Strategic retrieval, confabulations, and delusions: Theory and data

Abstract

Based on Moscovitch and Winocur's “working with memory” framework, confabulation is described as a deficit in strategic retrieval processes. The present paper suggests that only a confluence of deficits on multiple memory-related processes leads to confabulation. These are divided into three categories. Core processes that are unique to confabulation and required for its evolution include: (1) an intuitive, rapid, preconscious “feeling of rightness” monitoring, (2) an elaborate conscious “editor” monitoring, and (3) control processes that mediate the decision whether to act upon a retrieved memory. The second category is deficits on constitutional processes which are required for confabulation to occur but are not unique to it. These include the formation of erroneous memory representation, (temporal) context confusion, and deficits in retrieval cue generation. Finally, associated Features of confabulations determine the content “flavour” and frequency of confabulation but are not required for their evolution. Some associated features are magnification of normal reconstructive memory processes such as reliance on generic/schematic representations, and positivity biases in memory, whereas others are abnormal such as perseveration or source memory deficits. Data on deficits in core processes in confabulation are presented. Next, the apparent correspondences between confabulation and delusion are discussed. Considering confabulation within a strategic memory framework may help elucidate both the commonalities and differences between the two symptoms. Delusions are affected by a convergence of abnormal perception and encoding of information, associated with aberrant cognitive schema structure and disordered belief monitoring. Whereas confabulation is primarily a disorder of retrieval, mnemonic aspects of delusions can be described as primarily a disorder of input and integration of information. It is suggested that delusions might share some of the associated features of confabulation but not its core and constitutional processes. Preliminary data in support of this view are presented.

Keywords:

• Ventromedial prefrontal cortex
• Confabulation
• Memory monitoring
• Strategic retrieval
• Working With Memory

Acknowledgements

I would like to thank Drs. Morre Goldsmith and Vered Galon for the use of the QAP and Sarah Adler for testing schizophrenic patients. I would also like to thank Dr. Morris Moscovitch for his inspiring mentorship. AG is supported by an EU FP06 Marie Curie IRG grant MIRG-CT-2006-046544.