Psychiatric conditions such as depression and anxiety have far-reaching and devastating effects worldwide. Recent decades have seen a rapid advance in mental health science, in our understanding of the underlying neuroscience and cognitive science, and in the development of novel treatment approaches. Nevertheless, the situation remains challenging. While many treatments have proven effective, their effectiveness remains limited. To enhance the effectiveness of these interventions and develop new therapies, a deeper understanding of the mechanisms of action – why and how they work – is necessary. This thesis describes work aiming directly at this, with the hope that investigating the changes that arise from efficacious interventions can aid in improving treatment outcomes.
The thesis focuses on emotional and cognitive change processes in the treatment of de- pression and anxiety. It introduces a novel approach, which models the interplay between emotion self-reports, to understand the dynamics underlying mood maintenance and how they relate to a person’s psychological state and changes therein. We found stability and controllability features that distinguish patients with depression from healthy controls and are associated with self-reported depression scores. We applied control theory to under- stand how the environment can influence depressed and healthy states, by investigating the characteristics of externally elicited transitions between states. In a second study using the same modelling approach, we employed a new experimental design featuring a series of intense visual emotional stimuli paired with a randomized distancing intervention. This showed that a core psychotherapeutic intervention (distancing) influences both the intrinsic emotional processes and the responsiveness to external emotional inputs. Finally, along a third strand of work, we investigated the link between the effects of the SSRI ser- traline on reinforcement learning mechanisms and symptom improvement in a multi-site placebo-controlled randomized clinical trial acquiring the well-established Go/Nogo task. Our results suggest that both sertraline and anxiety are related to processing aversive re- inforcement. However, limitations in behavioral data quality highlighted the urgent need to develop robust tasks for clinical settings.
In summary, this thesis uses dynamical systems to gain an understanding of the impact of depression on mood dynamics, and how they are altered through treatments such as psychotherapy. Furthermore, it suggests that antidepressants affect aversive reinforcement learning in realistic clinical settings, but also emphasizes the importance of a renewed focus on measurement properties for translation of basic science insights to clinical applications.