Richard Morris, neuroscientist at the University of Edimburgh, has recently put forth a third model: the "scheme theory". Morris starts with the observation that it is very difficult for the brain to consolidate certain kinds of information which it is not ready to acquire. Let us take the example of an economist attending a lecture given by an astrophysicist. The economist will eventually understand the speaker's talk without having been briefed beforehand, but will retain hardly anything the minute he leaves the lecture hall. He will have retained even less the following day. On the other hand, Richard Morris continues, a specialist in the field will encode any new information directly in the cortex without going through the hippocampus.

In other words, there might be prior mnesic traces present in the cortex of the expert which could be slightly modified in order to make allowance for the new element. The hippocampo-cortical system would only come into play when the information is new and thus foreign to the mental structure of the individual. In this case, memory consolidation would take place in accordance with the classic theory.

The Passing of the Baton

Neuroscientists at the University of Liège CRC have based their research on the classic theory, which continues to have the most proponents today. They hypothesize that the hippocampus and thus the cortexes where the different constituents of a declarative memory are encoded are spontaneously reactivated, preferentially during sleep, resulting in the long-term reorganization of the mnesic trace in the cortical systems.

For their experimental work, Pierre Maquet and his team made up two groups of volunteers who were asked to memorize ninety pairs of easy-to-visualize words. Each pair of words belonged to the same lexical field, e.g. "tractor" and "steering wheel" but there was no obvious link between them as in "dog" or "cat" for instance. The two volunteer groups differed in that those in the second group were deprived of sleep during the night following the memorization of the words.

The volunteers' cerebral responses were measured by functional Magnetic Resonance Imagery (fMRI) during the training (encoding) phase and several times thereafter. As free verbal recall (of words) is impossible in the scanner - the simple fact of articulating a word makes the brain move and drastically distorts the image - the researchers presented the subjects with one word of each pair and asked them to press a button when they remembered the corresponding second term. To ensure that the experiment was valid, the subjects were asked to redo the free recall test once the scan was completed to verify whether the words they had indicated they remembered were indeed correct.