Confabulations are emotionally charged, but not always for the best

It's not a lie … If you believe it: Narrative analysis of autobiographical memories reveals over-confidence disposition in patients who confabulate

Humans perceive their personal memories as fundamentally true, and although memory is prone to inaccuracies, flagrant memory errors are rare. Some patients with damage to the ventromedial prefrontal cortex (vmPFC) recall and act upon patently erroneous memories (spontaneous confabulations). Clinical observations suggest these memories carry a strong sense of confidence, a function ascribed to vmPFC in studies of memory and decision making. However, most studies of the underlying mechanisms of memory overconfidence do not directly probe personal recollections and resort instead to laboratory-based tasks and contrived rating scales. We analyzed naturalistic word use of patients with focal vmPFC damage (N = 18) and matched healthy controls (N = 23) while they recalled autobiographical memories using the Linguistic Inquiry and Word Count (LIWC) method. We found that patients with spontaneous confabulation (N = 7) tended to over-use words related to the categories of 'certainty' and of 'swearwords' compared to both non-confabulating vmPFC patients (N = 11) and control participants. Certainty related expressions among confabulating patients were at normal levels during erroneous memories and were over-expressed during accurate memories, contrary to our predictions. We found no elevation in expressions of affect (positive or negative), temporality or drive as would be predicted by some models of confabulation. Thus, erroneous memories may be associated with subjectively lower certainty, but still exceed patients' report criterion because of a global proclivity for overconfidence. This may be compounded by disinhibition reflected by elevated use of swearwords. These findings demonstrate that analysis of naturalistic expressions of memory content can illuminate global meta-mnemonic contributions to memory accuracy complementing indirect laboratory-based correlates of behavior. Memory accuracy is the result of complex interactions among multiple meta-mnemonic processes such as monitoring, report criteria, and control processes which may be shared across decision-making domains.

Anomalies of Autobiographical Memory

OBJECTIVES

In this paper, I review three 'anomalies' or disorders in autobiographical memory: neurological retrograde amnesia (RA), spontaneous confabulation, and psychogenic amnesia.

METHODS

Existing theories are reviewed, their limitations considered, some of my own empirical findings briefly described, and possible interpretations proposed and interspersed with illustrative case-reports.

RESULTS

In RA, there may be an important retrieval component to the deficit, and factors at encoding may give rise to the relative preservation of early memories (and the reminiscence bump) which manifests as a temporal gradient. Spontaneous confabulation appears to be associated with a damaged 'filter' in orbitofrontal and ventromedial frontal regions. Consistent with this, an empirical study has shown that both the initial severity of confabulation and its subsequent decline are associated with changes in the executive function (especially in cognitive estimate errors) and inversely with the quantity of accurate autobiographical memories retrieved. Psychogenic amnesia can be 'global' or 'situation-specific'. The former is associated with a precipitating stress, depressed mood, and (often) a past history of a transient neurological amnesia. In these circumstances, frontal control mechanisms can inhibit retrieval of autobiographical memories, and even the sense of 'self' (identity), while compromised medial temporal function prevents subsequent retrieval of what occurred during a 'fugue'. An empirical investigation of psychogenic amnesia and some recent imaging studies have provided findings consistent with this view.

CONCLUSIONS

Taken together, these various observations point to the importance of frontal 'control' systems (in interaction with medial temporal/hippocampal systems) in the retrieval and, more particularly, the disrupted retrieval of 'old' memories.

Tell about yourself to improve your autobiographical memory: A study of Korsakoff's syndrome

Korsakoff's syndrome (KS) has been associated with a difficulty to retrieve specific autobiographical memories. We investigated whether this difficulty can be alleviated after the retrieval of statements describing self-images. KS patients and control participants were recruited and asked to retrieve autobiographical memories after providing statements to the question "Who am I?" and after a control condition consisting of verbal fluency. Analysis showed higher autobiographical specificity in the "Who am I?" than in verbal fluency condition in both patients with KS and control participants. At a theoretical level, our findings demonstrate how retrieval of information related to conceptual self may influence autobiographical memory in KS. At a clinical level, our procedures are important as they demonstrate how a simple task (i.e., "Who am I?" statements) may serve as a tool to cue specific autobiographical memories in patients with KS.

