GENERAL DISCUSSION 66 I 



coiivcrs^cncc centre (CC) and a neuronal network (NN) whereas we think that 

 convers;ences and associative interactions can take place within the same neuronal 

 assemblies as those present in the brani stem and non-specific diencephalic nuclei. 



hi mv opinion, the weakness ot all models so far proposed is that they usually 

 neglect the states of autogenic repetitive firing which are so common in the nerve 

 cells of cerebral structures in waking animals. This I tried to point out in m\- 

 paper. Utihzation of this general property would probably lead to a better under- 

 standing of the formation of new associative links as giving their functional role to 

 the plastic effects which are the consequence of repetitive bombardments. In that 

 respect, I do not believe that closed-loop circulations of impulses can be the onh', 

 not even the main cause of prolonged firings. Whatever the case, it is difficult to 

 believe that the process of reverberation, when it operates through multiple, 

 intricate, micro-loops and zig-zag recurrent pathways as are present in neuronal 

 networks, can have another net result than that ot deteriorating the informative 

 content of sensory messages, in one word of increasing 'noise' in the communica- 

 tion of signals within the brain. Shall we assume that learning mechanisms might 

 depend on such a process? 



Another essential mechanism to which sufficient credit has not been given in 

 models is that o( inhibition. 



In answer to Dr Segundo's remark, one can say that it would not be difficult to 

 introduce inhibitory functions into the models just alluded to. Habituation, extinc- 

 tion, differentiation, inasmuch as they involve inhibitory actions could thus receive 

 a good svmbolic representation. May I mention in addition that plastic changes like 

 post-tetanic potentiation apply to inhibitor)- synapses as well as to excitatory ones. 



Finally, I must confess that I have also been puzzled by the question raised by 

 Dr Doty and which he calls the 'temporal paradox of conditioning'. 



Could it not be that the conditional message being weak and ineffectual meanders 

 through neuronal networks, thereby losing time, so that finally it reaches the 

 associative synaptic fields at the moment when they receive the unconditional 

 signal? 



Hebb. The difficulty with these schemes is that they arc quite unrealistic, because 

 they depend on a gross selection of data. There are so many facts of behaviour, so 

 much psychological research that they do not comprehend, that they cannot be 

 taken seriously — at least until they show how the)' are going to handle some of the 

 main facts which have accumulated in the criticism ot exactly such theories as 

 these. Here is a hastily compiled list ot the sort ot thing they need to take into 

 account : 



There is the whole effect of early experience, as modifying the adult capacity to 

 associate stimulus with response. I don't need to go into the question of whether 

 Riesen's chimpanzees reared in darkness were able to see, organically; the chim- 

 panzee reared in unpatterned light was equally incapable in his visual learning. 



There is the factor ot attention. I may remind you again of Lashley's experiment 

 in which there was a failure of learning to discriminate patterns in 200 trials — 



