A Mechanism of Confabulation Revealed

Abstract & commentary

Source: Schneider A, Ptak R. Spontaneous confabulators fail to suppress currently irrelevant memory. Nat Neurosc 1999; 2(7):677-681.

Recognizing confabulation can be useful in the differential diagnosis of amnesia. A new study by Schneider and Ptak suggests that the mechanism underlying spontaneous confabulation may be a deficiency in the ability to suppress activated memory traces that are inappropriate to the current context. In essence, amnesiacs who confabulate may be activating too many memories rather than too few.

Schneider and Ptak examined six amnesiacs who acted according to invented stories (spontaneous confabulators) and compared them to 12 comparably amnesiac patients who did not confabulate, as well as 10 normal controls. Spontaneously confabulating patients had abnormalities in the basal forebrain or medical orbital frontal cortex. The subjects were shown various sets of pictures and asked to identify a target item that appeared recurrently among singly viewed distrators. The same sets of stimuli were used in multiple runs, with previous target items interchanged with distrators. Before each run, subjects were instructed to forget the pictures they had seen before, and to only identify recurrences within the given run.

On this continuous recognition task, all of the amnesiacs performed significantly worse than controls, making more false-positive responses. However, confabulating amnesiacs showed steeply increasing numbers of false-positive responses as the experiment progressed. When the interval between successive runs was increased to 30 minutes, the false-positive response rate of confabulators remained high. Interference by previously acquired information was most evident when the stimuli represented real world objects, but was also apparent when meaningless designs were used.

Schneider and Ptak concluded that confabulating and nonconfabulating amnesiacs did not differ in their ability to detect new target items, indicating that confabulation is not simply a consequence of failure to saliently represent incoming information. What did distinguish confabulators was their inability to suppress interference arising from previously acquired information. Confabulators may process information encountered 30 minutes ago as though it were part of their experience of the present moment.

Comment by Norman R. Relkin, MD

Schneider and Ptak provide a new slant on the memory-monitoring deficit hypothesis. Based on their findings, spontaneous confabulators may not be able to suppress mental associations that pertain to past events or fully distinguish them from those that arise in the present. Information acquired days, weeks, months, or even years before may intrude into their current thinking, leading to the bizarre, false ideas that constitute confabulations.

This appealing theory explains some but not all of the phenomenology observed in amnesiac confabulators. Perhaps the most famous bedside clinical test for confabulation is the "purple string test," in which the examiner pretends to stretch a string between their own hands, and asks the amnesiac whether they can see the nonexistent purple string. Confabulators, particularly those with acute Korsakoff’s syndrome, often state that they see the string and that it appears purple. If suggestive statements and a beguiling hand position are all that one needs to convince the amnesiac patient that they are seeing something that isn’t really there, it would seem that confabulation represents more than just a failure to distinguish past from present associations. Confusion about what has been seen in the past vs. the present could be invoked as a partial explanation.

The brain lesions identified in these confabulating amnesiacs were all located in fairly circumscribed locations, distinct from those found in amnesiacs who did not confabulate. The salient areas lie in or near midline structures comprising the anterior limbic system, including the medial orbital frontal cortex and hypothalamus. This confirms that confabulatory amnesia usually arises from dysfunction in midline brain structures, rather than medial temporal and lateral prefrontal areas involved in other forms of forgetfulness. (Dr. Relkin is Assistant Professor of Neurology, New York Hospital-Cornell Medical Center.)