Operators and scales: diagnostic and rating issues in psychiatric PET research

In psychiatric research that for various reasons has to restrict itself to a limited number of subjects, such as studies involving expensive positron emission tomography techniques, issues concerning the parsimonious description of patients gain in importance. The number of descriptive variables must be optimally small. This paper offers a conceptual back-ground for the choice of operators in operational diagnostic systems designed to delimit pathological types, and of rating scales designed to measure syndromal severity in a dimensional way. A practical suggestion in five tenets for the organization of clinical research of this kind is presented.