Classification of attempted hand movements from EEG signals using the Hadamard basis matching movement-related cortical potential

Brain-computer interface (BCI) is a communication link between the human brain and a device which transforms intended changes in electroencephalography (EEG) signals to control signals. The attempted movement is a technique used to generate changes in EEG signals. In this study, the movementrelated cortical potentials resulted from the attempted movement activities were categorised by utilizing the delta band extracted via Hadamard basis. The projections of EEG signals from the four pre-selected (C3,C4, F3 and F4) channels on these basis were assigned as features. The proposed method was applied to data set file 001–2019 from BNCI Horizon 2020, a Fisher linear discriminant and multiclass support vector machine classifiers were employed to categorise the five classes and achieved a classification accuracy of 28.5 and 30.1 respectively which is higher than the chance level of 20%.