Information Transfer Rate Research Articles

A survey of deep learning-based classification methods for steady-state visual evoked potentials

Purpose Steady-state visual evoked potential (SSVEP) based BCI has attracted great interests owing to the high information transfer rate (ITR) and little training requirement. The performance of SSVEP-based BCI heavily depends on the classification methods. Deep Learning (DL) technology provides an alternative avenue for the data classification in SSVEP-based BCI, and has received increasing interests in recent years. This review aimed to summarize the progress of DL-based classification methods for SSVEP data over the past decade.Materials and method The literature was searched and selected based on the research topics of DL and SSVEP. We categorized these methods into four classes, i.e., traditional neural network structures-based DL methods, traditional frequency recognition methods inspiring DL methods, attention mechanisms-based DL models, and transfer learning technology-based DL methods, and generative model-based recognition method. Moreover, we analyzed the current challenges and presented future research opportunities.Conclusions This study provides a systematic description of the current development status on DL-based SSVEP classification methods, and sheds insight on future researches.

Optimization of the Guaranteed Information Transmission Rate by Pseudorandom Signals with a Randomized Base in Conditions of Intentional Interference

The principles of analysis and optimization of the parameters and signal processing mode in an information transmission system with a reverse channel operating under the influence of intentional interference, the structure of which can be selected from the position of disruption (deterioration) of the system have been considered. An algorithm for the formation and processing of pseudorandom signals with a randomized base is proposed, which provides an increase in the speed of information transmission by the system in theclass of interference with limited average power while ensuring the initial reliability indicators. The methodology is presented, calculation examples are given, and graphs of the dependences of the guaranteed informationtransfer rate in a system with a randomized two-stage switching of the signal base from the average power of the optimized interference are shown.

A hybrid steady-state visual evoked response-based brain-computer interface with MEG and EEG

While recent developments in electroencephalogram (EEG)-based brain-computer interfaces (BCIs) have enabled a bridge between the brain and external devices with relatively high communication speed, there is still room for improvement. Notably, the phenomenon of “BCI illiteracy,” which refers to the 15%–30% of people who struggle to type or control devices using BCI, remains unsolved, limiting the practical application of BCI systems. The EEG-based BCIs performance is constrained by the low-quality scalp EEG signals due to the attenuation and distortion of the skull. To address these limitations, this study proposes a hybrid BCI system combining EEG with magnetoencephalogram (MEG), a neuroimaging technology not influenced by the volume conduction effect, to boost BCI performance by enhancing signal quality. Comparative experiments involving 22 subjects showed that the steady-state visual evoked response (SSVER) from MEG has a wider range of effective bandwidth and higher signal-to-noise ratio than EEG. Moreover, differences in the spectral and spatiotemporal characteristics of MEG and EEG explain better performance. Simultaneous MEG-EEG recording experiments suggested that the hybrid MEG-EEG BCI achieved a significantly higher information transfer rate than either modality alone (hybrid: 312 ± 17 bits/min, MEG: 272 ± 17 bits/min, EEG: 240 ± 27 bits/min). Moreover, the 40-target classification accuracy of “BCI illiterate” increased from 50% to 95% with the help of MEG. These results highlight the methodological advantages of a hybrid MEG-EEG BCI, suggesting a promising paradigm for implementing high-speed BCIs.

An online hybrid BCI combining SSVEP and EOG-based eye movements.

Hybrid brain-computer interface (hBCI) refers to a system composed of a single-modality BCI and another system. In this paper, we propose an online hybrid BCI combining steady-state visual evoked potential (SSVEP) and eye movements to improve the performance of BCI systems. Twenty buttons corresponding to 20 characters are evenly distributed in the five regions of the GUI and flash at the same time to arouse SSVEP. At the end of the flash, the buttons in the four regions move in different directions, and the subject continues to stare at the target with eyes to generate the corresponding eye movements. The CCA method and FBCCA method were used to detect SSVEP, and the electrooculography (EOG) waveform was used to detect eye movements. Based on the EOG features, this paper proposes a decision-making method based on SSVEP and EOG, which can further improve the performance of the hybrid BCI system. Ten healthy students took part in our experiment, and the average accuracy and information transfer rate of the system were 94.75% and 108.63 bits/min, respectively.

