I 70O HANDBOOK OF PHYSIOLOGY >-~ NEUROPHYSIOLOGY III 



Ant- 

 /tg. 



feedback 



Lockere tSewegungen 



Viosek 



Transient oscillatory 



Sleady.srate 



error 



• ■ Transienf 

 error 



i 



Time 



fig. 8. Upper left: Schema showing the principle of organization of .1 servomechanism unit. Lower 

 left: Distribution of the muscular activity in two antagonistic groups of muscles, the agonists I 

 and antagonists (Ant.), at four stages of increasing speed of movement. The electromyographic ac- 

 tivity of each group is schematized by a line the thickness of which varies with intensity. The me- 

 chanographic record shows the increasing tendency to oscillation when the speed of execution is in- 

 creasing. The final position is achieved only by a transient oscillatory movement. [From Wachholder 

 ( 124. 1 I ppei right: Diagram showing transient stability of a physical system with varying degrees of 

 damping. Curves 2, 3 and ./ are progressively underdamped and show increasing degrees of oscilla- 

 tory behavior. Curve / is overdamped and shows great stability at the expense of a long response 

 time A servomechanism shows a similar mode of functioning. Compare these curves with those 

 recorded by Wachholder. [From Brown & Campbell (19).] Lower right : Response of a human sub- 

 ject in a tracking experiment. The diagram illustrates the several kinds of observed errors between 

 the response (dashed lint > and the command ' \nlul hnr) during a sudden change in the latter. A steady 

 -•tale and transient errors, as well as transienl oscillatory behavior, are also characteristic of the 

 performance "i .1 servomechanism. [From Ruch (103).] 



.it each level of the nervous organization. We now 

 have tt) search mure precise!) for the principle of 

 organization of such ,i regulative action which 

 automatically leads the prearranged motor per- 

 formance i<> its correct achievement. 



mi 1 11 \ms\is 1 1 1 mi p-ADjusTMENT. .V .1 matter ol fact, 

 the receni new developments in control systems and 

 communications provide attractive analogies and 

 promising suggestions foi .1 better understanding of 

 this problem (19, ;•>, ,|, 1 13). For biologists, the 

 most interesting control svstcms are those commonly 

 called 'servomechanisms' or slave systems. Cont- 

 ents .ind characteristics of servomechanisms are 

 numerous and varied, bul their common feature is 



ih. 11 they possess Mime kind of controlling device 

 able to appreciate continuously the discrepancy 

 between die state of die machine realized al .1 given 

 moment and die final aim assigned to it by its con- 

 structor. Through a •feed-hack' circuit, die informa- 

 tion collected from an error-detecting device is al 

 ever) moment sent hack to die servomotor controlling 

 die output. H\ modifying the input command it 

 permits die output 10 be corrected lor the detected 

 discrepancy. Thus the 'behavior' of a sen omechanism 

 is not governed by a blind obediance to the order of 

 a predetermined program of action, but it presents a 

 kind of self-adjustment hv modifying the input 

 command of the svstem as a function of its output. 



Such 'teleological' mechanisms (36) are designed 



