Narrow Results

Achieving recognizable visual event-related potentials plays an important role in improving the success rate in telepresence control of a humanoid robot via N200 or P300 potentials. The aim of this research is to intensively investigate ways to induce N200 potentials with obvious features by flashing robot images (images with meaningful information) and by flashing pictures containing only solid color squares (pictures with incomprehensible information). Comparative studies have shown that robot images evoke N200 potentials with recognizable negative peaks at approximately 260ms in the frontal and central areas. The negative peak amplitudes increase, on average, from $1.2\mu$V, induced by flashing the squares, to $6.7\mu$V, induced by flashing the robot images. The data analyses support that the N200 potentials induced by the robot image stimuli exhibit recognizable features. Compared with the square stimuli, the robot image stimuli increase the average accuracy rate by 9.92%, from 83.33% to 93.25%, and the average information transfer rate by 24.56bits/min, from 72.18bits/min to 96.74 bits/min, in a single repetition. This finding implies that the robot images might provide the subjects with more information to understand the visual stimuli meanings and help them more effectively concentrate on their mental activities.

Increasing command generation rate of an event-related potential-based brain-robot system is challenging, because of limited information transfer rate of a brain-computer interface system. To improve the rate, we propose a dual stimuli approach that is flashing a robot image and is scanning another robot image simultaneously. Two kinds of event-related potentials, N200 and P300 potentials, evoked in this dual stimuli condition are decoded by a convolutional neural network. Compared with the traditional approaches, this proposed approach significantly improves the online information transfer rate from 23.0 or 17.8 to 39.1 bits/min at an accuracy of 91.7%. These results suggest that combining multiple types of stimuli to evoke distinguishable ERPs might be a promising direction to improve the command generation rate in the brain-computer interface.

Traditional training methods need to collect a large amount of data for every subject to train a subject-specific classifier, which causes subjects fatigue and training burden. This study proposes a novel training method, TrAdaBoost based on cross-validation and an adaptive threshold (CV-T-TAB), to reduce the amount of data required for training by selecting and combining multiple subjects’ classifiers that perform well on a new subject to train a classifier. This method adopts cross-validation to extend the amount of the new subject’s training data and sets an adaptive threshold to select the optimal combination of the classifiers. Twenty-five subjects participated in the N200- and P300-based brain–computer interface. The study compares CV-T-TAB to five traditional training methods by testing them on the training of a support vector machine. The accuracy, information transfer rate, area under the curve, recall and precision are used to evaluate the performances under nine conditions with different amounts of data. CV-T-TAB outperforms the other methods and retains a high accuracy even when the amount of data is reduced to one-third of the original amount. The results imply that CV-T-TAB is effective in improving the performance of a subject-specific classifier with a small amount of data by adopting multiple subjects’ classifiers, which reduces the training cost.