A new benchmark dataset for P300 ERP-based BCI applications

Because of its non-invasive nature, one of the most commonly used event-related potentials in brain-computer interface (BCI) system designs is the P300 electroencephalogram (EEG) signal. The fact that the P300 response can easily be stimulated and measured is particularly important for participants with severe motor disabilities. In order to train and test P300-based BCI speller systems in more realistic high-speed settings, there is a pressing need for a large and challenging benchmark dataset. Various datasets already exist in the literature but most of them are not publicly available, and they either have a limited number of participants or utilize relatively long stimulus duration (SD) and inter-stimulus intervals (ISI). They are also typically based on a 36 target (6×6) character matrix. The use of long ISI, in particular, not only reduces the speed and the information transfer rates (ITRs) but also oversimplifies the P300 detection. This leaves a limited challenge to state-of-the-art machine learning and signal processing algorithms. In fact, near-perfect P300 classification accuracies are reported with the existing datasets. Therefore, one certainly needs a large-scale dataset with challenging settings to fully exploit the recent advancements in algorithm design (machine learning and signal processing) and achieve high-performance speller results. To this end, in this article we introduce a new freely- and publicly-accessible P300 dataset obtained using 32-channel EEG, in the hope that it will lead to new research findings and eventually more efficient BCI designs. The introduced dataset comprises 18 participants performing a 40-target (5×8) cued-spelling task, with reduced SD (66.6 ms) and ISI (33.3 ms) for fast spelling. We have also processed, analyzed, and character-classified the introduced dataset and we presented the accuracy and ITR results as a benchmark. The introduced dataset and the codes of our experiments are publicly accessible at https://data.mendeley.com/datasets/vyczny2r4w.

A portable SSVEP-BCI system for rehabilitation exoskeleton in augmented reality environment

In order to strengthen the participation of stroke patients in rehabilitation training and weaken the dependence of steady-state visually evoked potential (SSVEP)-based brain-computer interface (BCI) on external stimuli equipment, we build an augmented-reality-based brain-computer interface (AR-BCI) system applied to rehabilitation exoskeleton. The system uses HoloLens to design a 4-category BCI, and adopts the sequential logic decoding method to control sixteen rehabilitation movements. In offline experiments, the performance of AR-BCI is compared with computer screen-based brain-computer interface (CS-BCI), and the effects of data length, number and position of electrodes on the performance of BCI are studied. Then, the instruction classification accuracy of AR-BCI and movement accuracy of exoskeleton are evaluated in online rehabilitation training. The average recognition accuracy of AR-BCI is 90.2% in offline experiments, which is smaller gap with CS-BCI. The recognition accuracy still reaches more than 90% when only Oz and O2 electrodes are used. The online results show that the instruction classification accuracy of AR-BCI is 88.9% and the averaged information transfer rates (ITR) is 30.01 bits min−1 under the data length of 2.5 s. The movement accuracy of exoskeleton is 91.12% with the ITR of 31.63 bits min−1, which is 2.2% higher than instruction recognition accuracy of AR-BCI. These results show that AR-BCI provides a high-performance and more friendly human–computer interaction method, and greatly improves the application potential of wearable BCI.

A deep learning method for classification of steady-state visual evoked potentials in a brain-computer interface speller

ABSTRACT The main drawback of SSVEP-based BCIs is the lack of a classifier that categorizes SSVEPs with high accuracy and Information Transfer Rate (ITR). Addressing this, we proposed a deep convolutional neural network (CNN) for classifying a 40-class SSVEP. Time windows of length 2 and 3.5 seconds were used for training and testing the model by leave-one-subject-out cross-validation, using nine, three, and single-channel EEG. The proposed model reached 88.5% average accuracy for the nine-channel EEG with the mean and max ITR of 72 and 91.23 bpm, respectively. It outperformed the previous deep learning methods for SSVEP-based BCIs, in terms of accuracy and ITR. In the three-channel experiment the mean accuracy and ITR were 76.02% and 40.1 bpm. In single-channel implementation, O1 channel achieved 77.38 % average accuracy (highest) and the mean ITR was 57.51 bpm. The model showed promising performance to put this technology forward and make it more practical.

