Objective. There have been no criteria found so far to quantify the wide range of EEG spectral and coherent indicators, which would allow discrimination between schizophrenia disorders and healthy human states. The goal of this research is to find objective EEG-based schizophrenia criteria through a discriminant analysis and to obtain a linear discriminant function (LDF) with sensitivity and specificity of at least 85%. Material and methods. The study covered 83 schizophrenia patients and 116 healthy individuals, with similar major characteristics. EEGs were recorded in EDF format using 16 leads under the 10—20 system and ear reference electrodes, with further spectral and coherent analysis. Results. A discriminant analysis has provided LDFs from which a formula with 5 predictors was extracted. The most valuable diagnostic predictors were beta2 and theta rhythm powers in the F3 lead. In the same manner, interhemispheric theta rhythm coherence in the pair T5—T6 and interhemispheric beta rhythm coherence in the pairs F3—C3 and T3—C3 were significant predictors. All of them being negative, the EEG supplied thereby most likely relates to the schizophrenia class. Conclusions. The novel methods used for selecting EEG features and choosing their combinations to obtain LDFs and verify the outcomes will give the researcher a powerful and flexible tool to perform an EEG discriminant analysis which allows obtaining linear discriminant functions in short terms without routine procedures, performance of preliminary diagnosis using an expert system based on the obtained LDF, evaluation of treatment efficacy by the LDF score dynamics. During treatment and changes of predicting modules function, variables may increase or decrease depending on the direction of predictors that may serve as an indicator of treatment efficacy.