A new psychophysiological method called ''cognitive pupillary oscillation hypothesis'' is introduced. Under the influence of a cognitive test (simple calculations) changes of the maximal amplitudes of pupillary oscillations, reflecting central nervous activation, are observed. In healthy subjects a retest reliability of 0.75-0.85 was calculated. Furthermore, the data after FFT (Fast Fourier Transform) of pupillary oscillations of two populations of patients (n = 509 and 396, respectively) showed a high stability. The validity of the method was determined by means of a procedure similar to cross validation. The use of the ''cognitive pupillary oscillation hypothesis'' for the differential diagnosis between patients with neurotic disorders (ICD 9: 300) (n = 73) and patients with organic brain syndrome (ICD 9: 290, ICD 291.2) (n = 34) is demonstrated.