j Abstract Objective EEG investigation in patients with an at risk mental state (ARMS) for psychosis and patients with a first episode of psychosis (FE) in comparison to healthy controls (HC) in a clinical follow up study of Early Detection of Psychosis. Method Seventy-three patients (42 ARMS, 31 FE) and 35 HC were investigated. ARMS patients were followed up in order to monitor transition to psychosis. Psychopathology was assessed with respect to positive and negative symptoms. At study baseline EEG was recorded using the 10/20 system. Two blinded neurologists analyzed the EEGs visually for presence of generalized or focal slowing and epileptiform discharges. EEG data were controlled for medication and substance abuse. For statistical analyses we used v 2 -tests, logistic regression, ANOVA, and receiver operating characteristics. Results Patients showed significantly more pathological EEG abnormalities than HC (P < 0.05), located more frequently in temporal or fronto-temporal regions (P < 0.01) of the brain, with twice as many pathologies in ARMS than in FE patients. The specificity of the prediction of psychosis could be increased from 59 to 73% by considering EEG pathology in addition to psychopathology alone. In contrast, sensitivity of prediction remained unchanged. Conclusions These results show that EEG investigation in patients at risk for psychosis can add to the identification of those patients who will not develop psychosis later on.