$Goal:$ Traditional visual brain-computer interfaces (BCIs) preferred to use large-size stimuli to attract the user's attention and elicit distinct electroencephalography (EEG) features. However, the visual stimuli are of no interest to the users as they just serve as the hidden codes behind the characters. Furthermore, using stronger visual stimuli could cause visual fatigue and other adverse symptoms to users. Therefore, it's imperative for visual BCIs to use small and inconspicuous visual stimuli to code characters. $Methods:$ This study developed a new BCI speller based on miniature asymmetric visual evoked potentials (aVEPs), which encodes thirty-two characters with a space-code division multiple access (SCDMA) scheme and decodes EEG features with a discriminative canonical pattern matching (DCPM) algorithm. Notably, the visual stimulus used in this study only subtended 0.5° of visual angle and was placed outside the fovea vision on the lateral side, which could only induce a miniature potential about 0.5 μV in amplitude and about -16.5 dB in signal-to-noise rate (SNR). Twelve subjects were recruited to use the miniature aVEP speller in both offline and online tests. $Results:$ Information transfer rates up to 63.33 bits/min could be achieved from online tests. $Conclusion:$ Experimental results demonstrate the feasibility of using very small and inconspicuous visual stimuli to implement an efficient BCI system, even though the elicited EEG features are very weak. $Significance:$ The proposed innovative technique can broaden the category of BCIs and strengthen the brain-computer communication.

Objective: Recently, electroencephalography (EEG)- based brain-computer interfaces (BCIs) have made tremendous progress in increasing communication speed. However, current BCI systems could only implement a small number of command codes, which hampers their applicability. Methods: This study developed a high-speed hybrid BCI system containing as many as 108 instructions, which were encoded by concurrent P300 and steady-state visual evoked potential (SSVEP) features and decoded by an ensemble task-related component analysis method. Notably, besides the frequency-phase-modulated SSVEP and time-modulated P300 features as contained in the traditional hybrid P300 and SSVEP features, this study found two new distinct EEG features for the concurrent P300 and SSVEP features, i.e. time-modulated SSVEP and frequency-phase- modulated P300. Ten subjects spelled in both offline and online cued-guided spelling experiments. Other ten subjects took part in online copy-spelling experiments. Results: Offline analyses demonstrate that the concurrent P300 and SSVEP features can provide adequate classification information to correctly select the target from 108 characters in 1.7 seconds. Online cued-guided spelling and copy-spelling tests further show that the proposed BCI system can reach an average information transfer rate (ITR) of 172.46±32.91 bits/min and 164.69±33.32 bits/min respectively, with a peak value of 238.41 bits/min (The demo video of online copy-spelling is enclosed and can be found at https://www.youtube.com/watch?v=EW2Q08oHSBo). Conclusion: We expand a BCI instruction set to over 100 command codes with high-speed in an efficient manner, which significantly improves the degree of freedom of BCIs. Significance: This study hold promise for broadening the applications of BCI systems.

Recent studies have shown that the feasibility of speech brain-computer interfaces (BCIs) as a clinically valid treatment in helping nontonal language patients with communication disorders restore their speech ability. However, tonal language speech BCI is challenging because additional precise control of laryngeal movements to produce lexical tones is required. Thus, the model should emphasize the features from the tonal-related cortex. Here, we designed a modularized multistream neural network that directly synthesizes tonal language speech from intracranial recordings. The network decoded lexical tones and base syllables independently via parallel streams of neural network modules inspired by neuroscience findings. The speech was synthesized by combining tonal syllable labels with nondiscriminant speech neural activity. Compared to commonly used baseline models, our proposed models achieved higher performance with modest training data and computational costs. These findings raise a potential strategy for approaching tonal language speech restoration.

Objective. Deep learning (DL) models have been proven to be effective in decoding motor imagery (MI) signals in Electroencephalogram (EEG) data. However, DL models’ success relies heavily on large amounts of training data, whereas EEG data collection is laborious and time-consuming. Recently, cross-dataset transfer learning has emerged as a promising approach to meet the data requirements of DL models. Nevertheless, transferring knowledge across datasets involving different MI tasks remains a significant challenge in cross-dataset transfer learning, limiting the full utilization of valuable data resources. Approach: This study proposes a pre-training-based cross-dataset transfer learning method inspired by Hard Parameter Sharing in multi-task learning. Different datasets with distinct MI paradigms are considered as different tasks, classified with shared feature extraction layers and individual task-specific layers to allow cross-dataset classification with one unified model. Then, Pre-training and fine-tuning are employed to transfer knowledge across datasets. We also designed four fine-tuning schemes and conducted extensive experiments on them. Main results: The results showed that compared to models without pre-training, models with pre-training achieved a maximum increase in accuracy of 7.76%. Moreover, when limited training data were available, the pre-training method significantly improved DL model’s accuracy by 27.34% at most. The experiments also revealed that pre-trained models exhibit faster convergence and remarkable robustness. The training time per subject could be reduced by up to 102.83 s, and the variance of classification accuracy decreased by 75.22% at best. Significance: This study represents the first comprehensive investigation of the cross-dataset transfer learning method between two datasets with different MI tasks. The proposed pre-training method requires only minimal fine-tuning data when applying DL models to new MI paradigms, making MI-Brain–computer interface more practical and user-friendly.

Objective: Currently, ensemble task-related component analysis (eTRCA) and task discriminative component analysis (TDCA) are the state-of-the-art algorithms for steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs). However, training the BCIs requires multiple calibration trials. With insufficient calibration data, the accuracy of the BCI will degrade, or even become invalid with only one calibration trial. However, collecting a large amount of electroencephalography (EEG) data for calibration is a time-consuming and laborious process, which hinders the practical use of eTRCA and TDCA. Methods: This study proposed a novel method, namely Source Aliasing Matrix Estimation (SAME), to augment the calibration data for SSVEP-BCIs. SAME could generate artificial EEG trials with the featured SSVEPs. Its effectiveness was evaluated using two public datasets (i.e., Benchmark, BETA). Results: When combined with SAME, both eTRCA and TDCA had significantly improved performance with a limited number of calibration data. Specifically, SAME increased the average accuracy of eTRCA and TDCA by about 12% and 3%, respectively, with as few as two calibration trials. Notably, SAME enabled eTRCA and TDCA to work well with a single calibration trial, achieving an average accuracy >90% for the Benchmark dataset and >70% for the BETA dataset with 1-second EEG. Conclusion: SAME is an effective method for SSVEP-BCIs to augment the calibration data, thereby significantly enhancing the performance of eTRCA and TDCA. Significance: We propose a new data-augmentation method that is compatible with the state-of-the-art algorithms of SSVEP-based BCIs. It can significantly reduce the efforts required to calibrate SSVEP-BCIs, which is promising for the development of practical BCIs.