Piezoelectric nanogenerators (PENGs) are electromechanical devices that convert physiological motion and tissue deformation into electrical signals, offering promising applications in self-powered neural interventions. This comprehensive review systematizes current research across mechanical inputs, piezoelectric transduction, neural responses, and functional restoration, while examining their applications in peripheral, central, and autonomic nervous systems, as well as neurodegenerative diseases and neural interfaces. The technology builds on foundational work by Wang et al. in 2006, which demonstrated ZnO nanowire arrays converting mechanical deformations into measurable electrical signals. The review evaluates neural-interface-compatible materials, device architectures, and powering strategies while addressing critical barriers including deep-tissue targeting precision, long-term implant safety, and standardized clinical translation frameworks. Key contributions include systematic comparison of PENGs with competing self-powered and wirelessly powered technologies, proposal of standardized reporting frameworks linking device output to biologically effective dosage, and identification of future directions in multimodal energy coupling, closed-loop feedback systems, and integration with genetic and metabolic reprogramming approaches. This work aims to bridge fundamental research and clinical translation in neural engineering applications.

Researchers have developed an open-source reinforcement learning framework for real-time, closed-loop optimization of electrical stimulation in engineered neuronal networks cultured on microelectrode arrays. The platform uses single-spike-resolution neural activity as continuous feedback to identify stimulation patterns that drive specific activity motifs in structurally constrained four-node circuits. Built on inkube, a low-cost electrophysiology system, the framework delivers stimulation with single-sample precision and millisecond-range latency, enabling multiple independent reinforcement learning agents to control different networks simultaneously. This addresses a critical challenge in neuroscience: the prohibitively large space of possible spatiotemporal stimulation patterns and the complexity of history-dependent evoked responses. While the study demonstrates solid methodological rigor and provides a reproducible, accessible tool for adaptive interaction with biological neural networks, further validation is required to determine how effectively these strategies generalize to larger, unstructured, or more conventional network architectures. The work has valuable implications for understanding neural circuit information processing and encoding through systematic perturbation under controlled conditions.

Brain-Computer Interfaces represent transformative neurotechnology enabling direct communication between the brain and external devices through EEG, ECoG, or intracortical arrays, offering substantial benefits for individuals with motor impairments, spinal cord injuries, or locked-in syndrome. While recent advances in high-resolution neural recording and signal processing have been significant, the field faces a critical challenge: transitioning from laboratory demonstrations to robust real-world applications. The primary obstacle is the "translation gap"—maintaining consistent performance across different users, mental states, and extended timeframes. Current systems require lengthy daily calibration and struggle with signal drift, neural plasticity, and complex co-adaptation dynamics between users and algorithms. This Research Topic addresses these computational challenges through innovative frameworks combining machine learning, adaptive filtering, and neural manifold alignment to develop accurate, adaptive "plug-and-play" interfaces. By integrating control theory and computational neuroscience insights, the initiative seeks to improve motor intent decoding and sensory feedback encoding, ultimately enabling high-bandwidth, bidirectional brain-machine communication. The collection emphasizes computational and mathematical foundations necessary for advancing BCI reliability and performance across diverse applications.

Researchers at China's National Center for Nanoscience and Technology and the Institute of Neuroscience have developed neurotassel elastocapillary filament arrays, a novel neural recording technology described in a March 2019 Science Advances publication. The innovation features flexible filaments that bundle into a rigid, insertable structure when withdrawn from molten polyethylene glycol at 120°C. The device incorporates chromium and gold conductors insulated between polyimide layers, with exposed recording sites at the filament tips. Platinum electrodeposition significantly reduces electrical impedance from over 1 megaohm to approximately 50 kiloohms at 1 kHz. Once implanted, body fluids dissolve the PEG coating, allowing the filaments to return to their flexible state for minimal tissue damage. Testing demonstrated over 95% electrical integrity post-implantation. The technology supports configurations ranging from 16 to 1,024 channels, using flip-chip-bonded flexible circuits and Omnetics connectors. Fabrication yields exceeded 80% with consistent impedance values of 54 ± 15 kiloohms after platinum treatment. This advancement represents a significant contribution to high-density neural interfacing, offering improved biocompatibility through its elastocapillary mechanism.