A high-performance SSVEP-based BCI using imperceptible flickers

Objective. Existing steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) struggle to balance user experience and system performance. This study proposed an individualized space and phase modulation method to code imperceptible flickers at 60 Hz towards a user-friendly SSVEP-based BCI with high performance. Approach. The individualized customization of visual stimulation took the subject-to-subject variability in cortex geometry into account. An annulus global-stimulation was divided into local-stimulations of eight annular sectors and presented to subjects separately. The local-stimulation SSVEPs were superimposed to simulate global-stimulation SSVEPs with 47 space and phase coding combinations. A four-class phase-coded BCI diagram was used to evaluate the simulated classification performance. The performance ranking of all simulated global-stimulation SSVEPs were obtained and three performance levels (optimal, medium, worst) of individualized modulation groups were searched for each subject. The standard-modulation group conforming to the V1 ‘cruciform’ geometry and the non-modulation group were involved as controls. A four-target phase-coded BCI system with SSVEPs at 60 Hz was implemented with the five modulation groups and questionnaires were used to evaluate user experience. Main results. The proposed individualized space and phase modulation method effectively modulated the SSVEP intensity without affecting the user experience. The online BCI system using the 60 Hz stimuli achieved mean information transfer rates of 52.8 ± 1.9 bits min−1, 16.8 ± 2.4 bits min−1, and 42.4 ± 3.0 bits min−1 with individualized optimal-modulation, individualized worst-modulation, and non-modulation groups, respectively. Significance. Structural and functional characteristics of the human visual cortex were exploited to enhance the response intensity of SSVEPs at 60 Hz, resulting in a high-performance BCI system with good user experience. This study has important theoretical significance and application value for promoting the development of the visual BCI technology.

Модернизация авиационной системы радиосвязи с целью повышения помехозащищенности радиоканала

The article considers the possible modernization of the well-known aeronautical radio station in order to increase the noise immunity of radio communications. Uninterrupted noise-immune radio communication between the aircraft and the air traffic controller is an important factor in ensuring flight safety. The radio station provides voice (telephone) communication and telecode digital communication, with voice information transmitted in analog format this causes radio interference distorting the transmitted messages. Increased noise immunity can be achieved by transmitting voice information in a digital form, the special noise protection measures relevant only for digital transmission being applied. Antinoise coding and pseudo-random software frequency hopping can be used as the preventive measures. The task was set to investigate the problem of modernizing a well-known aviation radio station with the minimal changes in its configuration. There has been proposed a method providing the digital radio communication with modification of a prefabricated aeronautical radio station. A digital speech signal developing in this case is provided due to the resources of the radio station used to transmit the telecode information. The information transfer rate increases from 16 to 128 kbps, which leads to the increased frequency deviation and, as a result, to the expanded signal spectrum. The noise immunity of the system was also assessed in a communication range of 100 km: the signal power, signal-to-noise ratio, noise spectral density at the receiver input, transmitting antenna directivity were determined. Based on the results of the calculations, it was inferred that due to realizing the proposed measures a manifold multiplication of the required power of the jamming

Brain–computer interface in an inter-individual approach using spatial coherence: Identification of better channels and tests repetition using auditory selective attention

Auditory Brain–Computer Interfaces (BCIs) have been studied with the main purpose of improving the quality of life of totally paralyzed people. BCIs based on Auditory Selective Attention (ASA) may have distinct features, such as the number of sound sources and most approaches presented in literature use individualized settings. This setup requires more training of individuals or adjustments of the internal classifiers. In this context, a generalized approach can be an interesting alternative for an interface tailored to application and without the need for exhaustive training. The present work investigates ASA using stimuli with AM modulation and spatial coherence to evaluate its value in the modulating frequency and the electrodes position. The objective is to identify better combinations of electrodes and evaluate the number of repetitions and the intervals between them in an inter-individual approach. The best result obtained using the modular classifier, proposed in an earlier work, reached average hit rates of 75% and information transfer rate (ITR) of 2.217 bits/min, considering three windows (5.1 s). This result was obtained using AM stimuli (500 Hz and 2 kHz carriers) and a combination of mostly frontal and prefrontal electrodes and considering 5 s interval between repetitions. With these results, an implementation of a vision-free BCI that covers inter-individual differences could allow users to communicate through selective auditory attention without previous training of subjects.

