Anatomy of a false memory

WE BELIEVE THAT memory provides us with a faithful record of past events. But in fact, it is well established that memory is reconstructive, and not reproductive, in nature. In retrieval, a memory is pieced together from fragments, but during the reconstruction errors creep in due to our own biases and expectations.

Generally, these errors are small, so despite not being completely accurate, our memories are usually reliable. Occasionally, there are too many errors, and the memory becomes unreliable. In extreme cases, memories can be completely false.

False memory, or confabulation, is completely unintentional, and can occur spontaneously due, for example, to the suggestive power of a leading question or a doctored photograph. It can also occur following frontal lobe damage due to tumours, head injuries, or ruptured arteries. 

An early description of confabulation is given by the Russian neuropsychiatrist Sergei Korsakoff, who recognized it in chronic alcoholics, and describes it (and the severe amnesia that is also associated with chronic alcoholism), as follows:

This mental disorder appears at times in the form of sharply delineated irritable weakness of the mental sphere, at times in the form of confusion with characteristic mistakes in orientation for place, time and situation, and at times as an almost pure form of acute amnesia, where the recent memory is most severely involved, while the remote memory is well preserved . . . Some have suffered so widespread memory loss that they literally forget everything immediately.

Studies of confabulation have providing confusing results, mainly because they have implicated several different regions of the frontal lobes. Not only have they failed to localize confabulation, but they have also reached no firm conclusions about the specific cognitive deficits involved.

A new study now reconciles these conflicting data, by showing that the different regions of the brain previously implicated are involved in different kinds of memory errors. It also pinpoints a specific region as being involved in false memories, and could help researchers better understand how the brain controls memory.

Case studies of amnesic patients have implicated two distinct regions of the frontal lobes: the orbitofrontal cortex, so named because it lies on the roof of the orbits of the eyes, and the ventromedial cortex, which is immediately adjacent. These studies have found no evidence of lateralization: confabulation was found to occur following damage to either the left or the right frontal lobes, or both.

Functional neuroimaging studies of the processes that control memory have provided conflicting results. They implicate areas on the lateral (outer) surface of the frontal lobe of the right hemisphere, areas which are believed to be involved in specifying the cues for memory retrieval and in monitoring the appropriateness of those cues.

Furthermore, the functional deficits that result in confabulation have not been determined - it could occur largely as a result of a memory deficit, or of dysfunction in the executive control centres in the frontal lobes, which are involved in the planning and execution of other cognitive processes, or of a comination of the two.

The new study, led by Martha Turner of the Institute of Cognitive Neuroscience at UCL, involved 38 patients with localized damage in some part of the frontal lobes. 12 patients with damage to other parts of the brain, plus 50 healthy subjects, were included as controls.

The participants performed a battery of tests designed to assess different types of memory. General semantic memory was tested with questions such as "What happened to President Kennedy", and personal semantic memory with questions such as "What is your address?". Orientation in time ("What month is it?") and space ("What city are you in?") were also tested, as was executive function.


Locations of lesions in brain-damaged patients. Shaded areas represent the proportion of patients with damage affecting 25% or more of the depicted region.

It was found that participants with frontal lobe damage consistently produced a greater number of confabulations than the controls. 3 of the patients with frontal lobe damage even produced spontaneous false memories frequently while not being tested. In general, the greater the extent of the damage, the more confabulations were produced, and the false memories were most often produced in response to questions testing personal episodic memory and orientation in time. By contrast, none of the patients with damage to other lobes of the brain produced significantly more false memories than the healthy controls.

Another pattern emerged when the test results were grouped according to the site of the lesion: those with damage to the orbital, medial or left lateral regions of the frontal lobes produced significantly more confabulations of personal episodic memory, whereas those with damage to the orbital, medial or right lateral frontal lobes produced more confabulations of orientation in time.

When the data were grouped according to the total number of confabulations, it was found that all the patients who produced significantly more false memories than controls had damage in the ventromedial prefrontal cortex, either in the orbital region of in the anterior cingulate gyrus. These "high confabulators" could be distinguished from "low confabulators" on the basis of 6 memory tests and 2 measures of executive functioning.

Thus, confabulation is strongly associated with damage to the ventromedial prefrontal cortex. Memory impairments seem to be necessary for the production of false memories, but executive function is less important. However, although the study localizes confabulation to the ventromedial prefrontal cortex, further work will be needed to determine whether it occurs as a result of impairments in memory or executive function, or both. 

If confabulation occurs following damage to the ventromedial cortex, what functions might this part of the brain be involved in? Some researchers have suggested that it normally suppresses memories that are not relevant to the current situation, while others argue that it acts as a monitoring system which normally rejects false memories that don't "feel right".


Turner, M.S., Cipolotti,, L., Yousry, T.A., Shallice, T. (2008). Confabulation: Damage to a specific inferior medial prefrontal system. Cortex 44: 637-648. [PDF]

More like this

Having had an exceptionally large meningioma excised 3 years ago that originated at the right temporal/parietal, then extended toward the vertex and to the central venous sinus, I noted something retrospectively. Before I was aware of its presence, I experienced VERY vivid, extremely detailed dreams that were entirely complete in all sensory aspects, to the extent that it became difficult to seperate them from actual events of distant past events. During the day, I would have what could probably be best described as 'absence seizures' wherein the same would occur during the day. It leaves me wondering if the increase in ICP could have been the underlying etiology of this phenomenon (I had > 35% displacement with a large midline shift), as I was not doing this with deliberate intent. My wife was aware of it as it was happening during the few days before I was diagnosed, as it became quite severe for the last 3-4 days. Your thoughts?

By i-have-half-a-… (not verified) on 15 Jun 2008 #permalink