[Neural network modeling of triiodothyronine concentrations by the levels of lymphocytic NADP-dependent dehydrogenases in patients with thyroid hypo- and hyperfunction].

Whether a computer model which calculates serum triiodothyronine concentrations (T3) can be constructed by the levels of metabolic enzymes in the immunocompetent cells by means of a neural network predictor of real numbers. For this, the concentration of serum T3 and the activity lymphocytic NADP-dependent dehydrogenases were determined by radioimmune assay and bioluminescence technique, respectively. It was found that the neural network prediction might be only in each examined group alone. Conceivably, this is associated with the presence of compensatory processes of intracellular metabolism at high and low concentrations of thyroid hormones. The involvement of the examined enzymes in the metabolic mechanisms of immuno-endocrine interaction both in health and in disease, as well as in the thyroid hypo- and hyperfunction can be judged from the value of the input parameters of the neural network model.