Abstract
It has been suggested that neuronal oscillations play a fundamental role for shaping the functional architecture of the working brain. This thesis investigates brain oscillations in rat, human healthy population and major depressive disorder (MDD) patients. A novel measurement termed cross-frequency directionality is developed to assess the directional interaction between slow and fast oscillations. The results suggest that bidirectional communication - the fast oscillations coordinating the slow oscillations, or vice versa - are both possible. Overall, this indicates that the brain coordinates information flow in a complex and flexibie manner. Moreover, disrupted brain oscillations are identified in MDD patients in comparison to healthy controls. Furthermore, information about patients’ brain oscillations information and an advanced machine learning technology are integrated to predict MDD depression severity. The proposed model could yield a quantitative and objective estimation for depression severity, which in turn has the potential to help psychiatrists with more precise diagnosis, earlier detection, better prevention, and treatment effect evaluation on MDD.
