Browsing by Author "Tammilehto, Olli"

Threat processing prepares animals to act in the presence of threat, with adequate responses depending on the physical and psychological distance to the threat. Mounting evidence from human and animal studies indicate that this processing occurs in phylogenetically conserved large scale networks that include amygdala with rodent studies pointing out the particular importance of its lateral, basal, and central subnuclei in fear conditioning and extinction. Previous fMRI studies have shown that human amygdala activates both in threatening situations and situations involving members perceived as belonging to an outgroup, suggesting a link between threat and prejudice. Not every outgroup receives the same prejudice: some are deemed more threatening than others. Amygdala studies taking multiple outgroups into account have been scarce. This thesis amends the situation by investigating the relationship between outgroup types, derived from stereotype content model, and amygdala responses by using machine learning to classify fMRI responses from virtual intergroup contact task that included threatening elements and events. The results showed that the classifiers were able to distinguish outgroups above chance only in few meaningful events using data from specific regions. Classifying different task events provided evidence that amygdala is sensitive to modulations of interpersonal distance. Despite being above chance, the classifiers performed modestly in both cases. The absolute differences to the chance level were only marginal and classifiers tended to confuse some of the categories. Basal subnucleus of amygdala and the situation involving an outgroup member starting to approach the perceiver were found particularly important for successful classification. Using MVPA methods to analyse amygdala subnuclei show promise but are possibly limited by fMRI resolution.