■958 



IIWIiHiKiK (iF PHYSIOLOGY 



NEUROPHYSIOLOGY III 



would have their thresholds lowered and, on sufficient 

 repetition, would constitute an enduring low-thresh- 

 old path the structural engrain demanded l>\ each 

 memory. A new wave of either of the initial types, but 

 no other kind of wave, on reaching this low threshold 

 region could set up the other initial wave type or 

 could re Meet hack on itself— a mechanism for con- 

 ditioning. Even more generally, two waves presenl 

 in a region could "recruit' cells not activated by 

 either but which would then link both a diffuse 

 engrain and a conditioning mechanism (Beurlc, 

 personal communication). Finally, if a last assumption 

 be made ih.it some recurrent fibers from the output 

 of the mass can bring excitation back to the input 

 region, recurrent waxes could travel through the 

 mass and so different outputs be 'explored.' This 

 model thus accounts for reason based on memory and 

 imagery, and can choose the 'good' response internally 

 rather than by behavioral trial and error. 



epitome. For comparison with further neurophysio- 

 logies! data, it will be convenient to epitomize the 

 model as follows, a) Given a random mass of neurons 

 with random connections, waves will arise in it, will 

 decrement or increment as they spread, and, in the 

 latter case, will be able to retravel a given region. 

 b) 11 special sets of cells send excitatory (E) or inhib- 

 itory (I i axons into the mass and if axons from the 

 mass act upon these sets so as to constitute a negative 

 feed-back system, then constant waves can travel 

 through the mass, and different waves can travel in 

 the same mass, using different actual cell patterns, 

 leaving different traces, reaching differenl end cells, 

 and in general retaining individuality, c) Activity 

 lowers excitation threshold; a given wave traversing 

 the cell mass will thus favor the subsequent passage 

 of that wave but of no other. il\ An input favorable 

 to the organism activates I. and therefore facilitates 

 the passage oi the associated wave, an unfavorable 

 one activates I ,m(] mi deflects the wave into new 

 paths and <•< a new output, i > Waves kepi al a con- 



Stant level bv the mtv omeeha iii~.ni of ( h I can reflect 



hum a region of difficult propagation, can cross one 

 .mother, cm leave enduring low threshold tracks 

 where wave fronts have once crossed, and can re- 

 generate themselves or others which have crossed 

 them hum this locus. More diffuse low-threshold cell 

 groups cm similarly form and later regenerate 



paired wav cs / i With a final assumption, of recurrent 

 flections fl the output of SUCh a mass lo the 



input, recurrent u,iu> can navel, the effects of a 



presumptive output can be sampled and problems 



i. iii le solved by reason rather than by behavioral 

 trial and error. 



EXPERIMENTAL SUPPORT. A random population with 

 random connections is a weak postulate; any addi- 

 tional structuring must give richer attributes. Indeed, 

 a powerful approach to the actual table of organi- 

 zation (or sociometrics I) of neurons is offered by the 

 departure of the system from that of a random net 

 (239). A good example of the specific and (he prob- 

 abilistic connections is given bv the spread of mvo- 

 tatic responses, both to well-channeled effector 

 outlets and to others, as a function of distance from 

 the input locus (Lloyd). The tendency of a wave to 

 fade out or to rise to a maximum is reminiscent of the 

 local excitatory process in nerve (156) and has also 

 been urged for cell m,is>es 171) as well as for ex- 

 plosive summation in cord and brain (Bartlev). 



That feed-back modulating mechanisms operate 

 on neural masses is hardly a postulate. Both excitatory 

 and inhibitory inflows to the cortex exist, as from the re- 

 ticular formation and from the cerebellum; and both E 

 and I influences from each of these centers play dow n- 

 stream, and the proportion of E and I output, for 

 example from the cerebellum, in turn is modulated 

 by the rate of input to this high-frequency syn- 

 chronously-discharging organ (Brookhart). Volleys 

 to the cortex via the nonspecific system sensitize the 

 deep cells to vollcvs via the specific one, and they 

 help control the timing and gating of cortical neuron 

 discharges, including spread within the cortex 

 (Jasper). These actions are associated with dendritic 

 potentials and presumably influence soma threshold 

 in the expected manner (Jasper, Chang). With 

 learning, electrical responses appear in deep Structures 

 (especially the hypothalamus) before the cortex, and 

 stimulation of these (especially the centromedian 

 nucleus) can speed conditioning manyfold (20) 

 Certainly waves of potential change of constant 



amplitude can sweep across (he mammalian cortex 

 (167, 1 79) and die amphibian forebrain ( 1 ~ 1 



There is also Mime evidence that different waves 



1,111 traverse the same neuron population. Changing 

 the parameters, such as the frequency of stimulation 

 to a given locus [medulla I ,'>. 898), thalamus 

 cerebellum (214)], can lead 10 quite different outputs. 

 Differenl sequences of complex motor acts are 

 elicited from nearby cerebellar regions despite their 

 rich neuronal linkage and joint participation in (he 

 1'ioievs (Brookhart). A given neuron can be made lo 

 fire at a higher or lower frequency bv the same 

 vestibular stimulation, depending on head position 



