Xiaogang Chen

High-speed spelling with a noninvasive brain-computer interface.

Significance Brain–computer interface (BCI) technology provides a new communication channel. However, current applications have been severely limited by low communication speed. This study reports a noninvasive brain speller that achieved a multifold increase in information transfer rate compared with other existing systems. Based on extremely precise coding of frequency and phase in single-trial steady-state visual evoked potentials, this study developed a new joint frequency-phase modulation method and a user-specific decoding algorithm to implement synchronous modulation and demodulation of electroencephalograms. The resulting speller obtained high spelling rates up to 60 characters (∼12 words) per minute. The proposed methodological framework of high-speed BCI can lead to numerous applications in both patients with motor disabilities and healthy people. The past 20 years have witnessed unprecedented progress in brain–computer interfaces (BCIs). However, low communication rates remain key obstacles to BCI-based communication in humans. This study presents an electroencephalogram-based BCI speller that can achieve information transfer rates (ITRs) up to 5.32 bits per second, the highest ITRs reported in BCI spellers using either noninvasive or invasive methods. Based on extremely high consistency of frequency and phase observed between visual flickering signals and the elicited single-trial steady-state visual evoked potentials, this study developed a synchronous modulation and demodulation paradigm to implement the speller. Specifically, this study proposed a new joint frequency-phase modulation method to tag 40 characters with 0.5-s-long flickering signals and developed a user-specific target identification algorithm using individual calibration data. The speller achieved high ITRs in online spelling tasks. This study demonstrates that BCIs can provide a truly naturalistic high-speed communication channel using noninvasively recorded brain activities.

Filter bank canonical correlation analysis for implementing a high-speed SSVEP-based brain–computer interface

OBJECTIVE Recently, canonical correlation analysis (CCA) has been widely used in steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) due to its high efficiency, robustness, and simple implementation. However, a method with which to make use of harmonic SSVEP components to enhance the CCA-based frequency detection has not been well established. APPROACH This study proposed a filter bank canonical correlation analysis (FBCCA) method to incorporate fundamental and harmonic frequency components to improve the detection of SSVEPs. A 40-target BCI speller based on frequency coding (frequency range: 8-15.8 Hz, frequency interval: 0.2 Hz) was used for performance evaluation. To optimize the filter bank design, three methods (M1: sub-bands with equally spaced bandwidths; M2: sub-bands corresponding to individual harmonic frequency bands; M3: sub-bands covering multiple harmonic frequency bands) were proposed for comparison. Classification accuracy and information transfer rate (ITR) of the three FBCCA methods and the standard CCA method were estimated using an offline dataset from 12 subjects. Furthermore, an online BCI speller adopting the optimal FBCCA method was tested with a group of 10 subjects. MAIN RESULTS The FBCCA methods significantly outperformed the standard CCA method. The method M3 achieved the highest classification performance. At a spelling rate of ∼33.3 characters/min, the online BCI speller obtained an average ITR of 151.18 ± 20.34 bits min(-1). SIGNIFICANCE By incorporating the fundamental and harmonic SSVEP components in target identification, the proposed FBCCA method significantly improves the performance of the SSVEP-based BCI, and thereby facilitates its practical applications such as high-speed spelling.

A high-ITR SSVEP-based BCI speller

Spelling is an important application of brain-computer interfaces (BCIs). Previous BCI spellers were not suited for widespread use due to their low information transfer rate (ITR). In this study, we constructed a high-ITR BCI speller based on the steady-state visual evoked potential (SSVEP). A 45-target BCI speller was implemented with a frequency resolution of 0.2 Hz. A sampled sinusoidal stimulation method was used to present visual stimuli on a conventional LCD screen. The online results revealed that the proposed BCI speller had a good performance, reaching a high average accuracy (84.1% for 2 s stimulation time; 90.2% for 3 s stimulation time) and the corresponding high ITR (105 bits/min for 2 s stimulation time, 82 bits/min for 3 s stimulation time) during the low-frequency stimuli, while 88.7% and 61 bits/min were achieved for a 4 s time window during the high-frequency stimuli.

Hybrid frequency and phase coding for a high-speed SSVEP-based BCI speller.

Steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) have potential to realize high-speed communication between the human brain and the external environment. Recently, multiple access (MA) methods in telecommunications have been introduced into the system design of BCIs and showed their potential in improving BCI performance. This study investigated the feasibility of hybrid frequency and phase coding methods in multi-target SSVEP-based BCIs. Specifically, this study compared two hybrid target-coding strategies: (1) mixed frequency and phase coding, and (2) joint frequency and phase coding. In a simulated online BCI experiment using a 40-target BCI speller, BCI performance for both coding approaches were tested with a group of six subjects. At a spelling speed of 40 characters per minute (1.5 seconds per character), both approaches obtained high information transfer rates (ITR) (mixed coding: 172.37±28.67 bits/min, joint coding: 170.94±28.32 bits/min) across subjects. There was no statistically significant difference between the two approaches (p>0.05). These results suggest that the hybrid frequency and phase coding methods are highly efficient for multi-target coding in SSVEP BCIs with a large number of classes, providing a practical solution to implement a high-speed BCI speller.

Brain-computer interface based on intermodulation frequency.

OBJECTIVE Most recent steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI) systems have used a single frequency for each target, so that a large number of targets require a large number of stimulus frequencies and therefore a wider frequency band. However, human beings show good SSVEP responses only in a limited range of frequencies. Furthermore, this issue is especially problematic if the SSVEP-based BCI takes a PC monitor as a stimulator, which is only capable of generating a limited range of frequencies. To mitigate this issue, this study presents an innovative coding method for SSVEP-based BCI by means of intermodulation frequencies. APPROACH Simultaneous modulations of stimulus luminance and color at different frequencies were utilized to induce intermodulation frequencies. Luminance flickered at relatively large frequency (10, 12, 15 Hz), while color alternated at low frequency (0.5, 1 Hz). An attractive feature of the proposed method was that it would substantially increase the number of targets at a single flickering frequency by altering color modulated frequencies. Based on this method, the BCI system presented in this study realized eight targets merely using three flickering frequencies. MAIN RESULTS The online results obtained from 15 subjects (14 healthy and 1 with stroke) revealed that an average classification accuracy of 93.83% and information transfer rate (ITR) of 33.80 bit min(-1) were achieved using our proposed SSVEP-based BCI system. Specifically, 5 out of the 15 subjects exhibited an ITR of 40.00 bit min(-1) with a classification accuracy of 100%. SIGNIFICANCE These results suggested that intermodulation frequencies could be adopted as steady responses in BCI, for which our system could be used as a practical BCI system.

Enhancing performances of SSVEP-based brain-computer interfaces via exploiting inter-subject information.

OBJECTIVE A new training-free framework was proposed for target detection in steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) using joint frequency-phase coding. APPROACH The key idea is to transfer SSVEP templates from the existing subjects to a new subject to enhance the detection of SSVEPs. Under this framework, transfer template-based canonical correlation analysis (tt-CCA) methods were developed for single-channel and multi-channel conditions respectively. In addition, an online transfer template-based CCA (ott-CCA) method was proposed to update EEG templates by online adaptation. MAIN RESULTS The efficiency of the proposed framework was proved with a simulated BCI experiment. Compared with the standard CCA method, tt-CCA obtained an 18.78% increase of accuracy with a data length of 1.5 s. A simulated test of ott-CCA further received an accuracy increase of 2.99%. SIGNIFICANCE The proposed simple yet efficient framework significantly facilitates the use of SSVEP BCIs using joint frequency-phase coding. This study also sheds light on the benefits from exploring and exploiting inter-subject information to the electroencephalogram (EEG)-based BCIs.

Enhancing Detection of SSVEPs for a High-Speed Brain Speller Using Task-Related Component Analysis