Neural recording ASICs are the integrated circuits within headstages and data acquisition systems that amplify, filter, and digitize electrode signals before data leaves the implant site. NeuroNexus provides a comprehensive guide comparing four major DAQ chip families used in neuroscience research. The Intan RHD family serves as the primary extracellular recording solution, with the 32-channel RHD2132 and 64-channel RHD2164 variants enabling acute and chronic recordings in rodents and large animals. The RHS2116 adds closed-loop stimulation capabilities by allowing independent recording or constant-current pulse delivery on each channel, making it ideal for intracortical microstimulation and deep brain stimulation research. Neuropixels probes integrate recording ASICs directly into probe shanks, enabling hundreds of recording sites across cortical layers in single penetrations. Neuropixels 2.0 offers improved ADC resolution and lower noise compared to version 1.0. The SiNAPS-1 custom ASIC enables thousands of recording sites through on-shank amplification and digitization. Selecting the appropriate chip depends on specific experimental requirements regarding channel density, stimulation needs, and recording scale.

Researchers have developed an innovative MRI-compatible three-dimensional soft carbon neural interface utilizing a depth-tunable laser drawing technique. This advancement addresses a critical limitation in current brain-computer interface technology: the incompatibility of traditional metallic neural electrodes with magnetic resonance imaging systems. The soft carbon construction eliminates ferromagnetic artifacts that typically interfere with MRI scanning, while the versatile laser drawing method enables precise fabrication of complex three-dimensional electrode geometries at controlled depths within biocompatible materials. This breakthrough is significant for clinical applications, as it allows patients with implanted neural interfaces to undergo MRI diagnostic procedures without device malfunction or safety risks. The development bridges neurotechnology and medical imaging, potentially expanding therapeutic possibilities for brain disorders and neurological conditions. By enabling simultaneous neural recording and high-resolution brain imaging, this technology represents a substantial advancement toward more comprehensive neurological diagnostics and treatment monitoring for patients requiring both implantable neural devices and periodic imaging assessments.

Paradromics Inc. announced in October 2026 that its brain-computer interface has achieved over four years of stable neural recording and decoding performance, based on studies of sheep implanted with its cortical modules—the electrode technology underlying the Connexus® BCI system. The milestone demonstrates that the platform maintains spike signal-to-noise ratio stability years after implantation, with no electrode-array migration or animal health complications observed. In a separate ongoing study of the complete Connexus BCI platform in six sheep, spike SNR has remained stable for nearly two years. CEO Matt Angle emphasized that this achievement refutes the common industry assumption that BCIs must trade peak performance for long-term durability, noting that Paradromics' platform combines the highest information transfer rates of any BCI with sustained stability. The system was intentionally engineered for longevity using medical-grade materials, surgical techniques that prevent electrode migration, and microwires thinner than human hair to minimize biological impact. The Connexus BCI is designed for individual neuron-resolution recording at high data rates with full implantability and years of stable performance, positioning Paradromics to develop durable clinical therapies.

This conceptual paper proposes an innovative solution to limitations in current deep brain stimulation (DBS) therapy by introducing a battery-less, biomarker-driven closed-loop system. Traditional DBS devices require periodic surgical battery replacement and operate continuously regardless of the patient's actual neurological state. The proposed architecture combines two key technologies: respiration-driven energy harvesting and cerebrospinal fluid (CSF) microbiosensing. Mechanical energy generated from diaphragmatic and thoracic wall motion during breathing is converted to electrical energy through piezoelectric or triboelectric transduction, then stored in a micro-supercapacitor or biodegradable cell. A biodegradable, antibody-functionalized nanowire biosensor positioned at the dural venous sinus or spinal epidural space monitors disease-relevant CSF biomarkers including amyloid-β, tau, and neurofilament light chain, triggering stimulation only when pathological thresholds are exceeded. This approach addresses significant clinical challenges by eliminating battery replacement surgeries, reducing infection risks, and enabling truly personalized stimulation based on real-time neurobiological markers. The system represents a substantial advancement toward autonomous, intelligent neurotechnology for treating neurological disorders.

