Principles of Stimulus Coding 3 



brane parts could be easily detected. In fact, the ratio of trans- 

 membrane/axoplasmic resistance decreases exponentially with 

 increasing distance along an axon from a source of potential, and 

 the preferential pathway for current is thus across the near mem- 

 brane, contributing to the progressive decline of longitudinal cur- 

 rent flow from the source point (fig. 2). Indeed, with the exception 

 of the largest fibers found in invertebrate preparations (up to i mm. 

 in diameter) the complete depolarization of an axon at one region 

 would be undetectable by conventional amplification techniques 

 only one centimeter away. Since many axons are longer than one 

 centimeter — up to a meter or more in vertebrate preparations — 

 variations in the resting potential of the neuronal membrane 

 cannot by themselves be used for analog replication of stimulus 

 intensity, and alternative mechanisms of information transfer have 

 necessarily evolved. A series of proportional voltage amplifiers 

 might foreseeably have been developed by organisms as an 

 alternative to propagated all-or-none type of transmission. As 

 Rushton'^ has pointed out, however, the number of separate 

 amplification points logically required in, for example, the length 

 of the nervous pathway involved in the transmission of informa- 

 tion from one's finger to the spinal cord, would tend to introduce 

 serious distortions of the amplitude of the original signal unless 

 the accuracy of each stage were unrealistically large. It is the 

 structural uniformity required of such a system that is so difficult 

 to visualize in practical terms. The spread of potential along an 

 axon, for example, depends to a great extent on fiber diameter. 

 Any variation in this parameter would necessarily have to be 

 precisely matched by compensatory alterations in the characteris- 

 tics of the segmental amplifiers if distortion-free signals were to be 

 transmitted. Now, although pulse-coded systems are theoretically 

 as prone to distortion as any other, in practice time-dependent 

 responding systems can be utilized with a great deal of precision. 

 If the nerve impulse and its recovery processes were determined 

 by the rate constants of one or more self-limiting reactions within 

 the membrane, identical electrical transients would appear at each 

 membrane region during the passage of an action potential down 

 a nerve; only during very high frequencies might the amplitude 

 and time course of an impulse undergo distortion or change, 

 possibly due to a too-rapid utilization of components involved in 



