MULTI-SCALE COMPUTATIONAL MODELING OF ASPECTS OF DEPRESSION: A REVIEW

A complete understanding of the pathophysiology of depression is still a major challenge for researchers. Such an understanding requires complex multi-scale models, which reflect not only the externally observed behaviour of subjects, but explain such behaviour in termsS62123of low level neuronal activity. With current knowledge and computational capabilities, such models are extremely hard to create. The required complexity is due to many factors, such as diversified symptomatology, complex etiology, involvement of nearly all brain areas and the varied contributions of a range of neurotransmitters and hormones. An alternative approach is to model various behaviours independently and subsequently integrate them into a single multiscale,multi-behavioural model.This review identifies some of the physiological systems (hypothalamic pituitary adrenal (HPA) axis, circadian system, sleep wake network, melatonergic system) that are involved in the pathogenesis of selected behaviours clinically observed in depression (sleep disorder,diurnal changes in mood and anhedonia). It considers the current state of the art in relevant computational models, with a focus on lower level representations. The latter emphasize the importance of electrophysiological Hodgkin Huxley (HH)-based models, which emulate limited details of neuronal behaviour. Computational models of populations of networks of such modified HH neurons, operating under the influence of neurotransmitters, hormones and intracellular pathways, may be useful in providing better insight into the disturbed physiology; the current status in development of such neuron models and networks is discussed.Finally the review discusses some hypotheses based on the neurobiological aspects of specific behaviors, and addresses the issues and challenges in developing an integrated multi-scale model of depression.