Misinformation, oder „Fake News“, ist in der Online-Kommunikation weit verbreitet und beeinflusst Diskurs und Zusammenleben. Es ist jedoch wenig darüber bekannt, wie wir auf individueller Ebene beeinflusst werden, wenn wir Meinungen bilden und Urteile ableiten. Diese Dissertation untersucht die kognitiven und Gehirn Mechanismen, die der Verarbeitung sozialer, personenbezogener Misinformation zugrunde liegen. Proband*innen wurden mit negativen, positiven oder relativ neutralen personenbezogenen Nachrichten konfrontiert, die entweder verbal als unglaubwürdig gekennzeichnet waren, z.B. "angeblich", oder aus bekannten Medienquellen stammten, die als glaubwürdig oder unglaubwürdig wahrgenommen wurden. Die Ergebnisse zeigten, dass soziale Urteile stark vom emotionalen Gehalt beeinflusst waren, unabhängig von der Glaubwürdigkeit. Elektrophysiologische Korrelate früher emotionaler und erregungsbezogener Prozesse sowie Korrelate späterer evaluierender Verarbeitung waren verstärkt für Personen, die mit emotionalen Inhalten assoziiert wurden—unabhängig von der Glaubwürdigkeit der Information. Emotionale Inhalte wirken also nicht nur auf die unwillkürliche und früher Reaktion auf Nachrichten, sondern sogar auf Prozesse, für die erwartet wurde, dass sie die Information aufgrund ihrer Glaubwürdigkeit evaluieren würden. Um zu intervenieren, bewerteten die Proband*innen vor der Konfrontation mit Schlagzeilen explizit die Glaubwürdigkeit der Quelle. Dies half teilweise, die Glaubwürdigkeit positiver Nachrichteninhalte zu verarbeiten. Die Einsicht in die fehlende Glaubwürdigkeit hatte jedoch keinen Einfluss auf die Effekte negativer Nachrichteninhalte. Unsere Ergebnisse zeigen, wie der emotionale Gehalt sozialer Misinformation das Gehirn und das Urteilsvermögen beeinflussen kann, selbst wider besseres Wissen über die fehlende Glaubwürdigkeit. Perspektivisch helfen diese Erkenntnisse, uns den Herausforderungen von Misinformation aus Sicht der individuellen Kognition zu stellen. / Misinformation, also called “fake news”, is highly prevalent in online communication affecting public discourse and social coexistence. However, little is known about how we are affected by it on the individual level when we derive opinions and judgments. This dissertation investigates the cognitive and brain mechanisms underlying the processing of social, person-related misinformation. Participants were exposed to negative, positive, or relatively neutral news about other persons that was either verbally marked as untrustworthy by adding e.g., “allegedly”, or stemmed from well-known media sources perceived as trusted or distrusted. We found that social person judgments strongly relied on the emotional content independent of the credibility, showing how social misinformation affects person evaluation although it is perceived as untrustworthy. Electrophysiological indexes of early emotional and arousal-related processes, as well as correlates of later evaluative processing were enhanced for persons associated with emotional contents regardless of the credibility of the information. This shows the pronounced influence of emotional contents not only on the initial and early response to news, but even on processes that were expected to evaluate the information on merit of its credibility. In a first attempt to intervene, participants explicitly evaluated the credibility of the source before reading the headlines. This helped to overcome the bias for positive news and process its credibility to some degree. However, the insight into the lack of credibility had no influence on the effects of negative news on brain responses and social judgments. Our results demonstrate how emotional content in social misinformation can affect mind, brain, and judgments even against better knowledge of its lacking credibility. In perspective, these insights help to face the challenges of misinformation from the perspective of the individual’s cognition.

Objective: Lack of insight into illness is common in patients with schizophrenia diagnoses. It is supposed to reflect deficits of executive functioning that are frequently assessed with the Wisconsin Card Sorting Test. Studies on the remediability of WCST deficits in schizophrenia, however, raise doubts about its construct validity and suggest the use of a dynamic pretest-training-posttest paradigm (WCSTdyn) and single-case analysis (Reliable Change Index, RCI) in studies on insight. Moreover, a multifactorial etiology with neurocognitive and motivational factors, as suggested by Startup's (1996) model, has to be taken into consideration. The model hypothesizes a quadratic function of the relationship between insight and cognition, which means that both cognitively impaired and cognitively intact patients with low insight are to be expected. Method: Three interrelated studies on WCSTdyn and insight were conducted. In study 1, the split-half reliability of the WCST-128 was investigated in a non-psychiatric sample (N = 110). Study 2 compared different RCI single-case tests of significance of intraindividual change on data from N = 400 patients with schizophrenia diagnoses. Furthermore, a typology with three homogenous subgroups was developed and first steps toward an external validation were taken. In study 3, the three types of test-takers were compared on two measures of insight (Item G12 of the Positive and Negative Syndrome Scale [PANSS]; Osnabrueck Scale of Therapeutic Attitudes and Identification of Psychological Problems in Schizophrenia [OSSTI]). The model by STARTUP (1996) was tested by means of regression and cluster analyses including scales on coping (Freiburg Questionnaire of Coping with Illness, FKV: MUTHNY, 1989) and defensiveness (Eppendorf Schizophrenia Inventory, ESI-FR: MAß, 2001; N = 85). Results: The WCST-64 was sufficiently stable (r_tt = .70 [Total Number Correct]). Concordances of different RCI methods were high (kappa = .72 - .90). 45 % of the sample consisted of high scorers, in 43 % a low initial score could be normalized by a short training intervention (learners), and in only 12 % of the sample the WCST deficits were not amenable to training (nonlearners). Insight of nonlearners was significantly reduced (G12: g = 0,45). Finally, three clusters were identified with configurations of insight and WCSTdyn scores consistent with the prediction by STARTUP (1996). Patients with intact neurocognition but low insight responded in a significantly more defensive manner (g = 0,38). They did not differ in self-rated coping, however. A quadratic relationship could not be confirmed. Conclusion: The RCI-based performance typology developed in studies 1 and 2 is a universally applicable analytic tool for future studies on insight-limiting neurocognitive deficits with the WCSTdyn. Results from study 3 point in the direction of a multifactorial etiology of lack of insight in schizophrenia with differential contributions of neurocognitive deficits (e.g., interference control) and defensiveness.