The article deals with the question, whether nonlinear analysis of the sleep EEG may give clues to understanding of disturbed information processing in psychiatric diseases. We point out how the nonlinear approach to EEG dynamics is motivated and review recent investigations concerning nonlinear dynamical properties of physiological systems. We present calculations of the correlation dimension D2 and the principal Lyapunov-exponent lambda 1 for sleep EEG data from 9 depressive and 11 schizophrenic inpatients compared to healthy controls. Combining the findings for D2 and lambda 1 our results point to altered nonlinear brain dynamics mainly during slow wave sleep in depression and during REM sleep in schizophrenia. Finally, possible interpretations of altered nonlinear EEG dynamics during sleep concerning disturbed information processing in depression and schizophrenia are discussed.