ABSTRACT:

This paper presents a new encoding/decoding approach to brain–computer interface (BCI) based on chromatic transient visual evoked potential (CTVEP). The proposed CTVEP-based encoding/decoding approach is designed to provide a safer and more comfortable stimulation method than the conventional VEP-based stimulation methods for BCI without loss of efficiency. For this purpose, low-frequency isoluminant chromatic stimuli are time-encoded to serve as different input commands for BCI control, and the superior comfortableness of the proposed stimulation method is validated by a survey. A combination of diversified signal processing techniques are further employed to decode the information from CTVEP. Based on experimental results, a properly designed configuration of the CTVEP-based stimulation method and a tailored signal processing framework are developed. It is demonstrated that high performance (at information transfer rate: 58.0 bits/min, accuracy: 94.9%, false alarm rate: 1.3%) for BCI can be achieved by means of the CTVEP-based encoding/decoding approach. It turns out that to achieve such good performance, only simple signal processing algorithms with very low computational complexity are required, which makes the method suitable for the development of a practical BCI system. A preliminary prototype of such a system has been implemented with demonstrated applicability.