulation improves.
Three main theoretical approaches have been influential in the understanding of the mechanisms behind spontaneous confabulation. The first attributes confabulations to disrupted editing of memory traces. The second proposes a relation of spontaneous confabulation more to confusion of the temporal order of memory traces. The third approach responds to the shortfalls of the others by proposing a multi-factorial, or multiple deficits model.
Disrupted editing
The disrupted editing approach was drawn from the idea that spontaneous confabulation relates to errors in memory retrieval [13], [50], [51]. As such, confabulations are recollected due to indiscriminate retrieval and filtering, and poor verification of memories after retrieval [52]. This approach incorporates the concept that frontal system executive deficits (e.g. impaired source monitoring, editing and reasoning) underpin spontaneous confabulation. Burgess and Shallice's model specifies that inappropriate and fantastic responses stem from dysfunction in any of three memory retrieval processes [50]: (i) the descriptor process, which identifies the to-be-retrieved memory trace; damage to the descriptor results in retrieval of too wide a net of memories, which are then conflated; (ii) the editor, which checks the retrieved memory against memory stores and task requirements; and (iii) the mediator, which determines the plausibility of retrieved memory components through reasoning and problem-solving. The authors suggest that it is the extent of damage to this mediator in addition to the editor that determines whether or not a confabulation will be fantastic. Burgess and Shallice base their suggestion on three spontaneous confabulators reported in the literature, who all performed poorly on the Cognitive Estimates Test (a measure of reasoning ability). But the authors failed to cite contradictory evidence from the Shapiro et al. sample of seven confabulators, in whom there was no correlation between performance on the Cognitive Estimates Test and the severity/fantasy of confabulation [49].
Schacter et al. similarly referred to dysfunction in the ‘constructive memory framework’ [53]. They specified the role of disrupted editing at retrieval, but also proposed that faulty encoding increases the possibility of confabulation at recall. Parkin proposed a similar view, suggesting that the variability in presentation of confabulations might be caused by deficits in the memory encoding or retrieval phase, with the underlying mechanism dependent on lesion site [54].
In their important study, however, Schnider and Ptak challenged this view, demonstrating that confabulating and non-confabulating amnesic patients could not be differentiated in terms of encoding abilities (or executive functioning for that matter) [14]. They proposed that spontaneous confabulation results from an inability to suppress currently irrelevant information. The authors presented six spontaneous confabulators, 12 non-confabulators with similarly severe amnesia, and 10 healthy controls with a series of 80 drawings among which some reappeared several times. Subjects were asked to indicate picture reappearances. They were then given a second run of the same task, but asked to forget that they had already seen the pictures, and to point out only recurrences within the second run. Thus, subjects were being asked to distinguish between recurrences in the second run (‘ongoing reality’) from memories of the drawings’ previous occurrence in the first run (‘currently irrelevant memories’). Spontaneous confabulators were found to have higher false positive responses in the second run, indicating an inability to suppress irrelevant memories. They also showed the same ability as non-confabulating amnesics to detect true positives in the second run, thus challenging the idea that spontaneous confabulators have difficulty with encoding. The Schnider and Ptak theory is further supported by the observation that improvement in the ability to suppress irrelevant memories parallels improvement in spontaneous confabulation [55], and by studies that showed that monkeys with orbitofrontal lesions failed to suppress habits that are no longer rewarding [56].
Gilboa et al. [57] built on the work of Schnider et al. They designed a study to test their hypothesis that deficits in strategic retrieval, rather than temporal confusion, were central to spontaneous confabulation. They performed experiments on four confabulating patients with ACoA aneurysms, eight non-confabulating patients with ACoA, and 16 normal age-matched controls. The rigorous design added to those used by Schnider et al. [12], [14], with tests of remote semantic and autobiographical memory. The results support the view that temporality confusion is less important than memory retrieval in confabulation, with the interesting finding that confabulation can involve semantic memory. Gilboa et al., echoing Hirstein's concept of ‘pathological certainty,’ further proposed that an abnormal pre-conscious ‘feeling of rightness’ results in confident acceptance of false memories. This makes intuitive sense, but Gilboa et al. later admitted that it may not be a full explanation, and that abnormality in multiple monitoring systems may be required for confabulation to occur.
