P300 signal is an endogenous event related potential component. It is mostly elicited from the frontal to parietal brain lobes. Electroencephalography is used for acquiring P300 signal from scalp. P300 signal is used for brain-computer interface systems. P300 based brain-computer interface systems are preferable since they have high overall performance. The most significant overall performance indicator is information transfer rate for P300 based brain-computer interface systems. P300 signal detection accuracy and P300 detection time are using for information transfer rate calculation. Hence, P300 signal classification accuracy is important for getting higher information transfer rate. In this study, it is aimed to investigate P300 detection model for higher classification accuracy. Thus, it is proposed 3-dimensional input convolutional neural network model for P300 detection. Moreover, the proposed model was applied with region based P300 speller which constituted audio and visual stimuli. In experiments, the participants were asked to spell desired words in two sessions which were offline and online session. Linear support vector machine, stepwise linear discriminant analysis, 2-dimensional input convolutional neural network, and the proposed method were compared in both online and offline sessions. It is reached highest average classification accuracy rate with the proposed method in both sessions. According to the online session result, average classification accuracy was 94.22% in 3-dimensional input convolutional neural network model. Furthermore, average information transfer rate was 5.53 bit/min in 3-dimensional input convolutional neural network model. We have also applied methods on BCI competition III-dataset II for 2 participants ''A'' and ''B'' for evaluating performance of algorithms. The proposed method had higher classification accuracy rate than linear support vector machine, stepwise linear discriminant analysis, 2-dimensional input convolutional neural network, and multi-classifier convolutional neural network which was used in other study on same dataset. INDEX TERMS Brain computer interface, deep Learning, human machine systems, P300 detection.

Electroencephalography-based brain computer interface systems could provide alternative communication methods for severely disabled people who cannot use their neuromuscular systems. The P300 signal is one of the event related potentials that are used for brain computer interface systems. The most important performance parameter of a P300 based brain computer interface system is information transfer rate that is calculated by using classification accuracy and P300 signal detection time. Moreover, P300 speller has a very critical role for classification accuracy and information transfer rate in a P300 based brain computer interface. Although most of studies are about row column based P300 speller in literature, region based P300 speller proved that has higher classification accuracy than row column based one. There are very few studies about region based P300 speller. This study aims to investigate methods for obtaining higher classification accuracy and information transfer rate with using region based P300 speller that constituted audio and visual stimulus. This is the first research that using audio and visual stimulus for a region based P300 speller in literature. Previous studies about region based P300 spellers focused on spellers with only visual stimulus types. Our new paradigm presents region based P300 spellers with only audio, only visual, and audiovisual stimuli. Audiovisual P300 speller structure is the newest model for region based spellers. The subject focused on the desired character stimulus. We used the stepwise linear discriminant analysis method for classification that either included the desired P300 signal or not. According to stepwise linear discriminant analysis, the mean classification accuracy value of the experiment was 90.31% with the audiovisual region based P300 speller. With this new paradigm, classification accuracy in the audiovisual P300 speller was improved 15.69% and 66,99% according to the visual only and audio only P300 speller that we used in the experiments, respectively.

No abstract

No abstract

Steady-state visual evoked potential, type of electroencephalography (EEG) signal, that is used for brain-computer interface systems are considered in this Letter. Steady-state visual evoked potential stimulator is needed for realising the signal on the scalp. Besides, information transfer rate is the most significant parameter to evaluate overall performance of a brain-computer interface. EEG signal classification methods, task completion time, and signal stimulator structure affect information transfer rate values. In this Letter, the authors aimed to reach a high information transfer rate by using the proposed signal stimulator and classification method that has new architectures. Eight flickering objects that provide 36 different characters to spell were used. This stimuli optimisation prevented the effect of eye fatigue on signal. Therefore, steady-state visual evoked potential was elicited dominantly. Moreover, 1D convolutional neural network for signal classification was proposed in this Letter. Online experimental data was also classified with canonical correlation analysis that is most commonly used in brain-computer interface systems. The authors compared results according to both of the classification methods. They have reached average value of information transfer rate as 50.67 bit/min with the proposed classification method. This result is significantly higher than similar studies.