Czekalla, Nora2026-04-072026-04-072026https://epub.uni-luebeck.de/handle/zhb_hl/3624Self-beliefs are essential to our identity and emotional well-being. They are formed through self-related feedback from the environment. The processing of feedback is influenced by self-related motivations and emotional experiences. This can lead to biased weighting of incoming information when updating beliefs depending on whether it is better or worse than expected. In depression, these biases can perpetuate maladaptive beliefs. This dissertation explores how novel self-beliefs are formed, using neuroscientific and computational methods to model belief formation in response to feedback. Across four studies, biases in self-belief formation were examined in individuals with and without depression, considering individual and contextual factors such as affective experiences, symptom burden, and stress. Participants engaged in a learning task with trial-by-trial performance feedback in relatively unfamiliar domains, designed to elicit the formation of novel self-beliefs. Self-belief updates were best described by prediction error valence, with consistent findings of a negativity bias specific to self-beliefs but absent for beliefs about others. Study 1 investigated audience effects on self-belief formation and links to global prior self-beliefs indicated by self-esteem. It showed that more negative global self-beliefs were associated with more negative self-belief formation. In a public context, belief formation was more negative in those with higher subclinical social anxiety. Study 2 combined functional magnetic resonance imaging (fMRI), pupillometry, and emotion ratings to demonstrate that more negative self-belief formation was associated with the experience of more embarrassment and less pride, as well as more arousal, as indicated by pupil dilation. Neurally, it was linked to a heightened activity to more negative prediction errors in the anterior insula, amygdala, midbrain regions, and medial prefrontal cortex. Study 3 examined individuals with depression and healthy controls using fMRI. Individuals with depression showed stronger reactivity in the insula to negative compared to positive prediction errors. Despite no group differences in behavior, individuals with a higher symptom burden exhibited more biased belief formation, displaying more negativity towards themselves and more positivity towards others. In individuals with depression, in particular, symptom burden was linked to reduced updates following positive prediction errors. Study 4 showed that self-belief formation was less negatively biased following social-evaluative threat, which was associated with better recovery from stress-induced negative affect. In summary, the studies show a negativity bias when forming beliefs about one’s ability. This may be related to motivations present in performance contexts like improvement motivation or motivation to avoid failure with a threat driven focus on negative feedback. The findings also highlight the critical role of individual factors like current affective experience, global prior self-beliefs, and symptom burden, alongside contextual factors like social evaluative stress in self-belief formation. Neurally, the studies identify the insula as central to processing affective, self-related feedback linked to biased updating. Notably, in individuals with depression, reduced self-belief updating following positive prediction errors with increased symptom burden suggests a diminished receptivity to corrective positive feedback, which may perpetuate maladaptive self-beliefs.enself-belief updatingdepressionstressnegativity biascomputational modelingneuroimaging500thesis.doctoralurn:nbn:de:gbv:841-2026040702