A Computer Model of Disease Patterns in Affective Disorders

A Computer Model of Disease Patterns in Affective Disorders 

To simulate the disease episodes and subthreshold oscillations we again refer to our neuronal modeling approaches (Fig. 7.2b). The algorithms have been implemented with a simple but physiologically plausible approach, i.e. with two nonlinear feedback loops, one positive and one negative. Depending on the parameter setting, such a system can attain stable dynamics but also can develop oscillations. 

In our case, the episode generator is comparably stable, far away from spontaneous episode generation. The subthreshold mechanisms are normally also in a steady state but can be tuned with a disease variable into oscillatory limit cycles with increasing frequency and amplitude. Anyhow, the structure of the equations is the same for the two episodes and the two subthreshold variables and the similarities to the neuronal models immediately become evident from the equations. For a more detailed description see Huber s work [2-4, 7, 25]. 

S is the disease variable and gw accounts for Gaussian white noise. 

The strength of the feedback and also whether it is positive or negative depends on the reference value x, or, more precisely, on its distance to the actual disease state x. Scaling parameters are given by the values w with corresponding activation variables 

---