Economic aspects of feature selection method based on the genetic algorithm paradigm adapted to the specific characteristics of Brain–Computer Interfaces This monograph is based on the doctoral dissertation entitled Feature selection method based on the genetic algorithm paradigm adapted to the specific characteristics of Brain–Computer Interfaces, which was submitted and successfully defended in 2024 at the Faculty of Computer Science at the West Pomeranian University of Technology in Szczecin. Developing a new algorithm proved to be a significant research challenge due to the specific nature of brain waves observed in individual users of the brain–computer interface (BCI). Extracting a large number of redundant features becomes necessary as each user exhibits unique brain wave patterns. From this vast set of features, only those exhibiting the highest discriminatory ability for a given user are selected. Consequently, the number of features extracted from the EEG signal during the estimation of the core classifier parameters in the BCI system can reach hundreds or even thousands. However, on the other hand, in order for the interface to operate in real-time-like mode, it is crucial to limit the number of features to be extracted from the signal for each subsequent control instruction generation. Another challenge arises from the high specificity of brain waves, necessitating separate calibration sessions with each user to estimate classifier parameters based on the data collected. This limitation results in relatively small calibration datasets, typically ranging from 100 to 200 records, which is significantly small compared to the thousands of possible features that can be extracted from the EEG signal. Therefore, the selection of a few essential features with the highest discriminatory abilities becomes an extremely important and challenging stage in constructing a motor imagery-based interface. From an economic perspective, the development of brain–computer interfaces (BCIs) constitutes a significant driver of growth in the advanced technology sector. This technology stimulates innovation, fosters the emergence of start-ups, and increases demand for highly skilled professionals in areas such as artificial intelligence, biomedical engineering, and computer science. The steadily growing value of the global BCI market makes it an attractive domain for investment and a source of competitive advantage in the digital economy. In this context, the optimization of feature selection plays a crucial role by reducing computational costs and shortening data processing time, thereby enhancing the efficiency of BCI system implementation. Reducing the number of input features decreases the demand for IT resources, energy, and technical personnel, which translates into lower operational costs and improved scalability. The GAAMmf algorithm developed in this study, based on the genetic algorithm paradigm, enables effective feature set reduction while maintaining high predictive quality, measured by the accuracy of brain activity pattern classification. The proposed approach has applications not only in BCI systems but also across other data-driven and AI-based sectors of the economy, such as finance, logistics, e-commerce, and risk management. In these domains, precise feature selection supports faster decision-making, reduces prediction costs, and improves resource allocation efficiency. As such, feature selection emerges as a tool of strategic economic importance, fostering the development of data analytics, the automation of decision-making processes, and the digital transformation of organizations. The monograph is an extension of a doctoral dissertation aimed at developing a feature selection method based on genetic algorithms, adapted to the specific characteristics of brain–computer interfaces. These characteristics include the need to preserve information about the original feature location and limit the feature space to only a few to several relevant features. The proposed feature selection method enabled the selection of a set of features with higher or comparable discriminative abilities, as measured by the accuracy of brain activity pattern classification, and with a smaller number of features than reference methods.