Fotopoulou et al. appear to subscribe to the disrupted editing theory [23]. They proposed that deficits in memory monitoring and irrelevant memory inhibitory mechanisms provide the soil for emotional motivational factors and ‘kernels of truth’ to grow into confabulatory material. This entails an amplification of normal motivational mechanisms in influencing memory recall, combined with abnormal inhibitory function. Fotopoulou et al. also suggested that temporality monitoring may be impaired in spontaneous confabulation.
Temporal order confusion
While closely related to the first, the second theoretical approach focuses on errors in the temporal ordering of memory caused by the confabulator's inadequate temporal frame of reference. As such, true memories may be transposed to the present or immediate past, or memory parts may be incorrectly recalled together. This is not a new idea, in fact Korsakoff refers to the relationship between temporal sequencing deficits and confabulation in cases reported in 1889 [20]. In a review article Parkin noted that confabulation featuring temporal ordering problems is often associated with frontal lobe dysfunction [54]. But this is not always the case. Other research suggests that spontaneous confabulators can have impaired ability to recognize the temporal order of stored information but have relatively preserved executive functioning compared to non-confabulating amnesic controls [12].
The Dalla Barba et al. slant on the temporal ordering approach refers to primary dysfunction of ‘temporal consciousness’. Such individuals retain awareness of a present, past and future, but when asked about recent or prospective memories they tend to reply with routine or habits from long-term memory stores, regardless of their present relevance [34]. They suggested that a temporal gradient plays a part in confabulation, and that recent memories are more vulnerable to confabulation than old memories because they are ‘weaker’. They reported the case of an amnesic woman who had an increased frequency of provoked confabulations for recent versus distant personal events [48]. Researchers adopting the disrupted editing approach, however, highlighted shortcomings of the temporal ordering approach, claiming that temporal ordering errors are symptoms rather than the cause of confabulation, with the underlying cause being poor strategic retrieval [13]. The Gilboa et al. findings in particular would suggest that temporality confusion is not pivotal in causation, although it may well be a symptom of confabulation.
Neither of the previous two approaches alone can account for all of the variable presentations of spontaneous confabulation. While some cases indicate inappropriate memory retrieval, others point to perseveration or unchecked responses. Furthermore, not all individuals with frontal damage confabulate, nor do all those with temporal ordering impairment.
Multiple deficits
A broader view underpins the third and final group of theories. Shapiro et al. had earlier presented a multiple deficits model, proposing that impairment in different aspects of frontal lobe and executive functioning could lead to spontaneous confabulation [49]. In their study of seven confabulating men, the four severe confabulators had high levels of perseveration, impaired self-monitoring, and failure to inhibit incorrect responses. Perseveration, in particular, was most closely related to confabulation and paralleled its resolution. Interestingly, the severe confabulators had some improvement when prompted by cues. Later, Johnson et al. related confabulations to a broader range of deficits involving the encoding and retrieval stages of memory and highlighted the influence of motivation and judgment processes [58]. Using her approach, confabulation relates to a disturbance of consciousness, with disruption to the sense of self and reality monitoring. As illustrated in their vignette of the patient ‘GS’, Johnson et al. elaborated their view by proposing that spontaneous confabulation relates to the interaction between a vivid imagination, an inability to retrieve autobiographical information systematically, and source monitoring deficits [16]. Due to the complexity of frontal lobe functioning, it is likely that the various deficits shown by confabulators reflect interacting dysfunction in different frontal mechanisms. The strength of Johnson's work lies in its consideration of normal memory processes, in particular, her ‘source monitoring framework’ and her concepts of ‘heuristic’ and ‘systematic’ processing of memory. More data from spontaneous and fantastic confabulators, however, are needed to support her hypotheses.