Confabulation: Developing the 'emotion dysregulation' hypothesis

Confabulations offer unique opportunities for establishing the neurobiological basis of delusional thinking. As regards causal factors, a review of the confabulation literature suggests that neither amnesia nor executive impairment can be the sole (or perhaps even the primary) cause of all delusional beliefs - though they may act in concert with other factors. A key perspective in the modern literature is that many delusions have an emotionally positive or 'wishful' element, that may serve to modulate or manage emotional experience. Some authors have referred to this perspective as the 'emotion dysregulation' hypothesis. In this article we review the theoretical underpinnings of this approach, and develop the idea by suggesting that the positive aspects of confabulatory states may have a role in perpetuating the imbalance between cognitive control and emotion. We draw on existing evidence from fields outside neuropsychology, to argue for three main causal factors: that positive emotions are related to more global or schematic forms of cognitive processing; that positive emotions influence the accuracy of memory recollection; and that positive emotions make people more susceptible to false memories. These findings suggest that the emotions that we want to feel (or do not want to feel) can influence the way we reconstruct past experiences and generate a sense of self - a proposition that bears on a unified theory of delusional belief states.

Is there a positive bias in false recognition? Evidence from confabulating amnesia patients

Although there is some evidence for a positive emotional bias in the content of confabulations in brain damaged patients, findings have been inconsistent. The present study used the semantic-associates procedure to induce false recall and false recognition in order to examine whether a positive bias would be found in confabulating amnesic patients, relative to non-confabulating amnesic patients and healthy controls. Lists of positive, negative and neutral words were presented in order to induce false recall or false recognition of non-presented (but semantically associated) words. The latter were termed 'critical intrusions'. Thirteen confabulating amnesic patients, 13 non-confabulating amnesic patients and 13 healthy controls were investigated. Confabulating patients falsely recognised a higher proportion of positive (but unrelated) words, compared with non-confabulating patients and healthy controls. No differences were found for recall memory. Signal detection analysis, however, indicated that the positive bias for false recognition memory might reflect weaker memory in the confabulating amnesic group. This suggested that amnesia patients with weaker memory are more likely to confabulate and the content of these confabulations are more likely to be positive.

Puzzling Confabulations – An Overview of Classifications and Theories

Abstract. Confabulations comprise a puzzling collection of false statements, produced without conscious intent to deceive. They have been classified according to the mode of elicitation (spontaneous vs. provoked), the content (fantastic vs. plausible), the memory domain in which they become manifest (episodic vs. semantic), their stability (stable vs. ephemeral), or their selectivity (monothematic vs. multithematic). All classifications appear to be problematic, because there are no clear-cut dichotomies, and confabulations often seem to fall into overlapping categories. There are, in fact, many theories, presumably explaining different kinds of confabulations; the best bet is to regard them as complementary rather than competing.

Memory consolidation in humans: new evidence and opportunities

We are endlessly fascinated by memory; we desire to improve it and fear its loss. While it has long been recognized that brain regions such as the hippocampus are vital for supporting memories of our past experiences (autobiographical memories), we still lack fundamental knowledge about the mechanisms involved. This is because the study of specific neural signatures of autobiographical memories in vivo in humans presents a significant challenge. However, recent developments in high-resolution structural and functional magnetic resonance imaging coupled with advanced analytical methods now permit access to the neural substrates of memory representations that has hitherto been precluded in humans. Here, I describe how the application of 'decoding' techniques to brain-imaging data is beginning to disclose how individual autobiographical memory representations evolve over time, deepening our understanding of systems-level consolidation. In particular, this prompts new questions about the roles of the hippocampus and ventromedial prefrontal cortex and offers new opportunities to interrogate the elusive memory trace that has for so long confounded neuroscientists.