'95 6 



HWDBOOK OF PHYSIOLOGY 



Nil klll'HYSIULOGY III 



EFFECTORS 



fig. 2. Broadbrnt's information- 

 flow diagram for the organism. For 

 description sec the text. [From Broad- 

 bent (32) 



reverberation prior to perception and that of temporal 

 integration at the synaptic decision points.) 



Figure 2, Broad bent's information-flow diagram 

 for the organism, covers research by himself and 

 others on immediate memory, on learning, on 

 anticipation and refractoriness, and on noise, multi- 

 channel listening and prolonged performance. It is 

 obvious that this model, reached exclusively from 

 studies of behavior of intact organisms, mainly man, 

 is subject to easy translation into a neurophysiological 

 model, based on direct manipulation of the nervous 



. In . If tion Wave Model 



An especially impressive model has been offered 

 In Heinle (24) who starts with assumptions based 

 upon reasonable neurophysiological knowledge, de- 

 velops from these a formal model, and demonstrates 

 as mathematical consequences a set of behavioral 



• Beurle writes me see also Beurle (25)] on the lit of his 



model (below i will] Broadbent's as lollows 



'"It is easy to conceive .1 reverberator) store arising out ol .1 

 network in which there is a large random factor, but the 

 problem is that there would be .1 random relationship between 

 the output and the information stored. I Ins is very easily 

 solved it the output from the reverberatory store is fed directly 

 to the input ol my original model since this latter is essentially 

 1 apable ol dealing with, and sorting out, information arriving 

 along .1 set ol randoml) connected fibres. 



III. combination ol short term memorj with the original 

 long t<i in memoi'v model would go l.u towards explaining man) 



'il the well known features ol 01 ) li would explain our 



ability to 'telescope 1 the time scale of p. 1st events so that when 

 we iei .ill .1 train ol memories 11 is only the important highli ' 

 that 1 in 1 ie I >.u k 1 1 would also explain how single, but important 



events 1 fti a be remembered well, .1 fa< 1 that is less eas) to 



account 1 .1 basis ol the original simple long term memory 



nil. del alone 



properties that parallel many facets of complex 

 human behavior. It will be interesting to present 

 first a summary of Beurle's formulation and then to 

 marshall some of the material in this Handbook that 

 fits in with and supports it. 



FORMULATION. Wans in a neuron mass. Assume a 

 population of neurons, the cells distributed randomly 

 in a mass and making random connections with 

 other cells, except thai connections with others 

 decrease .is some, not critical, function of the distance 

 between them. Impulses travel from cell to cell along 

 the connections and a given cell fires when a sufficient 

 number of connections, about 10 assumed, sum their 

 activities. A brief synaptic delay intervenes between 

 effective summation and discharge. After a cell has 

 been fired ('used'), it will remain inactive for a time 

 and, aside from such a "refractory period,' its threshold 

 will remain constant or, on later assumptions, will 

 be lowered cumulatively by previous activity and 

 will be temporarily raised by inhibitory or lowered 

 by excitatory connections. A wave entering such .1 

 cell mass does not spread from cell to cell, as in 

 simple chains, and need mil activate all cells as it 

 travels; rather propagation requires the convergence 

 of many cell influences on each, and the divergence "I 

 each cell's influence on many; and the --pre. id of the 

 wave is nonlinear in relation in excitation level and 

 cell thresholds. 



A wave front of active cells .11 time, I, advances 

 during the time increment, r, bv activating additional 

 cells in front of il. The proportion l/'i of the cells in 

 any cross-section that become active as the wave 

 passes through it will remain constant /•',.. - / 



unlv for a critical initial value ol /•", /•",,. Above this 



level, activity rises to a saturation of active cells; 

 below ihis, the wave decrements to extinction. 