In summary, research stemming from various theoretical approaches is helpful. The various definitions identified by these theories may plausibly reflect dysfunction to components of frontal control mechanisms. These approaches, however, are yet to yield an encompassing understanding of the mechanisms behind spontaneous confabulation. This is well illustrated in the Burgess and McNeil case study in which no one theory could fully account for the content-specific form of spontaneous confabulation they observed [47]. The Fotopoulou et al. contributions attempt to address the issue of content in spontaneous confabulation [23]. Theories continue to be refined, and, most recently, Schnider et al. have proposed that the defect behind spontaneous confabulation precedes conscious memory processing, and involves various mechanisms including decreased verification and monitoring of memories, and vague encoding [59]. Similarly, in a useful review of current thinking, Metcalf et al. offered a ‘three factor cognitive-neuropsychological model’ of confabulation, consisting of (i) impaired retrieval; (ii) motivational biases; and (iii) impaired editing [60]. They argue that impaired retrieval and impaired editing are likely in most cases of spontaneous confabulation, while motivational biases and the general organization of autobiographical memory influence confabulatory content. Their contribution is perhaps the best articulated version of the multiple deficits model.
Course and treatment
In most cases, spontaneous confabulation is transient, with duration dependent on lesion site and nature [61]. Anatomical and cognitive factors have been proposed as underpinning the resolution of confabulations. In confabulation following acquired brain injury with frontal involvement, symptoms may decrease as perfusion to the damaged regions normalizes [45]. Likewise, objectively measured resolution of the ability to suppress memories that are irrelevant to the current situation corresponds with a reduction in confabulations [41]. Because spontaneous confabulation is most often transient, and resolves without intervention, treatment and rehabilitation approaches have gained little attention in the literature [41]. But when symptoms persist and become chronic or are associated with behaviours of concern, intervention may be required.
Behavioural interventions focusing on self-monitoring have been shown to be effective in managing the negative effects of spontaneous confabulations and reducing their occurrence. The actual confabulation is approached indirectly, by reducing associated behaviours. As a general rule, it is better to avoid confrontation because it is likely to upset the confabulator, whose conception of reality is sincerely held [41]. Dayus and van den Broek reported on a patient who developed several spontaneous confabulations associated with increased anger and swearing, after injury, that persisted for 6 years [62]. They used a self-monitoring technique combined with positive reinforcement, an approach that had previously been found to be effective in managing dysexecutive behavioural problems [63]. The patient was taught to monitor the frequency of swear words during conversation and given positive reinforcement as accuracy increased. The resultant reduction in instances of swearing was associated with a reduction in his confabulatory outbursts. The authors claimed that because poor self-monitoring is proposed to be behind confabulation, the general training in self-monitoring reduced instances of confabulations by side-stepping the symptoms of poor executive functioning. Burgess and McNeill also reported on the efficacy of a self-monitoring intervention in reducing spontaneous confabulations [47]. Their patient had a very specific confabulation that persisted for 12 weeks after injury and was becoming problematic. He believed that he had to perform a stocktake for a local shop and his wife's role in preventing him from acting on this was causing conflict between them. He was taught to record all daily events in a diary. This was effective in improving his monitoring of memories, and in the 5 weeks of treatment the behaviours of concern stopped.
While most reported treatments for confabulation are behavioural, there is one case report of it responding to pharmacological treatment. For their patient, Pihan et al. found that risperidone was effective in eliminating spontaneous confabulations that had persisted for 4 months after injury [21]. The confabulations resolved in a dose-dependent fashion, with associated improvement in the ability to adapt thought and behaviour to ongoing reality. The authors suggested that the dopamine antagonist may have acted to reduce the patient's tendency to maintain irrelevant mental associations by acting on frontal systems functioning.
Conclusion
Spontaneous confabulation is a striking clinical presentation, but remains a phenomenological concept continually evolving as our understanding of neurocognition improves. Spontaneous confabulation remains ill defined, but seems closely associated with amnestic disorder and frontal dysfunction. There is no current model that sufficiently explains aetiological mechanisms, although a multiple deficits model seems most likely. The course of spontaneous confabulation is largely related to the associated clinical picture in which it presents. Although difficult to research, future study will further illuminate our understanding.