Аннотация:Brain-computer interface (BCI) enables an individual to control devices directly using their EEG. The control in BCI is based on the EEG-detection of attentional focuses to externally presented stimuli. Nevertheless the specific EEG-response to the stimulus can be obtained even without conscious attention involvement. We are interested in the development of the BCI-based system for the detection of covert emotional focuses of attention. This system can be used in psychological and psychiatric research for diagnosis and monitoring of different diseases (autism, schizophrenia), as well as in education and marketing research. In this study we used photographic images of human faces or various scenes as stimuli in the visual oddball paradigm. 14 healthy adults were asked to view faces (mode 1) and pictures (mode 2) passively or to concentrate on the target stimuli (mode 3). Then participants were asked to estimate the stimuli according to the scale “does not generate emotion (0)/generates emotion (100)”. The average classification accuracy for the most emotional stimuli in a trial was 33.6% (faces) and 42.9% (pictures) reaching 80.0% by some individuals while the random level was 16.7%. The average classification accuracy in mode 3 which reflects the classical BCI system was 72.1%. Our results provide evidence that further work on experimental conditions will lead to the development of a system for detection of covert emotional focuses of attention that will able to recognize covert intentions and psycho-emotional states.