them from my hand, with attendant pain! This doesn't happen of 

 course, because my trained system knows all about the danger. So 

 even when my attention is otherwise directed toward a conversation 

 while I am eating, my system automatically prevents me from in- 

 voking and activating a finger-biting schema. I'm sure that I did bite 

 my finger a few times during infancy, however. I suspect that the 

 system incorporates knowledge of such adverse experiences (and of 

 course their positive counterparts) into its repertoire during dream- 

 ing, and it is interesting to note that infants spend much more time 

 than adults in the dreaming state. In his James Arthur Lecture, Matt 

 Cartmill (1996) opted for the ingenious theory of dreaming put for- 

 ward by Francis Crick and Graeme Mitchison (1983), which pos- 

 tulates that it serves to eliminate accidentally formed — and un- 

 wanted — memories. I find it difficult to believe that a system prone 

 to such erroneous storage would have survived in the evolutionary 

 arena. 



My view appears to harmonize with results reported by Robert 

 Stickgold (1998). who found evidence for consolidation and inte- 

 gration of memories during sleep, including its dreaming phase, and 

 highlighted the interaction of the hippocampus (see fig. 2) with the 

 rest of the cerebral cortex. Brenda Milner (1967) had demonstrated 

 the major role played by the hippocampus in the laying down of 

 long-term memory, through her observations of the patient H.M. 

 following bilateral removal of his hippocampus. These observations 

 were subsequently reinforced and extended through Larry Squire's 

 ( 1992) studies of members of several different mammalian species, 

 all of which had had the hippocampus removed. It is now widely 

 believed that the hippocampus serves as a temporary depository for 

 short-term memories that are candidates for longer-term memory 

 storage, the actual storage being elsewhere — and widely distribut- 

 ed — in the cerebral cortex. Alexel Egorov, Bassam Hamam, Erik 

 Fransen, Michael Hasselmo and Angel Alonso (2002) have recently 

 discovered how this might be coded for, because they have found 

 hippocampal neurons whose activity level depends in a stepwise 

 manner on how much input they have received during the previous 

 few seconds. 



Dreaming sleep is now more commonly referred to as rapid-eye- 



22 