A high-speed hybrid brain-computer interface with more than 200 targets

Objective. Brain-computer interfaces (BCIs) have recently made significant strides in expanding their instruction set, which has attracted wide attention from researchers. The number of targets and commands is a key indicator of how well BCIs can decode the brain’s intentions. No studies have reported a BCI system with over 200 targets. Approach. This study developed the first high-speed BCI system with up to 216 targets that were encoded by a combination of electroencephalography features, including P300, motion visual evoked potential (mVEP), and steady-state visual evoked potential (SSVEP). Specifically, the hybrid BCI paradigm used the time-frequency division multiple access strategy to elaborately tag targets with P300 and mVEP of different time windows, along with SSVEP of different frequencies. The hybrid features were then decoded by task-discriminant component analysis and linear discriminant analysis. Ten subjects participated in the offline and online cued-guided spelling experiments. Other ten subjects took part in online free-spelling experiments. Main results. The offline results showed that the mVEP and P300 components were prominent in the central, parietal, and occipital regions, while the most distinct SSVEP feature was in the occipital region. The online cued-guided spelling and free-spelling results showed that the proposed BCI system achieved an average accuracy of 85.37% ± 7.49% and 86.00% ± 5.98% for the 216-target classification, resulting in an average information transfer rate (ITR) of 302.83 ± 39.20 bits min−1 and 204.47 ± 37.56 bits min−1, respectively. Notably, the peak ITR could reach up to 367.83 bits min−1. Significance. This study developed the first high-speed BCI system with more than 200 targets, which holds promise for extending BCI’s application scenarios.

Comparison of Two Paradigms Based on Stimulation with Images in a Spelling Brain-Computer Interface.

A P300-based speller can be used to control a home automation system via brain activity. Evaluation of the visual stimuli used in a P300-based speller is a common topic in the field of brain-computer interfaces (BCIs). The aim of the present work is to compare, using the usability approach, two types of stimuli that have provided high performance in previous studies. Twelve participants controlled a BCI under two conditions, which varied in terms of the type of stimulus employed: a red famous face surrounded by a white rectangle (RFW) and a range of neutral pictures (NPs). The usability approach included variables related to effectiveness (accuracy and information transfer rate), efficiency (stress and fatigue), and satisfaction (pleasantness and System Usability Scale and Affect Grid questionnaires). The results indicated that there were no significant differences in effectiveness, but the system that used NPs was reported as significantly more pleasant. Hence, since satisfaction variables should also be considered in systems that potential users are likely to employ regularly, the use of different NPs may be a more suitable option than the use of a single RFW for the development of a home automation system based on a visual P300-based speller.

Transfer learning of an ensemble of DNNs for SSVEP BCI spellers without user-specific training

Objective. Steady-state visually evoked potentials (SSVEPs), measured with electroencephalogram (EEG), yield decent information transfer rates (ITRs) in brain-computer interface (BCI) spellers. However, the current high performing SSVEP BCI spellers in the literature require an initial lengthy and tiring user-specific training for each new user for system adaptation, including data collection with EEG experiments, algorithm training and calibration (all are before the actual use of the system). This impedes the widespread use of BCIs. To ensure practicality, we propose a novel target identification method based on an ensemble of deep neural networks (DNNs), which does not require any sort of user-specific training. Approach. We exploit already-existing literature datasets from participants of previously conducted EEG experiments to train a global target identifier DNN first, which is then fine-tuned to each participant. We transfer this ensemble of fine-tuned DNNs to the new user instance, determine the k most representative DNNs according to the participants’ statistical similarities to the new user, and predict the target character through a weighted combination of the ensemble predictions. Main results. The proposed method significantly outperforms all the state-of-the-art alternatives for all stimulation durations in [0.2–1.0] s on two large-scale benchmark and BETA datasets, and achieves impressive 155.51 bits/min and 114.64 bits/min ITRs. Code is available for reproducibility: https://github.com/osmanberke/Ensemble-of-DNNs. Significance. Our Ensemble-DNN method has the potential to promote the practical widespread deployment of BCI spellers in daily lives as we provide the highest performance while enabling the immediate system use without any user-specific training.

A Precise Frequency Recognition Method of Short-time SSVEP Signals Based on Signal Extension.