Researchers at Harvard University have demonstrated a breakthrough solution to two critical challenges in developing long-term brain-computer interfaces. The study, published in Nature Biomedical Engineering by Hao Shen, Siyuan Zhao, Arnau Marin-Llobet and colleagues in Jia Liu's group, addresses representational drift—the gradual reorganization of neural firing patterns—and the physical instability of conventional neural probes. Traditional silicon or metal electrodes provoke immune responses in brain tissue, causing signal degradation within weeks. The team developed tissue-like flexible mesh electrode arrays with mechanical properties similar to brain tissue itself, allowing the probes to move with the brain rather than against it. Implanted in mouse visual cortex and hippocampus, these arrays successfully tracked individual neurons for up to five months, confirmed through spike sorting and waveform analysis. The researchers also implemented a geometric transformation algorithm to keep decoders locked onto drifting neural codes. This dual approach—combining biocompatible hardware with adaptive software—represents a significant advance toward reliable brain-computer interfaces that maintain functionality over extended periods, potentially transforming clinical applications for patients with neurological conditions or paralysis.

# Summary This research paper presents a comprehensive framework for developing whole-brain interfaces leveraging high-density recording and stimulation technologies. The work addresses the fundamental challenge of creating neural interfaces capable of simultaneously monitoring and modulating activity across multiple brain regions at unprecedented scale. The study discusses advanced electrode array designs and signal processing methodologies that enable simultaneous recording from thousands of neural sites, moving beyond traditional single-region approaches. Key technical contributions include novel multiplexing strategies, improved noise reduction algorithms, and integrated stimulation capabilities that maintain recording fidelity. The research demonstrates the feasibility of acquiring meaningful neural data across distributed brain networks while maintaining spatial and temporal resolution necessary for decoding complex neural computations. This work matters significantly for neuroscience and medical applications, as whole-brain interfaces could enable more sophisticated brain-computer interfaces for treating neurological disorders, advancing fundamental brain research, and ultimately supporting therapeutic interventions requiring comprehensive neural monitoring and control across multiple interconnected regions.

Researchers at the University of Massachusetts Amherst say they’ve found a way to integrate electronics with human cells in the lab and harvest the natural electricity generated by moving tissue. Using cardiac cells grown around a thin, flexible, ceramic-and-polymer mesh, the developers described the approach as the start of a reliable source of power that could do away with batteries.

Researchers at the University of Pennsylvania have developed an AI-based interface enabling real-time bidirectional communication between patients and epilepsy monitoring devices, as reported in Nature Medicine. In a clinical study, patients used a smartphone application to query their physiological data from electroencephalography (EEG) devices, which record brain activity indicative of seizures. The system learned from patient feedback to improve seizure detection algorithms, reducing false alarms by approximately 80%. Traditionally, EEG data is reviewed by experts only daily or multiple times weekly, limiting patient access to insights about their condition. This innovation shifts paradigm by placing patients in an active role in managing their epilepsy. The bidirectional learning model harnesses patients' real-time knowledge of their symptoms and experiences to correct device misclassifications of benign brain activity, while patients gain immediate access to personalized health information previously unavailable to them. According to Amy Bany Adams, acting director of the NIH's National Institute of Neurological Disorders and Stroke, this proof of concept demonstrates potential to transform epilepsy management and possibly other diseases toward more adaptive, patient-centered approaches.

Closed-loop neuromodulation represents a significant advancement in treating neurological and neuropsychiatric disorders including Parkinson's disease, epilepsy, treatment-resistant depression, and obsessive-compulsive disorder. Unlike conventional open-loop systems that deliver stimulation on fixed schedules regardless of patient status, closed-loop approaches continuously monitor neural biomarkers such as local field potentials and oscillatory activity, then adjust stimulation parameters in real time based on the patient's fluctuating brain state and symptoms. This adaptive methodology enhances therapeutic efficacy while reducing side effects from inappropriate stimulation levels. Established devices like the NeuroPace RNS system for epilepsy and emerging adaptive deep brain stimulation platforms demonstrate clinical promise, though they raise critical questions regarding biomarker identification, algorithm design, device engineering, and ethical implications of continuous brain monitoring. A comprehensive research initiative seeks to consolidate current knowledge on closed-loop system design, validation, and clinical implementation across diverse conditions. Priority areas include neural biomarker validation, real-time feedback algorithms, clinical trial outcomes, comparative open-loop versus closed-loop studies, machine learning applications, hardware innovations in brain-computer interfaces, and regulatory and ethical considerations governing continuous neural monitoring and adaptive stimulation technologies.

