Research on invasive and non-invasive brain–computer interfaces (BCIs) is developing rapidly, especially in the field of medicine. This article reviews BCI software, hardware, related neural mechanisms, and clinical applications. The principles of invasive and non-invasive BCI differ; however, it is widely used in various scenarios. Based on different brain functional areas, a variety of cutting-edge clinical applications could be derived in future.

Small-world networks, with a high degree of clustering and a small characteristic path length, are the universal nature of the human brain network. This review summarizes trends in small-world networks for different neurological and psychiatric disorders, with the brain network topology analysis in treating and diagnosing mental and neurological disorders.

Peripheral nerve disease is a common disorder that can lead to sensory, motor, and autonomic dysfunction, with common symptoms including pain, loss of sensation, lack of coordination, muscle weakness, and atrophy. Defensins (which can be divided into mammal, insect, and plant defensins), as regulatory factors of the immune system, can influence peripheral nerve diseases by participating in inflammatory processes, immune responses, and pathogen resistance. As therapeutic targets, they have the potential to restore the quality of life of patients and reduce the healthcare burden.

This study proposes a model of consciousness based on information theory, positing contra the Integrated Information Theory (IIT) that perception arises from uncertainty minimization (negative entropy). Unlike the causal information definition of IIT, this model adopts the framework of standard information theory, emphasizing the role of quantum coherence in consciousness and offering a new perspective for understanding the computational mechanisms of consciousness.

This review explores brain–computer interfaces (BCIs) to address amyotrophic lateral sclerosis (ALS) communication barriers. Recent advances include implantable electrocorticographic systems that enable high-accuracy (90%–99%) speech synthesis via adaptive algorithms (convolutional neural and long short-term memory networks), and non-invasive electroencephalography integrated with eye-tracking/functional near-infrared spectroscopy. By enhancing accessibility and reliability, BCIs restore independence and bridge communication gaps, transforming assistive communication technologies for ALS.

Large language models, including DeepSeek and ChatGPT, have the potential to significantly advance brain-computer interfaces and brain-inspired computing by enhancing the accuracy of brain signal decoding and optimizing user interaction. In brain-computer interfaces, these models facilitate more precise and responsive communication, while in brain-inspired computing, they enable realistic simulation of neural networks and improved hardware energy efficiency. However, substantial challenges remain, particularly in healthcare applications and other broader fields.

This study investigates the transcriptional profiles of gliomas across different grades (WHO II-IV) and clinical states (primary vs. recurrent). Utilizing RNA-seq data from public databases (e.g., GEO), we analyzed low-grade gliomas and high-grade gliomas, including oligodendrogliomas, glioblastomas, and other glioma subtypes. Key analyses encompassed differential gene expression, glioma subpopulation characterization (e.g., glioma-associated microglia/macrophages), regulatory network construction (WGCNA and transcription factor activity), and cell state analysis comparing primary and recurrent gliomas. Our findings reveal distinct transcriptional signatures and identify potential biomarkers associated with glioma progression and recurrence.

This review summarizes recent advances in the fundamental materials, functional designs, and application schemes of hydrogels for peripheral nerve repair. The hydrogels discussed here can be used to repair peripheral nerve injury, including filling the damaged area, transmitting electrical stimulation signals, and delivering therapeutic ingredients. The advances summarized in this review can guide the future direction of this biomaterial and provide a foundation for next-generation functionalized hydrogels.

A steady-state visually evoked potential (SSVEP) is a response to visual stimuli frequencies, but brain–computer interfaces (BCIs) struggle to distinguish it from spontaneous signals. This study proposes the variational mode decomposition–based filter bank canonical correlation analysis (VMD-FBCCA) algorithm to improve recognition accuracy by addressing overlapping frequency coverage between evoked and spontaneous electroencephalography signals. The algorithm has great prospects for developing improved BCI systems and offers a novel perspective for SSVEP-BCI research. Future research should enhance the robustness of the VMD-FBCCA algorithm and apply it to the development of SSVEP-BCI systems.