Improving the Information Transfer Rate (ITR) is a popular research topic in steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs). The higher recognition accuracy of short-time SSVEP signal is critical to improving ITR and achieving high-speed SSVEP-BCIs. However, the existing algorithms have unsatisfactory performance on recognizing short-time SSVEP signals, especially for calibration-free methods. This study for the first time proposed improving the recognition accuracy of short-time SSVEP signals based on the calibration-free method by extending the SSVEP signal length. A signal extension model based on Multi-channel adaptive Fourier decomposition with differ-ent Phase (DP-MAFD) is proposed to achieve signal exten-sion. Then the Canonical Correlation Analysis based on signal extension (SE-CCA) is proposed to complete the recognition and classification of SSVEP signals after exten-sion. The similarity study and SNR comparison analysis on public SSVEP datasets demonstrate that the proposed signal extension model has the ability to extend SSVEP signals. The classification results show that the pro-posed method outperforms Canonical Correlation Analysis (CCA) and Filter Bank Canonical Correlation Analysis (FBCCA) significantly in the measure of classification accu-racy and information transmission rate (ITR), especially for short-time signals. The highest ITR of SE-CCA is improved to 175.61 bits/min at around 1s, while CCA is 100.55 bits/min at 1.75s and FBCCA is 141.76 bits/min at 1.25s. The signal extension method can improve the recognition accuracy of short-time SSVEP signals and further improve the ITR of SSVEP-BCIs.

Fifty-selective SSVEP-BCI Speller with CCA

In this study, we used canonical correlation analysis (CCA), frequency component correlation method (FCCM) and ensemble model to develop a steady state visual evoked potential brain-computer interface (SSVEP-BCI) with fifty-selective. For the proposed fifty-selective SSVEP-BCI with only CCA, it was not possible to obtain sufficient SSVEP induction. In a previous study where a similar problem occurred, the maximum accuracy was 79.53%, and the information transfer rate (ITR) was 45.16 bits/min. Therefore, we proposed FCCM and ensemble model in CCA to improve the accuracy even when the SSVEP induction was not sufficient. We used the proposed method to achieve the highest accuracy of 93.23% and the highest ITR of 58.88 bits/min. The system also achieved an average accuracy of 71.01% and an average ITR of 40.79 bits/min, demonstrating the usefulness of the system. Also, the maximum accuracy and ITR in additional experiments were 98.53% and 65.41 bits/min.

Decoding Imagined Musical Pitch From Human Scalp Electroencephalograms.

Brain-computer interfaces (BCIs) can restore impaired cognitive functions in people with neurological disorders such as stroke. Musical ability is a cognitive function that is correlated with non-musical cognitive functions, and restoring it can enhance other cognitive functions. Pitch sense is the most relevant function to musical ability according to previous studies of amusia, and thus decoding pitch information is crucial for BCIs to be able to restore musical ability. This study evaluated the feasibility of decoding pitch imagery information directly from human electroencephalography (EEG). Twenty participants performed a random imagery task with seven musical pitches (C4-B4). We used two approaches to explore EEG features of pitch imagery: multiband spectral power at individual channels (IC) and differences between bilaterally symmetric channels (DC). The selected spectral power features revealed remarkable contrasts between left and right hemispheres, low- (< 13 Hz) and high-frequency ( 13 Hz) bands, and frontal and parietal areas. We classified two EEG feature sets, IC and DC, into seven pitch classes using five types of classifiers. The best classification performance for seven pitches was obtained using IC and multiclass Support Vector Machine with an average accuracy of 35.68±7.47% (max. 50%) and an information transfer rate (ITR) of 0.37±0.22 bits/sec. When grouping the pitches to vary the number of classes (K = 2-6), the ITR was similar across K and feature sets, suggesting the efficiency of DC. This study demonstrates for the first time the feasibility of decoding imagined musical pitch directly from human EEG.

Optimizing Stimulus Frequency Ranges for Building a High-Rate High Frequency SSVEP-BCI.

The brain-computer interfaces (BCIs) based on steady-state visual evoked potential (SSVEP) have been extensively explored due to their advantages in terms of high communication speed and smaller calibration time. The visual stimuli in the low- and medium-frequency ranges are adopted in most of the existing studies for eliciting SSVEPs. However, there is a need to further improve the comfort of these systems. The high-frequency visual stimuli have been used to build BCI systems and are generally considered to significantly improve the visual comfort, but their performance is relatively low. The distinguishability of 16-class SSVEPs encoded by the three frequency ranges, i.e., 31-34.75 Hz with an interval of 0.25 Hz, 31-38.5 Hz with an interval of 0.5 Hz, 31-46 Hz with an interval of 1 Hz, is explored in this study. We compare classification accuracy and information transfer rate (ITR) of the corresponding BCI system. According to the optimized frequency range, this study builds an online 16-target high frequency SSVEP-BCI and verifies the feasibility of the proposed system based on 21 healthy subjects. The BCI based on visual stimuli with the narrowest frequency range, i.e., 31-34.5 Hz, have the highest ITR. Therefore, the narrowest frequency range is adopted to build an online BCI system. An averaged ITR obtained from the online experiment is 153.79 ± 6.39 bits/min. These findings contribute to the development of more efficient and comfortable SSVEP-based BCIs.