Objective: This study proposes and evaluates a novel data-driven spatial filtering approach for enhancing steady-state visual evoked potentials (SSVEPs) detection toward a high-speed brain-computer interface (BCI) speller. Methods: Task-related component analysis (TRCA), which can enhance reproducibility of SSVEPs across multiple trials, was employed to improve the signal-to-noise ratio (SNR) of SSVEP signals by removing background electroencephalographic (EEG) activities. An ensemble method was further developed to integrate TRCA filters corresponding to multiple stimulation frequencies. This study conducted a comparison of BCI performance between the proposed TRCA-based method and an extended canonical correlation analysis (CCA)-based method using a 40-class SSVEP dataset recorded from 12 subjects. An online BCI speller was further implemented using a cue-guided target selection task with 20 subjects and a free-spelling task with 10 of the subjects. Results: The offline comparison results indicate that the proposed TRCA-based approach can significantly improve the classification accuracy compared with the extended CCA-based method. Furthermore, the online BCI speller achieved averaged information transfer rates (ITRs) of 325.33 ± 38.17 bits/min with the cue-guided task and 198.67 ± 50.48 bits/min with the free-spelling task. Conclusion: This study validated the efficiency of the proposed TRCA-based method in implementing a high-speed SSVEP-based BCI. Significance: The high-speed SSVEP-based BCIs using the TRCA method have great potential for various applications in communication and control.

An online hybrid BCI system based on SSVEP and EMG.

OBJECTIVE A hybrid brain-computer interface (BCI) is a device combined with at least one other communication system that takes advantage of both parts to build a link between humans and machines. To increase the number of targets and the information transfer rate (ITR), electromyogram (EMG) and steady-state visual evoked potential (SSVEP) were combined to implement a hybrid BCI. A multi-choice selection method based on EMG was developed to enhance the system performance. APPROACH A 60-target hybrid BCI speller was built in this study. A single trial was divided into two stages: a stimulation stage and an output selection stage. In the stimulation stage, SSVEP and EMG were used together. Every stimulus flickered at its given frequency to elicit SSVEP. All of the stimuli were divided equally into four sections with the same frequency set. The frequency of each stimulus in a section was different. SSVEPs were used to discriminate targets in the same section. Different sections were classified using EMG signals from the forearm. Subjects were asked to make different number of fists according to the target section. Canonical Correlation Analysis (CCA) and mean filtering was used to classify SSVEP and EMG separately. In the output selection stage, the top two optimal choices were given. The first choice with the highest probability of an accurate classification was the default output of the system. Subjects were required to make a fist to select the second choice only if the second choice was correct. MAIN RESULTS The online results obtained from ten subjects showed that the mean accurate classification rate and ITR were 81.0% and 83.6 bits min(-1) respectively only using the first choice selection. The ITR of the hybrid system was significantly higher than the ITR of any of the two single modalities (EMG: 30.7 bits min(-1), SSVEP: 60.2 bits min(-1)). After the addition of the second choice selection and the correction task, the accurate classification rate and ITR was enhanced to 85.8% and 90.9 bit min(-1). SIGNIFICANCE These results suggest that the hybrid system proposed here is suitable for practical use.

A study on dynamic model of steady-state visual evoked potentials

OBJECTIVE Significant progress has been made in the past two decades to considerably improve the performance of steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI). However, there are still some unsolved problems that may help us to improve BCI performance, one of which is that our understanding of the dynamic process of SSVEP is still superficial, especially for the transient-state response. APPROACH This study introduced an antiphase stimulation method (antiphase: phase [Formula: see text]), which can simultaneously separate and extract SSVEP and event-related potential (ERP) signals from EEG, and eliminate the interference of ERP to SSVEP. Based on the SSVEP signals obtained by the antiphase stimulation method, the envelope of SSVEP was extracted by the Hilbert transform, and the dynamic model of SSVEP was quantitatively studied by mathematical modeling. The step response of a second-order linear system was used to fit the envelope of SSVEP, and its characteristics were represented by four parameters with physical and physiological meanings: one was amplitude related, one was latency related and two were frequency related. This study attempted to use pre-stimulation paradigms to modulate the dynamic model parameters, and quantitatively analyze the results by applying the dynamic model to further explore the pre-stimulation methods that had the potential to improve BCI performance. MAIN RESULTS The results showed that the dynamic model had good fitting effect with SSVEP under three pre-stimulation paradigms. The test results revealed that the parameters of SSVEP dynamic models could be modulated by the pre-stimulation baseline luminance, and the gray baseline luminance pre-stimulation obtained the highest performance. SIGNIFICANCE This study proposed a dynamic model which was helpful to understand and utilize the transient characteristics of SSVEP. This study also found that pre-stimulation could be used to adjust the parameters of SSVEP model, and had the potential to improve the performance of SSVEP-BCI.