Stahl671 bmp

THE ROLE OF MODELS IN THEORETICAL BIOLOGY 195

in the context of biological control are reviewed by Iberall and Cardon (1964), who emphasize the differences between continuous and dis-continuous or discrete controller strategies, linear and nonlinear govem-ing eąuations, etc.

The most generał kind of invariant property for control systems of this class is what can be called a “performance score.” A model and prototype are considered similar when they yield the same score in a certain well-defined test situation. They need not in any way resemble each other in design or appearance. One can envisage also a model of the respiratory control system which functions like the latter over a certain rangę of pulse codę freąuency-modulated inputs and even has a “wiring diagram” revealing certain topological homologies with the actual nerve center. Real neuronal action however, is evidently based on molecular mechanisms, whose naturę is not presently well understood (see Stark et al., 1965), so that alleged correspondence of “wiring diagrams” may be morę apparent than real. In any event, existing Computer models of “bug brains” have connection charts that are remotely suggestive of their natural prototypes.

A. Cardiovascular Control Models

A number of studies deal with regulation of cardiovascular function. Warner (1959) and Warner and Cox (1962) formulated complex analog models of cardiovascular control by the autonomie nervous systems. These models produce rather lifelike results when subjected to simul-taneous inputs from a /ariety of receptora, such as the carotid sinus and aortic blood pressure transducers. A model of blood pressure control based on servomechanism theory is presented by Scher (1962), who studied blood pressure level after presentation of step and periodic stimuli at the carotid sinus receptor. Certain generał deductions about the natural system, e.g., damping properties, are offered on the basis of model performance. A somewhat morę complex analysis of blood pressure regulation is provided in a report by Lercari and co-workers (1964). These authora use generał linear system analysis to analyze gain and time constants in various portions of the natural regulator system. It is pointed out that the latter, in generał, involves a number of redundant nerve circuits and complex summation of many inputs. Localized vas-cular autoregulatory mechanisms, constituting part of the whole cardio-yascular regulatory machinery, have been analyzed in a model study by Koch (1964).