Sleep : the architect of memory
Scientists at the University of Liège's Cyclotron Research Centre have tried for several years to elucidate the links between sleep and memory. In two recent articles published in PNAS and PLOS Biology respectively, the researchers have shed new light on long-term memory consolidation; one of these papers pays particular attention to emotional memories.
Pierre Maquet, Director of Research at the FNRS, and his team have been studying the relation between sleep and memory for several years. The results of their research, carried out for the most part at the University of Liège's Cyclotron Research Centre, have been reported in different articles published in prestigious journals such as Nature Neuroscience, Neuron or The Journal of Neuroscience. In their study, the researchers show how sleep plays a unique role in consolidation. Sleep deprivation during the night after diurnal learning significantly decreases our mnestic performance.
Results of new research conducted by Pierre Maquet's team have recently been published in PNAS (Proceedings of the National Academy of Sciences of the United States of America – November 2007) and PLOS Biology (October 2007).
The article published in PNAS addresses the link between sleep and long-term memory consolidation. If the precise nature of this research is to be understood, it is necessary to provide some background information on the case of H.M., a famous patient in the field of neuropsychology. H.M. suffered from severe epilepsy, the focus of which was localized in the two hippocampi, one per each cerebral hemisphere. The patient underwent a surgical resection of these two regions in the 1950's which caused the epilepsy to disappear, but at a heavy price indeed. H.M. began to suffer from both anterograde amnesia (the inability to acquire new memories) and retrograde amnesia, which has prevented him since from having any recollection of events that took place before the temporal lobectomy. Two types of memories have not been affected, however: the oldest ones and those involving procedural learning.
The case of H.M.shows, Pierre Maquet explains, that the hippocampus is very closely related to a particular form of memory called declarative memory and that our oldest memories are no longer stored in the hippocampus.
H.M.'s lost memories covered a period of several years previous to the hippocampus resection. Other case studies have been done since of patients diagnosed with retrograde amnesia, in particular following herpetic encephalitis that damages the hippocampus of each cerebral hemisphere. It has been observed that the period of total memory loss varies from case to case and may cover a few weeks up to several years. In other words, our memories are dependent for a while on the hippocampus which, according to the prevailing hypothesis, could play a transitory role in declarative memory.
Let us consider a declarative memory of a strictly episodic nature such as : Yesterday, I had a nice steak at a restaurant with my cousin. Each component of the memory is likely encoded in a different region of the cortex, my cousin's face, for instance, in the fusiform gyrus. Each cortical area, then, would be activated so as to participate in the encoding of a particular component of the memory. According to the prevailing theory of mnesic consolidation of the episodic memory, the hippocampus plays the role of orchestra conductor, linking the encoded elements together to form a unique episode.
According to this theory, a memory of this type would need to be progressively reorganized within the cerebral circuits in order to be retained long term. A possible reorganization mechanism would consist in the reactivation of the hypocampus and then of the cortical areas involved in the memory, which would discharge simultaneously. According to Hebb's postulate, the regions of the cortex that become active synchronously connect with each other. The classic theory of mnesic consolidation of episodic memory hold that as reactivations continue, intracortical connections get stronger and hippocampo-cortical connections become weaker, so that episodic memory becomes essentially dependent on the cortex in the long run.