DFG Research Unit 5389
Contextual influences on dynamic belief updating in volatile environments: Basic mechanisms and clinical implications
Being able to extract regularities from noisy input, recognize when they change and update beliefs accordingly is crucial to mental health. Optimal learning in volatile environments requires dynamic up- and down-regulation of the learning rate depending both on the stochasticity within a state and on the probability of a state change having occurred. If the updating process goes awry this will result in biased internal representations of the state of the world that eventually give rise to maladaptive behaviour. In mental disorders, biased internal representations become apparent when patients either hold on to beliefs about themselves, others or the world that no longer match the observable experiences or fail to form sufficiently stable representations of their environment. However, the assumption that psychopathology could arise from difficulties in dynamic belief updating (DynBU) has not been systematically tested. Also, the neurocognitive mechanisms underlying DynBU are only just beginning to be unravelled. We assume that fundamental changes in the environmental state trigger a cascade of higher-level surprise, boosts of arousal and a cortical network reset. Identifying aberrancies of this cascade that affect the learning rate will inform our understanding of the origins of aberrant belief updating. We also assume that the ability to learn from new outcome contingencies will vary throughout ontogeny and in response to environmental input. Being able to pinpoint clinically relevant contextual influences on DynBU will provide a better understanding of basic adaptive processes and will help to understand the development of emerging psychopathology.
In nine innovative projects, this Research Unit aims to 1) unravel underlying neurocognitive mechanisms of DynBU in volatile environments, 2) identify clinically relevant developmental and environmental contextual influences on DynBU, and 3) use this knowledge to specify problems in DynBU related to psychopathology. To enable direct cross-project comparisons, every project will use a common change-point task, employ a shared clinical assessment and take a behavioural, computational, and neuroscientific approach. Data will be integrated by advanced computational modelling via shared data-analysis. To realize the ambitious aims, the interdisciplinary and internationally oriented Research Unit includes thirteen established researchers and a Mercator fellow with the required expertise in behavioural modelling, development, learning mechanisms, neural processes, and psychopathology. The fundamental insight into mechanisms that underlie DynBU and the developmental and environmental contexts that shape it, will significantly advance our understanding of how people learn and adapt. This will enable us to identify aberrancies that give rise to psychopathology and it can open up new avenues for mechanism-based and developmentally informed intervention built on a new understanding of uncertainty processing.