Krohne LG, Wang Y, Hinrich JL, et al. - In this fMRI study, researchers used supervised machine learning to categorize individuals according to the extent of social anhedonia they display. More precisely, 70 individuals were given a social cognition task, during which both spatial and temporal network features were extracted and utilized to support vector machines for classification. A notable classification of individuals with elevated levels of social anhedonia when using the times series extracted using multi-subject archetypal analysis, showing that temporal dynamics carry vital information for the classification of social anhedonia was observed. In a task classification of the theory of mind condition, a similar time series gave the greatest classification performance. In conclusion, the spatial network corresponding to that time series involved both prefrontal and temporal-parietal regions as well as insula activity, which formerly had been associated to schizotypy and the development of schizophrenia.