Background: Nonlinear dynamics are currently proposed to explain the course of recurrent affective disorders. Such a nonlinear disease model predicts complex interactions with stochastic influences, in particular, because both disease dynamics and stochastic influences, such as psychosocial stressors, will vary during the course of the disease. We approach this problem by investigating general effects of noise intensity on different disease states of a nonlinear model for recurrent affective disorders. Methods: A recently developed neurodynamic model is studied numerically. Results: Noise can cause unstructured randomness or can maximize periodic order. The frequency of episode occurrence can increase with noise but it can also remain unaffected or even can decrease. The observed effects, thereby, depend critically on both the noise intensity and the internal nonlinear dynamics of the disease model. Conclusions: Our findings indicate that altered stochastic influences can significantly affect the outcome of a dynamic disease. To evaluate the effects of noise, it is essential to know about the underlying dynamics of respective disease states. Therefore, characterization of low-dimensional dynamics might become valuable for disease prediction and control.