Computational analysis of the response of the monoaminergic neurotransmitter system and stress- and sex-steroid hormone systems to administration of antidepressant drugs

Fewer than half of depressed patients who take antidepressants experience complete remission of symptoms after 4-6 weeks of daily administration. The reason for this clinically observed heterogeneity in antidepressant response is still unclear. The majority of antidepressant drugs today are designed to enhance the brain’s production of one or more of the monoaminergic neurotransmitters (serotonin [5HT], norepinephrine [NE], and dopamine [DA]). In order to better understand how the brain adapts to chronic antidepressant administration, we developed a computational model that represents the known interactions of the neurobiology of depression. This model was expanded with further knowledge extraction, and also re-structured as computational tools were improved. Our model is based on the neuroadaptation hypothesis, whereby the brain homeostatically adapts to chronic antidepressant administration by adjusting the strengths of transmitter-system components (TSCs) in order to return brain-region activity levels back toward their pre-drug baselines. The main finding was that the model can adapt through many different pathways, arriving at many different TSC-strength configurations but not all of the adapted configurations are also associated with therapeutic elevations in the monoamines. These results provide insight into the heterogeneity among individuals in response to chronic antidepressants. We expanded this model to incorporate the stress-hormone response and the male sex-steroid system, and postulated that if only a subset of adapted configurations to chronic antidepressant are therapeutic, then it is possible that individual, pairs, or subsets of TSCs are responsible for mediating the therapeutic state. Through several modes of analyses, including sensitivity, correlation, and linear temporal-logic, we found that therapeutic neuroadaptation to chronic antidepressant is an overdetermined process that depends on multiple TSCs, providing a potential explanation for the clinical finding that no single antidepressant alleviates depressive symptoms in all patients. Our models can be used to systematically facilitate the clinical practice of antidepressant augmentation by providing the means to computationally screen for antidepressant drug/hormone combinations that could potentially be more therapeutic than single drugs by themselves