Multimodal Ultra-Flexible Neural Electrodes with Bioelectromechanical Compatibility for Brain-Computer Interface

Abstract

Brain-computer Interface (BCI) is a core technology that is used to connect the brain and external devices. Its performance relies on the interface compatibility between neural electrodes and brain tissue. Currently, implanted neural electrodes face two major bottlenecks: 1. Mechanical mismatch: The stiffness of traditional rigid electrodes (such as quartz optical fibers and silicon substrates) is 6-7 orders of magnitude higher than that of brain tissue. Even electrodes with better flexibility have a stiffness 3-4 orders of magnitude higher than brain tissue. Long-term implantation easily triggers glial scar hyperplasia, leading to a significant reduction in the signal-to-noise ratio (SNR) of neural signals. 2. Single function: Most electrodes can only realize neural electrical signal recording and cannot synchronously detect brain tissue micromotion (micrometer-scale displacements induced by physiological states). However, brain tissue micromotion cause relative positional deviation between electrodes and neurons, further interfering with signal quality. In addition, as a tool for precisely regulating neural activity, optogenetic technology is still limited by electrode substrate materials in its integration with neural electrodes (forming optogenetic electrodes), lacking an ultra-flexible design compatible with brain tissue. To address the above issues, this study takes the bioelectromechanical compatibility at the interface between ultra-flexible neural electrodes and brain tissue as the core design concept. It first focuses on the key technology for brain tissue micromotion detection-the piezoresistive effect. Since the microscopic mechanism of the piezoresistive effect in current flexible neural electrodes is still unclear (e.g., how strain is accurately converted into electrical signals and the state of environmental stability), directly proceeding with the electrode design may lead to unreliable functionality. Therefore, it is necessary to first reveal its essence through material-level preliminary research. Then, based on this mechanism, a practical ultra-flexible neural electrode is developed. Finally, the optogenetic function is expanded. On the whole, a complete research chain covering mechanism, design, fabrication, and verification is established, aiming to provide hardware support and theoretical basis for next-generation high-performance BCIs.

To reveal the microscopic essence of the piezoresistive effect in flexible neural electrodes, an in-situ opto-electromechanical characterization platform under the transmission electron microscope (TEM) was established. Using molybdenum disulfide (M_OS₂) nanoflakes as the functional layer and silicon nitride (Si₃N₄) as the substrate, ultra-flexible micro-cantilever chips (with a thickness of 100-180 nm and a bending angle＞45°) were fabricated to systematically investigate their piezoresistive mechanism and performance. Through high resolution TEM (HRTEM) and energy band analysis, it was confirmed that the piezoresistive effect of M_OS₂ originates from strain-modulated energy band structure: Under tensile strain, the bandgap of M_OS₂ narrows from 1.3 eV to 1.1 eV, and the carrier concentration increases by approximately 2 orders of magnitude. Under compressive strain, nano-wrinkles form in the channel, leading to enhanced carrier scattering. Meanwhile, the quantitative relationship between strain and electrical signals was revealed: the piezoresistive sensitivity reaches 225.8 pA/nm under tensile strain and 114.3 pA/nm under compressive strain, providing a theoretical basis and performance reference for the piezoresistive function of electrodes. Its core value lies in verifying the physical essence that the piezoresistive effect can quantitatively detect strain. However, its all-dry transfer process makes it difficult to mass-produce multi-channel electrodes. Additionally, the piezoresistive mechanism of M_OS₂ (energy band modulation) is extremely sensitive to material purity and number of layers (the piezoresistive effect of multi-layer M_OS₂ is weakened). In contrast, the piezoresistive mechanism of gold (geometric deformation (bending induces changes in cross-sectional area, which in turn leads to changes in resistance) is more stable, unaffected by factors such as brain tissue fluid and temperature fluctuations, making it more suitable for long-term implantation. Therefore, in subsequent work, materials such as polyimide (biocompatible) and gold (chemically stable) should be selected to transform the piezoresistive mechanism into implantable electrode functions.

Based on the core design principle of "linear correlation between strain and electrical signals" established in the preliminary research on M_OS₂, and to achieve perfect mechanical compatibility between the flexible neural electrode and brain tissue, this study proposes a new concept of ultra-flexible neural electrodes with bioelectromechanical compatibility. Taking zero relative displacement at the interface between the neural electrode and brain tissue as the criterion, an electrode-brain tissue model was established, and finite element simulation analysis was conducted on the electrode-brain tissue interface. By integrating electrode materials and electrode geometric properties, the three-dimensional parameters of the electrode were determined (40 mm in length, 164 μm in width, and 3 μm in thickness), with its equivalent stiffness reduced to 0.023 N/m. Using micro-fabrication technology, an ultra-flexible neural electrode with a 16-channel array was prepared (each electrode contains 4 recording points, with each recording point sized 14 μm × 24 μm and the distance between adjacent electrode recording point being 500 μm). Combined with a nanorobotic manipulator under a scanning electron microscope (SEM), the mechanical properties and piezoresistive properties of the electrode were accurately characterized under simulated brain tissue micromotion (the interfacial force of 575 nN and piezoresistive sensitivities of 6.4 pA/mm (length) and 10.2 pA/µm (displacement)). In addition, through electrophysiological characterization of mouse brain tissue, local field potential (LFP) signals of mice were successfully collected. The experiment showed that the intensity of the LFP signals was effectively enhanced at a specific frequency of 2.72 Hz, which corresponds to the mouse's respiratory state. This confirms bioelectromechanical compatibility of the ultra-flexible neural electrode, ultimately realizing the integration of three functions of the electrode (mechanical compatibility, brain tissue micromotion detection and electrophysiological recording).

Based on the research on ultra-flexible neural electrodes with bioelectromechanical compatibility, a blue-light LED was integrated into the neural electrode substrate, and an ultra-flexible optogenetic neural electrode with the dual functions of optical stimulation and electrophysiological recording was developed. While enabling the electrode to deliver targeted blue-light stimulation and record electrophysiological signals, this design also guarantees the mechanical compatibility between the electrode and brain tissue. Through in vivo electrophysiological experiments on the mouse, the blue-light stimulation was applied to hippocampal neurons. Synchronized recording showed that a significant increase in the intensity of local field potential (LFP) signals was detected. This verifies the synergy of the electrode’s triple functions (mechanical compatibility, optical stimulation and electrophysiological recording).

This study has developed an innovative framework encompassing material strain mechanism, device flexible design, biological interface compatibility and multifunctional integration. Furthermore, this framework provides theoretical, technical, and application-oriented support for next-generation implantable neural electrodes, and is anticipated to drive the advancement of BCIs toward an adaptive closed-loop paradigm.

Date of Award	15 Apr 2026
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Lixin DONG (Supervisor)