Researchers have developed nanomaterial-enabled flexible sensors capable of continuous biomedical monitoring and closed-loop intervention, representing a significant advancement in wearable medical technology. These sensors leverage nanotechnology to create devices that conform to the body's contours while maintaining robust sensing capabilities. The flexible sensor platform enables real-time tracking of physiological parameters through its integration of nanomaterials that enhance sensitivity and responsiveness. A key innovation involves the sensors' ability to facilitate closed-loop intervention, meaning they can not only monitor health metrics but also trigger automated therapeutic responses based on detected conditions. This dual functionality—continuous monitoring combined with immediate intervention capability—addresses a critical gap in current biomedical devices. The technology has substantial implications for personalized medicine, chronic disease management, and emergency response, potentially reducing hospitalization rates and improving patient outcomes. Published in September 2026 as an open-access review, this work synthesizes current research in flexible bioelectronics, offering healthcare professionals and biomedical engineers a comprehensive resource for understanding how nanomaterial integration enhances wearable sensor performance and enables next-generation patient care solutions.

Brain-Computer Interface (BCI) technology enables users to operate external devices without physical movement. Electroencephalography (EEG) based BCI systems are being actively studied due to their high temporal resolution, convenient usage, and portability. However, fewer studies have been conducted to investigate the impact of high spatial resolution of EEG on decoding precise body motions, such as finger movements, which are essential in activities of daily living. Low spatial sensor resolution, as found in common EEG systems, can be improved by omitting the conventional standard of EEG electrode distribution (the international 10–20 system) and ordinary mounting structures (e.g., flexible caps). In this study, we used newly proposed flexible electrode grids attached directly to the scalp, which provided ultra-high-density EEG (uHD EEG). We explored the performance of the novel system by decoding individual finger movements using a total of 256 channels distributed over the contralateral sensorimotor cortex. Dense distribution and small-sized electrodes result in an inter-electrode distance of 8.6 mm (uHD EEG), while that of conventional EEG is 60 to 65 mm on average. Five healthy subjects participated in the experiment, performed single finger extensions according to a visual cue, and received avatar feedback. This study exploits mu (8–12 Hz) and beta (13–25 Hz) band power features for classification and topography plots. 3D ERD/S activation plots for each frequency band were generated using the MNI-152 template head. A linear support vector machine (SVM) was used for pairwise finger classification. The topography plots showed regular and focal post-cue activation, especially in subjects with optimal signal quality. The average classification accuracy over subjects

Researchers conducted a comprehensive network meta-analysis examining brain-computer interface (BCI) training effectiveness for post-stroke upper-limb motor recovery across different feedback modalities. The study analyzed 22 randomized controlled trials involving 956 patients at various recovery stages, comparing BCI interventions classified as multimodal feedback (BCI-MM), robotic or exoskeleton-assisted feedback (BCI-Robot), functional electrical stimulation feedback (BCI-FES), and visual feedback (BCI-visual) against usual rehabilitation. Using the Fugl-Meyer Assessment for Upper Extremity as the primary outcome, alongside secondary measures including the Action Research Arm Test and Wolf Motor Function Test, the analysis revealed that BCI-Robot demonstrated superior efficacy with a weighted mean difference of 6.28 points compared to usual rehabilitation, followed by non-BCI robotic training and BCI-MM. Notably, BCI-FES showed no significant difference from standard rehabilitation, while BCI-visual achieved the highest surface under the cumulative ranking curve value at 77.6 percent. These findings provide evidence-based guidance for selecting optimal BCI feedback modalities in post-stroke rehabilitation, with robotic-assisted approaches showing particular promise for motor recovery outcomes.

