Guangyu Bin

An online multi-channel SSVEP-based brain-computer interface using a canonical correlation analysis method

In recent years, there has been increasing interest in using steady-state visual evoked potential (SSVEP) in brain-computer interface (BCI) systems. However, several aspects of current SSVEP-based BCI systems need improvement, specifically in relation to speed, user variation and ease of use. With these improvements in mind, this paper presents an online multi-channel SSVEP-based BCI system using a canonical correlation analysis (CCA) method for extraction of frequency information associated with the SSVEP. The key parameters, channel location, window length and the number of harmonics, are investigated using offline data, and the result used to guide the design of the online system. An SSVEP-based BCI system with six targets, which use nine channel locations in the occipital and parietal lobes, a window length of 2 s and the first harmonic, is used for online testing on 12 subjects. The results show that the proposed BCI system has a high performance, achieving an average accuracy of 95.3% and an information transfer rate of 58 +/- 9.6 bit min(-1). The positive characteristics of the proposed system are that channel selection and parameter optimization are not required, the possible use of harmonic frequencies, low user variation and easy setup.

VEP-based brain-computer interfaces: time, frequency, and code modulations [Research Frontier]

A brain computer interface (BCI) translates human intentions into control signals to establish a direct communication channel between the human brain and external devices. Because a BCI does not depend on the brains normal output pathways of peripheral nerves and muscles, it can provide a new communication channel to people with severe motor disabilities. Electroencephalograms (EEGs) recorded from the surface of the scalp are widely used in current BCIs for their non-invasive nature and easy applications. Among EEG based BCIs, systems based on visual evoked potentials (VEPs) have received widespread attention in recent decades. We described the three stimulus modulation approaches used in current VEP based BCIs: time modulation (t-VEP), frequency modulation (f-VEP), and pseudorandom code modulation (c-VEP). We then carried out a detailed comparison of system performance between an f-VEP BCI and a c-VEP BCI. The results show that an f-VEP BCI has the advantage of little or no training and simple system configuration, while the c-VEP based BCI has a higher communication speed. The stimulus modulation design is the crux of VEP based BCI systems.

A high-speed BCI based on code modulation VEP

Recently, electroencephalogram-based brain-computer interfaces (BCIs) have attracted much attention in the fields of neural engineering and rehabilitation due to their noninvasiveness. However, the low communication speed of current BCI systems greatly limits their practical application. In this paper, we present a high-speed BCI based on code modulation of visual evoked potentials (c-VEP). Thirty-two target stimuli were modulated by a time-shifted binary pseudorandom sequence. A multichannel identification method based on canonical correlation analysis (CCA) was used for target identification. The online system achieved an average information transfer rate (ITR) of 108 ± 12 bits min(-1) on five subjects with a maximum ITR of 123 bits min(-1) for a single subject.

A study of the existing problems of estimating the information transfer rate in online brain–computer interfaces

OBJECTIVE Today, the brain-computer interface (BCI) community lacks a standard method to evaluate an online BCIs performance. Even the most commonly used metric, the information transfer rate (ITR), is often reported differently, even incorrectly, in many papers, which is not conducive to BCI research. This paper aims to point out many of the existing problems and give some suggestions and methods to overcome these problems. APPROACH First, the preconditions inherent in ITR calculation based on Wolpaws definition are summarized and several incorrect ITR calculations, which go against the preconditions, are indicated. Then, the issues affecting ITR estimation during the test of online BCI systems are discussed in detail. Finally, a task-oriented online BCI test platform was proposed, which may help BCI evaluations in real-world applications. MAIN RESULTS The guidelines for ITR calculation in online BCIs testing are proposed. The platform executed in the Beijing BCI Competition 2010 shows that it can be used as a common way to compare the online performances (including the ITR) of existing BCI paradigms. SIGNIFICANCE The proposed guidelines and task-oriented test platform may reduce the uncertainty and artifacts of online BCI performance evaluation; they provide a relatively objective way to compare different BCIs performances in real-world BCI applications, which is a forward step toward developing standards for BCI performance evaluation.

A half-field stimulation pattern for SSVEP-based brain-computer interface

A novel stimulation pattern has been designed for brain-computer interface (BCI) using steady-state visual evoked potential (SSVEP) signals. Each target is composed of two flickers placed on right-and-left visual fields. The user is expected to concentrate his or her sight on the fixation point which is located in the middle of the two flickers modulated at specific frequencies respectively. Considering the role of optic chiasm, the two frequency components could be extracted from contralateral occipital regions. Canonical correlation analysis (CCA) was applied to distinguish the electroencephalography (EEG) frequency components from right-and-left visual cortex. The attractive feature of this method is that it would substantially increase the number of targets by a combination of frequencies. Based on this technique a nine-target SSVEP-based BCI system was designed using only three different frequencies. The test results with 8 subjects showed a classification accuracy between 40.0% and 96.3%.

