Abstract:

The aim of the experiments was to investigate differences in the electroencephalographic (EEG) signal when playing a simple computer game using standard control and reversed control. The data collected allowed the neural network to be trained on both trials. Classification of movements in reversed control with the network trained in normal control showed an efficiency close to 50%.The EEG signal was taken via the Unicorn Hybrid Black device. The collected data will allow the creation of two sets for the neural network - a test set and a learning set. The data was normalized for the purposes of the experiment. Deep learning (DL) better categorizes the signal for individual respondents. The classification performed by the network for individual cases performs much better than for the whole. This suggests that in order to find a general solution, adaptive elements should be added during the configuration process.