Evaluation von Gruppenunterschieden bei fMRI-Daten an einem Beispiel aus der psychiatrischen Forschung

Schizophrenia can be characterized by an impairment in emotional processing. Whereas emotional experience of patients with schizophrenia might function normally, the ability to perceive emotional stimuli is disturbed. The accomplishment to perceive and process emotional facial expressions is probably one of the most important components of social interaction that is impaired in patients with schizophrenia. Therefore the present study dealt with the topic, if differences between healthy controls and patients with schizophrenia in the processing of facial expressions would also become manifest in different brain activity during the task. Seventeen men and women with schizophrenia have been investigated by means of functional magnetic resonance imaging, whilst performing an event-related emotion discrimination task. The stimuli consisted of happy, sad, angry, fearful and neutral facial expressions. The emotion-specific activations have been compared to those of 17 healthy control subjects, that performed the same task. An impairment of the patients' specificity could be observed at the behavioral level, that has been reported for a very similar task before. The healthy volunteers exhibited higher levels of activation than the patients in several areas that have been linked to processing of facial expressions. These included e.g. right fusiform gyrus and right superior temporal gyrus. The former is associated with the processing of invariant aspects of facial expressions, whereas the latter has been linked to dynamic aspects of such stimuli. Although the amygdala plays a key role in any kind of emotion, the difference between the groups was smaller than expected. Furthermore the activation of the amygdala within groups fluctuated around zero in most conditions. In the cingulate cortex a larger activation was observed for the control group as compared to the patients. This region has also been linked to motional processing, particularly when cognitive processes were involved as well. The ventrolateral prefrontal cortex also differed between groups. Here it was observed that the activity of the patients assimilated to that of the controls, when threatening faces (angry and fearful) were presented. The basal ganglia were more active in the patients than in controls, which might come along with the overactive dopaminergic system in schizophrenia patients. Taken together the presented results point to a deficient processing of facial expressions on a neuronal level in patients with schizophrenia. These operations seemed to be impaired at very early stages of information processing and might be related to the problems these patients demonstrate at emotional and social levels. Another topic that has been addressed in relation to this dataset was the multiple comparison problem, which leads to inferential challenges, particularly in functional imaging. Differences in activation patterns were reported between four prevalent statistical thresholds.The problem of defining a suitable threshold is not resolvable in an objective manner without prior knowledge. This challenge has been discussed in the framework of functional imaging and in relation to the data at hand