The SSVEP topographic scalp maps by Canonical correlation analysis

As the number of electrodes increases, topographic scalp mapping methods for electroencephalogram (EEG) data analysis are becoming important. Canonical correlation analysis (CCA) is a method of extracting similarity between two data sets. This paper presents an EEG topographic scalp mapping -based CCA for the steady-state visual evoked potentials (SSVEP) analysis. Multi-channel EEG data and the sinusoidal reference signal were used as the inputs of CCA. The output linear combination was then employed for mapping. Our experimental results prove the topographic scalp mapping-based CCA can instruct for the improvement of SSVEP-based brain computer interface (BCI) system.

Analysis of phase coding SSVEP based on canonical correlation analysis (CCA)

Steady-state visual evoked potential (SSVEP) has been widely applied in brain computer interface (BCI) systems. The amplitude and phase features of SSVEP were commonly extracted by Fourier analysis method from single-channel EEG data. In the multichannel case, canonical correlation analysis (CCA) has been utilized for the analysis of frequency coding SSVEP. This paper presents the analysis of phase coding SSVEP using CCA. The phase coding scheme consists of six targets flickering at 10Hz, with a 60° phase difference between any two sequential targets. For each target, 20 trials of 8s EEG signal were acquired. Using CCA, we achieve channel selection and extraction of phase features; a classification accuracy of above 80% is obtained, with the length of the time window up to 4s. The results demonstrate that phase coding SSVEP analysis based on CCA is feasible.

Idle State Detection in SSVEP-Based Brain-Computer Interfaces

In recent years, the rapid development of Brain-Computer Interfaces in the laboratory has prepared a solid foundation for its application to real life situations. Among the techniques developed, the Steady-State Visual Evoked Potential (SSVEP)-based BCI is a promising one. Its stability and speed make it applicable in the near future. To realize its practicability, a workable method needs to be worked out to detect the idle state. In this paper, a method using C0 complexity, Principal Component Analysis (PCA) and Singular Spectrum Analysis (SSA) is proposed. This method can be called Principal-Component Co Complexity (PCC0). The results show that the idle state can be determined using this method with 90% accuracy when SSVEP can be detected with an average accuracy of 80%. This approach can be further developed for use in online asynchronous BCI systems.

A coded VEP method to measure interhemispheric transfer time (IHTT)

Interhemispheric transfer time (IHTT) is an important parameter for research on the information conduction time across the corpus callosum between the two hemispheres. There are several traditional methods used to estimate the IHTT, including the reaction time (RT) method, the evoked potential (EP) method and the measure based on the transcranial magnetic stimulation (TMS). The present study proposes a novel coded VEP method to estimate the IHTT based on the specific properties of the m-sequence. These properties include good signal-to-noise ratio (SNR) and high noise tolerance. Additionally, calculation of the circular cross-correlation function is sensitive to the phase difference. The method presented in this paper estimates the IHTT using the m-sequence to encode the visual stimulus and also compares the results with the traditional flash VEP method. Furthermore, with the phase difference of the two responses calculated using the circular cross-correlation technique, the coded VEP method could obtain IHTT results, which does not require the selection of the utilized component.

Space distribution of EEG responses to hanoi-moving visual and auditory stimulation with Fourier Independent Component Analysis.

Background and objective: The relationship between EEG source signals and action-related visual and auditory stimulation is still not well-understood. The objective of this study was to identify EEG source signals and their associated action-related visual and auditory responses, especially independent components of EEG. Methods: A hand-moving-Hanoi video paradigm was used to study neural correlates of the action-related visual and auditory information processing determined by mu rhythm (8–12 Hz) in 16 healthy young subjects. Independent component analysis (ICA) was applied to identify separate EEG sources, and further computed in the frequency domain by applying-Fourier transform ICA (F-ICA). Results: F-ICA found more sensory stimuli-related independent components located within the sensorimotor region than ICA did. The total number of independent components of interest from F-ICA was 768, twice that of 384 from traditional time-domain ICA (p < 0.05). In the sensory-motor region C3 or C4, the total source signals intensity distribution values from all 14 subjects was 23.00 (Mean 1.64 ± 1.17) from F-ICA; which was more than the 10.5 (Mean 0.75 ± 0.62) from traditional time-domain ICA (p < 0.05). Furthermore, the intensity distribution of source signals in the C3 or C4 region was statistically significant between the ICA and F-ICA groups (strong 50 vs. 92%; weak 50 vs. 8% retrospectively; p < 0.05). In the Pz region, the total source signal intensity distribution from F-ICA was 12.50 (Mean 0.89 ± 0.53); although exceeding that of traditional time-domain ICA 8.20 (Mean 0.59 ± 0.48), the difference was not statistically significant (p > 0.05). Conclusions: These results support the hypothesis that mu rhythm was sensitive to detection of the cognitive expression, which could be reflected by the function in the parietal lobe sensory-motor region. The results of this study could potentially be applied into early diagnosis for those with visual and hearing impairments in the near future.