Aran Safir

A model-based method for computer-aided medical decision-making

A general method of computer-assisted medical decision-making has been developed based on causal-associational network (CASNET) models of disease. A CASNET model consists of three main components: observations of a patient, pathophysiological states, and disease classifications. As observations are recorded, they are associated with the appropriate states. States are causally related, forming a network that summarizes the mechanisms of disease. Patterns of states in the network are linked to individual disease classifications. Recommendations for broad classes of treatment are triggered by the appropriate diagnostic classes. Strategies of specific treatment selection are guided by the individual pattern of observations and diagnostic conclusions. This approach has been applied in a consultation program for the diagnosis and treatment of the glaucomas.

Glaucoma consultation by computer

Abstract This paper describes a computer-based system for consultation in the diagnosis and therapy of glaucoma. The reasoning procedures interpret the findings of a particular patient in terms of a causal-associational network (CASNET) model that characterizes the pathophysiological mechanisms and clinical course of treated and untreated diseases. The major new features of this program are: (a) generation of complex interpretations from a qualitative model of a disease process; (b) reasoning about detailed follow-up management of a patient; (c) incorporation of alternative expert opinions about subjects under debate; and (d) its testing and updating by a collaborative computer-based network of glaucoma researchers.

Distribution of Cone Orientations as an Explanation of the Stiles–Crawford Effect*,†

The retinal directional effect of several subjects was measured. Flicker brightness matches were made using monochromatic lights. The experiment was performed in two parts. The first measured the Stiles–Crawford effect across only the central 6 mm of the pupil. Objective statistical methods were used to derive the best-fitting parabolas by least-squares criteria. Variances of the data were calculated from the replications and indices of goodness of fit were derived. On these grounds, the parabola was found to be unacceptable. The second part of the experiment included replicate matches made at points across the entire width of the dilated pupil. Similar methods were again used to derive the best-fitting gaussian and parabolic functions. The gaussian proved to be statistically acceptable as a description of the change of brightness with position of pupil entry while the parabola did not. The consequences of accepting the gaussian function are far-reaching. It is argued that the shape of the Stiles–Crawford-effect curve as measured across the pupil is the result of the normal distribution of cone angulations, each cone having inherent directional sensitivity. It follows from this that the directional sensitivity of individual cones cannot be as broad as the observed Stiles–Crawford-effect curve. It is probable that the individual cone will accept energy incident upon it within only a very small angle from its long axis.

The retinal directional effect: A model based on the gaussian distribution of cone orientations ☆

Abstract Experiments are presented in which the retinal directional effect was measured in several subjects. Mathematical analysis of the data leads to the conclusion that the gaussian function, Y = K 2 + Ae − B ( x − c )2 , satisfactorily describes the phenomenon; the parabola and cosine function do not. A three-dimensional gaussian model is given, explaining the characteristics of the retinal directional effect, including its spectral variation and hue-shift.

Concepts of computer-based modeling for consultation in optics and refraction

In this paper we discuss general principles of computer science applied to problems of medical consultation. We describe some of the design considerations for a computer-based consultation program in optics and refraction. Since the visual system is the input pathway for most of the data entering the central nervous system, the process of refraction serves important purposes beyond the generation of a prescription for a pair of eyeglasses. Refraction, the measurement and correction of optical deviations of the eye from the desired norm, evaluates what is often the weakest link in the chain of events constituting the visual process. Our purpose is to present the diversity of clinical objectives and constraints that affect the design of such a clinical consultation system, and to describe how different types of models are needed for different aspects of the consultation.