Latest Publications

Robo-pigeons, a novel class of hybrid robotic systems developed using brain-computer interface technology, hold marked promise for search and rescue missions due to their superior load-bearing capacity and sustained flight performance. However, current research remains largely confined to laboratory environments, and precise control of their flight behavior, especially flight altitude regulation, in a large-scale spatial range outdoors continues to pose a challenge. Herein, we focus on…

There has been incremental progress in moving BCI out of the laboratory environment and into the homes of those who would benefit most, especially children living with severe physical disabilities. Practical issues, such as available computational resources and long calibration times, have slowed down the adoption of such systems. To develop an efficient and scalable machine learning framework consistent with early approaches that facilitate at-home BCI use, this study provides valuable insights…

State-of-the-art intracortical neuroprostheses currently enable communication at 60+ words per minute for anarthric individuals by training on over 10K sentences to account for phoneme variability in different word contexts. There is limited understanding about whether this performance can be maintained in decoding naturalistic speech with 40K+ word vocabularies across elicited, spontaneous, and conversational speech contexts. We introduce a vocal-unit-level generalization test to explicitly…

The existing Riemannian geometry-based approaches for brain computer interface (BCI) employ fixed time windows. However, the inherent variability and dynamic changes among subjects necessitate robust and adaptive solutions for time window optimization. Recognizing the current limitations of Riemannian classifiers, we propose a time window selection confidence metric (TWSCM) based on Riemannian geometry. This metric operates on the manifold of symmetric positive definite (SPD) matrices, providing…

The brain-computer interface (BCI) based on rapid serial visual presentation (RSVP) provides a novel approach for efficiently optimizing traditional machine-based target detection, revealing a broad application prospect in security, entrainment, monitoring, etc. A bottleneck of current RSVP-BCI is that its detectable result is limited to a binary way, i.e., target vs. non-target, more detailed and important information about targets, such as the precise position, remains undetectable. To solve…

Automated emotion recognition using electroencephalogram (EEG) signals has gained substantial attention. Although deep learning approaches exhibit strong performance, they often suffer from vulnerabilities to various perturbations, like environmental noise and adversarial attacks. In this paper, we propose an Inception feature generator and two-sided perturbation (INC-TSP) approach to enhance emotion recognition in brain-computer interfaces. INC-TSP integrates the Inception module for EEG data…

ObjectiveElectroencephalogram (EEG) based brain-computer interfaces (BCIs) have shown tremendous promise in facilitating direct non-invasive brain-control over external devices. However, their practical application is hampered due to errors in command interpretation. A promising strategy for improving BCI accuracy is based on detecting error-related potentials (ErrPs), which are EEG potentials evoked in response to errors. Thus, performance can be improved by undoing actions that…

The supernumerary robotic finger (SRF) can expand human hand abilities to achieve motor augmentation, and integrate with brain computer interface (BCI) to free the occupation of inherent body degrees of freedom. However, the neuro remodeling mechanisms of brain-actuated SRF training is not clear. In this study, a BCI-actuated SRF was used to investigate the concurrent changes in behavior and brain activity. After 4 weeks BCI-SRF training, the novel sequence operation accuracy rate enhanced by…

This paper introduces a novel decoding approach for Brain Machine Interface (BMI) that enhances the estimation accuracy and stability of muscle activity by incorporating domain knowledge of motor control. Our approach uniquely integrates domain knowledge, focusing on the relationship between torque direction and muscle activity in isometric wrist tasks. We demonstrate the effectiveness of our approach through decoding analysis with non-human primates performing a wrist torque tracking task. By…

After prolonged paralysis, paraplegic spinal cord injury (SCI) patients typically lose the ability to generate the expected electroencephalogram (EEG) α/β modulation associated with leg movements. Brain computer interface (BCI)-controlled ambulation devices have emerged as a way to restore brain-controlled walking, but this loss of EEG signal modulation may impede the ability to operate such systems and prolonged training may be necessary to restore this physiologic phenomenon. To address this…

