Motor imagery brain-computer interfaces (MI-BCIs) remain challenging for practical use due to their reliance on multi-channel EEG devices and long calibration. To address these limitations, we proposed a wearable system for calibration-free online decoding using a lightweight, few-channel EEG headband, enabling portability, ease of use, and rapid setup. Specifically, we first built a large-scale wearable MI-EEG dataset from 100 healthy subjects to train a subject-independent model. We then developed a CNN-based temporal convolutional network (CTCNet) for online MI detection, which reduced computational complexity while maintaining high decoding performance. Furthermore, we introduced a supervised self-training (SST) strategy that leverages labeled online data and progressively fine-tunes a pre-trained subject-independent model, enabling calibration-free BCI operation without offline calibration. Four online experiments were conducted, involving 25 healthy subjects (Experiments II-IV) and 10 stroke patients (Experiment V). With the SST strategy, the accuracy of the subject-independent model improved from 69 % initially to 81 % after the first update and further increased to 86% after the second update, surpassing the subject-specific model (80%). Stroke patients exhibited a similar improvement trend. Moreover, simulated experiments confirmed the superiority of the subject-independent model compared to training from scratch. These findings demonstrate the effectiveness of the wearable MI-BCI system based on SST and CTCNet for online MI detection and highlight its substantial potential for motor recovery in stroke patients.

Conventional P300-based brain–computer interfaces (BCIs) commonly rely on two-dimensional (2D) visual flashing, which may induce visual fatigue and limit immersion, thereby restricting long-term usability and system performance. To address these limitations, this study proposes an immersive P300-BCI framework integrating a three-dimensional morphological stimulation paradigm, termed 3D-Morph, with self-adaptive Bayesian linear discriminant analysis (SA-BLDA). Instead of using color or luminance flickering, the proposed paradigm employs dynamic 2D-to-3D morphological transformations of virtual objects in a virtual reality environment to enhance target-related event-related potentials while preserving visual immersion. SA-BLDA further adjusts the number of stimulation rounds according to classification confidence to balance accuracy and interaction efficiency. Experiments with 24 participants showed that the proposed system outperformed the conventional 2D paradigm. In offline analysis, the proposed method achieved an average classification accuracy of 94.17% and an information transfer rate (ITR) of 25.50 bits/min, significantly outperforming the 2D paradigm (87.29% accuracy, 22.75 bits/min ITR, both $p<0.001$, Cohen’s $d\ge 1.22$). In online experiments, the 3D-Morph paradigm achieved an average accuracy of 91.46% and an ITR of 37.23 bits/min, compared with 83.96% and 28.74 bits/min for the conventional 2D paradigm (both $p<0.01$, Cohen’s $d\ge 1.14$). The average response time was reduced by 0.46 s ($p<0.01$, Cohen’s $d=0.78$), and the processing time per stimulation round (PT) of SA-BLDA was significantly reduced from $48.54\pm 10.47$ ms in the 2D paradigm to $26.40\pm 9.41$ ms in the 3D-Morph paradigm ($p<0.01$, Cohen’s $d=2.34$), corresponding to a 45.61% reduction in computational time per round. NASA-TLX evaluations indicated a significantly lower subjective workload across all dimensions (all $p<0.05$, Cohen’s $d\ge 0.76$). These results demonstrate that combining 3D-Morph stimulation with SA-BLDA can significantly improve classification performance, interaction efficiency, and user experience, providing a feasible framework for immersive and practical P300-BCI applications.

Brain–computer interfaces (BCIs) enable direct communication between the brain and external devices, providing a bio-inspired link between humans and artificial systems. However, electroencephalography (EEG)-based motor imagery (MI) decoding continues to pose challenges, due to limited temporal exploitation and insufficient spatio-temporal modeling. To alleviate these limitations, a complementary dual-branch receptive field network (DBRFNet) is proposed for cross-channel-density MI-EEG decoding. Specifically, two complementary branches with distinct temporal receptive fields are designed to capture multi-scale temporal dynamics. Each branch employs a progressive spatio-temporal module integrating temporal, depthwise spatial, and dilated convolutions for modeling short- and long-range temporal dependencies. The features from the two branches are fused using element-wise addition and further encoded for classification. The proposed method is evaluated on three public datasets representing low-, medium-, and high-density EEG configurations. DBRFNet achieves decoding accuracies of 80.48%, 84.37%, and 92.24% under cross-validation analysis, and 87.76%, 80.94%, and 94.64% under hold-out analysis, demonstrating competitive performance compared with state-of-the-art methods. Ablation studies further validate the effectiveness of each component within the proposed architecture. These results indicate that DBRFNet effectively learns discriminative spatio-temporal representations and provides a robust solution for MI-BCI systems.