ANALOGIES 305 



axon 



nucleus 



neuron 

 Figure 11-4. A Neural Network. 



chemical step of the crossing of the synapse, and the smaller, graded, at- 

 tenuating potential induced at the far side of the synapse. Because other 

 properties such as a slow wave of electrical activity on the neuron itself, vari- 

 able spike amplitude, varying wave form and overshoot of the spike, and 

 shifting baseline potentials are ignored, the simulations are still approxi- 

 mate. Replicated by such simulation have been: (a) intensity of electrical 

 activity as a function of time; (b) burst firing; (c) repetitive firing; (d) ac- 

 commodation, and change in excitability. Further, the simulated circuits 

 have disclosed certain conditions under which the firing frequency of the net- 

 work will shift. This is a clue from the machine about a phenomenon which 

 has not yet been observed experimentally by neurophysiologists. Thus 

 workers in the field hopefully look forward to advances in man's understand- 

 ing of his brain through its simulation by machines. The reader is en- 

 couraged to study the papers by Bullock, 18 and of Harmon, 19 and to treat 

 himself to the optimism of Reiss, 17 and the careful analysis of Farley, 17 

 thereby to prepare for himself a proper perspective of this exciting new 

 aspect of biophysics. 



We turn now to an outline of the principles upon which are based the 

 two great classes of computers, digital and analog. 



ANALOGIES 



The Digital Nature of Nervous Propagation 



The electrochemical burst arising at the point of stimulation and moving 

 rapidly along the nerve, and called the impulse, was discussed in Chap- 

 ter 10. To a first approximation, the nerve is either stimulated into action 

 or it is not. This is the "all-or-none" property. The stimulation must be 



