A linear discriminant analysis transformation-based approach to the classification of three different motor imagery types for brain–computer interfaces was considered. The study involved 16 conditionally healthy subjects (12 men, 4 women, mean age of 21.5 years). First, the search for subject-specific discriminative frequencies was conducted in the task of movement-related activity. This procedure was shown to increase the classification accuracy compared to the conditional common spatial pattern (CSP) algorithm, followed by a linear classifier considered as a baseline approach. In addition, an original approach to finding discriminative temporal segments for each motor imagery was tested. This led to a further increase in accuracy under the conditions of using Hjorth parameters and interchannel correlation coefficients as features calculated for the EEG segments. In particular, classification by the latter feature led to the best accuracy of 71.6%, averaged over all subjects (intrasubject classification), and, surprisingly, it also allowed us to obtain a comparable value of intersubject classification accuracy of 68%. Furthermore, scatter plots demonstrated that two out of three pairs of motor imagery were discriminated by the approach presented.

It is quite clear that a person can make mistakes in the process of speech: badly pronounced letters, “missing” syllables. This situation poses certain difficulties for developments in the field of mental or inner speech-based BCI technology. The inability to separate errors from correct speech can significantly reduce the effectiveness of BCIs. In this study, we used words representing the direction in space and pseudowords made randomly from individual syllables, that is, pseudowords were phonetically close to words. The aim of the study was to explore the possibilities of neural network classification of words and pseudo-words-related EEG coherence patterns in the task of mental and spoken speech. It was shown that the preparation and pronunciation of pseudowords takes a long time and is performed non-lexically with sublexical components. It is shown that the structure of EEG patterns with high coherence values registered in both hemispheres in the gamma range is homotopically equivalent and seems to be related to attention mechanisms. The EEG patterns with low coherence are significantly different, which indicates the specifics of right hemisphere involvement in speech processes. The most stable coherence patterns with a significant difference between words and pseudo-words are found in the left hemisphere in gamma frequencies. The main conclusion is that the neural network classification of inner speech-related EEG coherence patterns based on the multilayer perceptron (MLP) demonstrates an accuracy up to 90% in the recognition of words and pseudowords.