50 BRAIN MECHANISMS AND LEARNING 



learning these large numbers that Dr Hcbb was discussing, but very complex when 

 one compares it with the kind ot learning task the insect normally has to achieve. 

 1 think that there is a very real problem here, namely that of explaining the very 

 rapid mastery by the insect brain ot a substantial task in serial learning. And then of 

 course one must remember that there are no synaptic knobs, so far described, in 

 the insect's nervous system which could be acting m the way that Milncr proposed. 



EccLES. If I understood Dr Hebb correctly, he said he couldn't see how he could 

 explain some of the digital learning processes on the basis of synaptic trace changes. 



Hebb. Only when dealing with synapses already so highly developed as must be 

 the case in repeating a sequence as much practised as 1-2-3. Here a single exposure, 

 a single repetition, could hardly make a significant additional change. 



EccLES. Against that concept I would like to say that in the very simplest cerebral 

 actions w^e are using millions of neurones in the most complicated imaginable 

 patterns. In, say the 1-2-3 sequence, we don't have a group of neurones that are 

 related to i, another group for 2, and another for 3. I think that in each digital 

 association the most complicated neuronal network is in operation. With a 2-1-3 

 sequence, we would have a quite different asseinblage of neurones in activity. It is 

 not sutf iciently realized what an immense number of neurones we have to draw 

 upon for the simplest memory. Lashley says that in the simplest engram there are 

 millions of neurones involved. I would go much further than that myself There is 

 evidence from EEG record. During mental arithmetic there may be disturbance of 

 the EEG over a wide region of the cerebral cortex. Thus immense assemblage of 

 neurones are used doing some unusual multiplication, say 23 X43. Therefore I 

 don't subscribe to Dr Gerard's idea that there is a functional neurone reserve. I like 

 to think that we are using all the neurones in some kind of pattern or other and 

 using them thousands of times over in patterns, but that we can still use them many 

 more thousand times over. It is the pattern that is important, not the neurones. 



Gerard. This is not contrary to my notion, except that the more neurones you 

 have the more patterns you can deal with. 



Eccles. I agree entirely. One final point, and that is if there is electrical interac- 

 tion, and we have seen from Dr Estable's work the complexity of connections, 

 and we now know from the electronmicroscopists that there is no free space, only 

 200 A. clefts, everywhere in the central nervous system, then everything should be 

 electrically interacted with everything else. I think this is only electrical background 

 noise and, that when we lift with specific chemical connections above that noise we 

 get a significant operational system. I would say that there is electrical interaction 

 but it is just a noise, a nuisance. 



Hebb. I think we are in agreement, at least in part, if I understood this last com- 

 ment. When you say that the whole cortex is thrown into action when somebody 

 multiplies 23 X 43 you are suggesting that most of the activity contributes nothing 

 positive. The Jasper, Ricci and Doane experiments at least suggest strongly that the 

 course of learning is throwing out the neurones that are irrelevant, keeping them 

 out of the way wliile the others do the job. And this may mean that the course of 

 learning involves many more neurones than would be desirable. It may be that 

 though the operation is very complex, it could be done better if there were no other 

 neurones present, for that operation. But the great characteristic of the human 

 brain is the extraordinary variety and complexity of things it can deal wih. 



Gerard. I might add, since, the work of Jasper's group has been mentioned so 



