Studying these processes is crucial for several reasons: they are connected to pivotal cognitive processes such as memory consolidation; maladaptive thought processes significantly contribute to various psychiatric disorders; and most importantly for my research, they could hold the key to training multipurpose RNNs. It is noteworthy that catastrophic forgetting is currently a highly researched topic in artificial intelligence, representing for instance, one of the aspects of the Third Wave of AI in Darmstadt. Given these considerations, exploring the formation, progression, and potential control of thought trajectories is essential for understanding the human brain and behavior, promising substantial advancements in disciplines including artificial intelligence, psychology, and mental disorders study.

In our previous research, we developed a data-driven mathematical framework based on network control theory, conceptualizing spontaneous thoughts as a networked linear time-invariant dynamical system, with nodes representing memory properties. We validated this model using data from 82 healthy participants, revealing that the model accurately represented the true distribution of thought probes and that memories called into consciousness exhibited a temporal pattern governed by the networked dynamical system's energy landscape.