StairMed has developed the WRS, an ultra-flexible wireless brain-computer interface combining penetrating electrodes with wireless implant technology. The system exists in multiple generations: the WRS01/WRS64 represents the first-generation chronic implant with 64 channels, while the newer WRS02 offers 256 channels. Research published in Nature Communications in 2026 describes intraoperative arrays featuring 128 and 256 channels with 2-micrometer shank thickness, though these experimental versions may differ from the commercial implant. StairMed's product specifications indicate 1-micrometer electrode thickness. The WRS64 is registered through ClinicalTrials.gov for chronic implantation studies. This innovation addresses the critical challenge of wireless neural recording by integrating ultra-flexible electrode arrays that minimize tissue damage while maintaining high channel density for comprehensive brain activity monitoring. The wireless capability eliminates percutaneous connectors, reducing infection risk and improving patient mobility—significant advantages for long-term brain-computer interface applications in clinical and research settings.

Researchers in South Korea have developed an innovative wireless power transfer system for implantable bioelectronics, addressing a critical challenge that has hindered the deployment of battery-free medical devices. Led by Dae-Hyeong Kim at Seoul National University and collaborating institutions, the team published their findings in Nature Electronics, demonstrating a wearable-to-implant power link that maintains stable energy delivery despite misalignment, tissue stretching, and fluctuating electrical loads. The system employs a nonlinear parity-time symmetric circuit architecture, building on Stanford's 2017 approach, which enables the transmitter to continuously adjust its operating parameters. This dual-pronged solution combines circuit innovation with specialized materials to overcome the resonance frequency drift problem that has plagued conventional inductive coupling systems since the 1980s. The researchers successfully validated their technology by pacing large animal hearts without external connections, demonstrating practical clinical potential. This advancement represents a significant step toward realizing truly implantable, maintenance-free medical devices that can reliably deliver therapy or monitoring from within the body without wires or replacement surgeries.

Brain-computer interfaces (BCIs) establish direct communication pathways between the brain and external devices, offering promising rehabilitative applications for children with neurodevelopmental disorders including cerebral palsy, autism spectrum disorder, attention deficit hyperactivity disorder, and perinatal brain injury. This comprehensive review examines five major BCI paradigms—P300, motor imagery, steady-state visual evoked potentials, functional near-infrared spectroscopy, and multimodal hybrid systems—evaluating their pediatric adaptability and optimization strategies. The analysis covers four primary clinical applications: augmentative and alternative communication, motor function rehabilitation, ADHD attention training, and ASD social-emotional intervention. Key variables affecting BCI performance in children include cognitive and physiological factors, alongside device and environmental considerations. Preliminary evidence suggests non-invasive BCIs are feasible for children above five years old, though evidence remains limited by small sample sizes and short-term studies. Significant barriers to adoption include user experience, clinical workforce readiness, and ethical considerations specific to pediatric populations. While evidence is more established for augmentative and alternative communication and motor rehabilitation in cerebral palsy, substantial long-term data are lacking for ADHD and autism interventions. Future research must prioritize child-specific lightweight wearable devices and multicenter longitudinal trials to advance pediatric BCI implementation.

# Professional Summary Brain-computer interfaces utilizing electroencephalography (EEG) for motor imagery represent a significant advancement in neurotechnology applications. Motor imagery-based BCIs enable users to control external devices or systems by imagining limb movements while EEG sensors detect and interpret corresponding brain activity patterns. This technology integrates sophisticated signal processing algorithms with machine learning techniques to decode neural signals and translate intentional motor thoughts into actionable commands. The field encompasses multiple algorithmic approaches designed to enhance classification accuracy and reduce training time for end users. Key applications include assistive devices for individuals with motor disabilities, rehabilitation systems for stroke recovery, and prosthetic limb control. The development of robust EEG-based motor imagery BCIs addresses critical challenges in real-world implementation, including signal variability, noise reduction, and user adaptation. These systems demonstrate potential for restoring independence and improving quality of life for paralyzed or severely motor-impaired populations. Continued refinement of signal processing methodologies and algorithmic frameworks promises enhanced system reliability and broader clinical deployment across neurological rehabilitation and assistive technology sectors.