The incorporation of artificial intelligence (AI) into neuroscience has the potential to significantly enhance our comprehension of brain function and facilitate more effective diagnosis and treatment of neurological disorders. Artificial intelligence (AI) techniques, particularly deep learning and machine learning, offer transformative solutions by improving the analysis of complex neural data, facilitating early diagnosis, and enabling personalized treatment approaches. A bibliometric analysis…

In 2024 and early 2025, several successful surgeries involving brain-computer interfaces (BCIs) gained media attention, including those conducted by Elon Musk's company Neuralink, which implanted BCIs in three paralyzed volunteers, allowing them to control computers through thought alone. While the concept of merging humans with machines dates back to the 1960s, BCI technology has now entered the clinical trial stage, with a focus on restoring communication, mobility, and sensation in…

The fluidic memristor has attracted growing attention as a promising candidate for neuromorphic computing and brain-computer interfaces. However, a fluidic memristor with ion selectivity as that of natural ion channels remains a key challenge. Herein, inspired by the structure of natural biomembranes, we developed an ion-shuttling memristor (ISM) by utilizing organic solvents and artificial carriers to emulate ion channels embedded in biomembranes, which exhibited both neuromorphic functions and…

Decoding algorithms used in invasive brain-computer interfaces (iBCIs) typically convert neural activity into continuously varying velocity commands. We hypothesized that putting constraints on which decoded velocity commands are permissible could improve user performance. To test this hypothesis, we designed the discrete direction selection (DDS) decoder, which uses neural activity to select among a small menu of preset cursor velocities. Approach. We tested DDS in a closed-loop…

The study objective was to characterise indices of learning and patterns of connectivity in two neurofeedback (NF) paradigms that modulate mu oscillations in opposite directions, and the relationship with change in excitability of the corticospinal tract (CST). Approach: Forty-three healthy volunteers participated in 3 NF sessions for upregulation (N=24) or downregulation (N=19) of individual alpha (IA) power at central location Cz. Brain signatures from multichannel…

Adding supernumerary robotic limbs (SRLs) to humans and controlling them directly through the brain are main goals for movement augmentation. However, whether neural patterns that are distinct from the traditional inherent limbs motor imagery (MI) paradigm can be extracted, which is essential for the high-dimensional control of external equipment. In this study, a novel type of MI paradigm based on SRLs was proposed, consisting of "the sixth-finger", "the third-arm" and "the third-leg", and…

Brain-computer interfaces (BCIs) suffer from limited accuracy due to noisy electroencephalography (EEG) signals. Existing denoising methods often remove artifacts such as eye movement or use techniques such as linear detrending, which inadvertently discard crucial task-relevant information. To address this issue, we present BGNet, a novel deep learning framework that leverages underutilized baseline EEG signals for dynamic noise mitigation and robust feature extraction to improve motor imagery…

Due to inter-subject variability in electroencephalogram (EEG) signals, the generalization ability of many existing brain-computer interface (BCI) models is significantly limited. Although transfer learning (TL) offers a temporary solution, in scenarios requiring sustained knowledge transfer, the performance of TL-based models gradually declines as the number of transfers increases-a phenomenon known as catastrophic forgetting. To address this issue, we introduce a novel domain-incremental…

In this study, a novel open-source brain-computer interface (BCI) platform was developed to decode scalp electroencephalography (EEG) signals associated with sustained attention. The EEG signal collection was conducted using a wireless headset during a sustained visual attention task, where participants were instructed to discriminate between composite images superimposed with scenes and faces, responding only to the relevant subcategory while ignoring the irrelevant ones. Seven volunteers…

Electroencephalography (EEG) is crucial for monitoring brain activity in neuroscience and clinical applications. However, the multitude of channels recorded by scalp electrodes poses challenges, including impractical usage and high model complexity. This paper addresses the challenges of high dimensionality in EEG data and introduces an innovative EEG channel selection algorithm, LSvT-NI, based on model training and noise injection, achieving substantial reductions in channels, model size, and…