664 BRAIN MECHANISMS AND LEARNING 



An example of the phenomenon of the first type may be the well-known EEG 

 'conditioned reflex', when using in a constant succession the 'sound-light' stnnuli. 



An example of the second type can be the generalized desynchronization of the 

 electric activity of the cortex which is linked to a stimulation of all the vegetative 

 organs and which occurs in the case of a biological reinforcement, i.e. which ends 

 by an emotional discharge. 



Taking all this into account, one should, in my view, first consider by which 

 mechanisms emotional discharges do contribute to the selective consolidation of 

 fully determined synaptic connections, corresponding to the succession precisely 

 of those stimuli which led to this emotional discharge. 



I believe that the solution to the problem of the conditioned reflex requires first 

 of all the characteristics of the biological condition for the closing of the circuit 

 itself and only this aspect will enable us to succeed. 



Without it any synaptic combination of hypotheses for explaining the condi- 

 tioned reflex cannot lead us to a solution of this difficult problem. 



Gerard. Dr Eccles has taken up the several points raised, and my answers are 

 much the same as his. Certainly time sequences are involved in behaviour, but it 

 does not really work backward in conditioning, and the modifications in neurones 

 and synapses are quite explicitly able to account for conditioned responses. 



Turning from Dr Doty's question to Dr Segundo's, the issue is now a reversal 

 in reinforcement rather than in time. Habituation could be easil)- understood in 

 two ways, facilitation of inhibitory pathways with a feedback action, or effective 

 channelization of impulses into appropriate outputs, with the corresponding 

 diminution of irradiation to non-efiectivc units. This parallels the recent evidence 

 regarding learning; widespread cortical activity occurs while a correct response is 

 being learned, this mostly disappears as highly channelized activity develops with a 

 learned response, but reappears when tlie problem is changed or an error is made 

 in the response. 



Dr Hebb's many points, far from posing difficulties for the model, are among the 

 very things that it effectively does explain. The Broadbent phenomenon and model 

 fit excellently with that of Beurle ; the Lashlcy experiment fits with the feedback 

 mechanisms indicated and so for the other specific points raised. As I see it, the 

 neurophysiological mechanisms now available for interpreting behaviour can 

 account, in principle, for everything except the basic value hierarchy of the indivi- 

 dual, wliich things matter to it; and this I suspect is largely the residue of earlier 

 experience, with its reinforcements and punishments. I can only refer l^r Hebb to 

 the chapter already mentioned for a fuller discussion of these problems. 



