Abstract
We present a novel approach based on deep learning for decoding sensory information from non-invasively recorded Electroencephalograms (EEG). It can either be used in a passive Brain-Computer Interface (BCI) to predict properties of a visual stimulus the person is viewing, or it can be used to actively control a BCI application. Both scenarios were tested, whereby an average information transfer rate (ITR) of 701 bit/min was achieved for the passive BCI approach with the best subject achieving an online ITR of 1237 bit/min. Further, it allowed the discrimination of 500,000 different visual stimuli based on only 2 seconds of EEG data with an accuracy of up to 100%. When using the method for an asynchronous self-paced BCI for spelling, an average utility rate of 175 bit/min was achieved, which corresponds to an average of 35 error-free letters per minute. As the presented method extracts more than three times more information than the previously fastest approach, we suggest that EEG signals carry more information than generally assumed. Finally, we observed a ceiling effect such that information content in the EEG exceeds that required for BCI control, and therefore we discuss if BCI research has reached a point where the performance of non-invasive visual BCI control cannot be substantially improved anymore.
Highlights
A brain-computer interface (BCI) is a device that translates brain signals into output signals of a computer system
As we found that linear methods can not appropriately model the visual evoked potentials (VEPs) response, this paper presents an approach to combine the EEG2Code method with deep learning to create a nonlinear model that predicts arbitrary stimulation patterns based on the VEP response, and we show how this method can be used in a BCI
The synchronous BCI control was simulated, which resulted in an average accuracy of 95.9% using the optimized stimulation patterns and a trial duration of 1 s, which in turn corresponds to an information transfer rate (ITR) of 183.1 bit/min including the inter-trial time of 0.5 s
Summary
A brain-computer interface (BCI) is a device that translates brain signals into output signals of a computer system. The BCI output is mainly used to restore several functionalities of motor disabled people, e.g., for prosthesis control or communication [1]. Besides the use of BCIs that gives the user the ability to actively control a device, passive BCIs have been accepted as a different kind of BCIs that do not have the purpose of voluntary control [2]. In the area of BCIs for communication purposes, BCIs based on visual evoked potentials (VEPs) have emerged as the fastest and most robust approach for BCI communication. Different approaches were demonstrated that use VEPs for BCI control. The majority of VEP-based BCI systems is based on frequency-modulated SSVEPs. The highest
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