Phase-locked Time-shift Data Augmentation Method for SSVEP Brain-Computer Interfaces.

Steady-state visual evoked potential (SSVEP) based brain-computer interfaces (BCIs) have achieved an information transfer rate (ITR) of over 300 bits/min, but abundant training data is required. The performance of SSVEP algorithms deteriorates greatly under limited data, and the existing time-shift data augmentation method fails to improve it because the phase-locked requirement between training samples is violated. To address this issue, this study proposes a novel augmentation method, namely phase-locked time-shift (PLTS), for SSVEP-BCI. The similarity between epochs at different time moments was evaluated, and a unique time-shift step was calculated for each class to augment additional data epochs in each trial. The results showed that the PLTS significantly improved the classification performance of SSVEP algorithms on the BETA SSVEP datasets. Moreover, under the condition of one calibration block, by slightly prolonging the calibration duration (from 48 s to 51.5 s), the ITR increased from 40.88 ±4.54 bits/min to 122.61 ± 7.05 bits/min with the PLTS. This study provides a new perspective on augmenting data epochs for training-based SSVEP-BCI, promotes the classification accuracy and ITR under limited training data, and thus facilitates the real-life applications of SSVEP-based brain spellers.

Adaptive Window Method Based on FBCCA for Optimal SSVEP Recognition.

In the conventional studies related to steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs), the window length (detection time) was typically predetermined through the offline analysis, which had limitations of practical applicability of a BCI system due to the inter-subject/trial variability of electroencephalography (EEG) signals. To address these limitations, this study aims to automatically optimize the window length for each trial based on training-free approaches and proposes a novel adaptive window method (ANCOVA-based filter-bank canonical correlation analysis, ABFCCA) for SSVEP-based BCIs. The proposed method is based on analysis of covariance (ANCOVA) which is applied after feature extraction by the conventional training-free SSVEP recognition approaches. To evaluate the performance of the proposed method, conventional fixed window and recent adaptive window methods were compared using two open-access datasets. In the Benchmark dataset, the average information transfer rate (ITR) was 146.81 bits/min, the average accuracy 93.55%, and the average window length 1.53 s. In the OpenBMI dataset, the average ITR was 119.01 bits/min, the average accuracy 83.50%, and the average window length 0.65 s. The proposed method significantly outperformed the conventional approaches with fixed window in terms of the accuracy and ITR, and is applicable to various SSVEP-based BCI paradigms based on the criterion of significance level without offline analysis to find optimal hyper-parameters. ABFCCA is enabled the practical use of various BCI systems by automatically optimizing the window length independently.

A Hybrid Method Fusing Frequency Recognition With Attention Detection to Enhance an Asynchronous Brain-Computer Interface.

One critical problem in controlling an asynchronous brain-computer interface (BCI) system is to discriminate between control and idle states. This paper proposes a hybrid attention detection and frequency recognition method based on weighted Dempster-Shafer theory (ADFR-DS), which integrates information of different aspects of the task from two brain regions, to enhance asynchronous control performance of a steady-state visual evoked potential (SSVEP)-based BCI system. The ADFR-DS method utilizes a hybrid architecture to process electroencephalogram (EEG) data from different channels simultaneously: an individualized frequency band based optimized complex network (IFBOCN) algorithm processes neural activity from the prefrontal area for attention detection, and an ensemble task-related component analysis (eTRCA) algorithm processes data from the occipital area for frequency recognition. The ADFR-DS method then fuses their classification results at decision level to generate the final output of the BCI system. A novel weighted Dempster-Shafer fusion method was proposed to enhance the fusion performance. This study evaluated the proposed method using a 40-target dataset recorded from 35 participants. The proposed method outperformed the eTRCA algorithm in the true positive rate (TPR), true negative rate (TNR), accuracy (ACC) and information transfer rate (ITR). Specifically, ADFR-DS improved the average ACC of eTRCA from 62.71% to 69.30%, and improved the average ITR from 184.28 bits/min to 216.89 bits/min (data length 0.3 s). The results suggest that the proposed ADFR-DS method can enhance asynchronous SSVEP-based BCI systems.