Advancing design and fabrication of high-density neural interfaces for long-term neural recordings

Abstract

Neural interfaces are critical tools for studying and modulating activity in the central and peripheral nervous systems for both scientific and clinical applications. Neural signals span a vast range of spatial and temporal scales, from micron-scale subcellular structures (e.g., dendrites, axons, and spines) to whole-brain networks spanning tens of centimeters. These signals also vary in frequency, from extracellular action potentials (~1000 Hz) to local field potentials with frequencies as low as 1–3 Hz. Effective investigation of neural activity requires high-resolution sampling in both space and time. While research has advanced high-channel-count electrode arrays and recording systems, several engineering challenges remain regards to their clinical translation. Fabricating high-density electrode arrays requires novel materials and techniques that must meet strict biocompatibility and reliability standards. Long-term implantable systems demand fully integrated electronics to minimize infection risks, necessitating new strategies to protect recording and stimulation circuits while enabling high-density feedthroughs. Current packaging methods, such as bulky ceramic and titanium enclosures, limit device miniaturization. One critical clinical application of neural interfaces is the treatment of refractory epilepsy. Accurate seizure onset zone mapping, essential for surgical planning, relies on electrocorticography (ECoG) or stereoencephalography (sEEG) recordings from numerous brain locations. Enhancing spatial resolution and long-term implantation capabilities could improve patient outcomes but requires overcoming significant engineering barriers. In my dissertation, I address these challenges by developing novel design, fabrication, and packaging techniques for microelectrode arrays to increase density and enable long-term recordings. My work builds on prior research exploring liquid crystal polymer (LCP) as a thin-film substrate for neural interfaces. LCP is a promising material due to its compatibility with lithographic fabrication, low water absorption, and thermoplastic properties that enable monolithic encapsulation of embedded electronics. I designed two high-density neural interfaces: a micro-electrocorticographic (µECoG) electrode array for cortical surface recordings and a micro-stereoelectroencephalographic (µSEEG) electrode array for deep brain recordings. Both devices utilize an LCP substrate and feature significantly higher channel counts and spatial densities than current clinical versions. For the micro-sEEG device, I transitioned from outsourcing LCP electrode fabrication to developing an in-house process at the Duke University cleanroom. The fabrication process required overcoming several challenges, including metal adhesion to LCP, creation of vertical access interconnects (VIA) in the 25-micron LCP sheet, and dealing with the opaqueness of the material for mask alignment between layers. I developed a two-metal-layer photolithographic procedure using thin-film LCP sheets, which were then rolled into 1 mm diameter cylindrical shanks with high-density contacts arranged in four columns along the outer walls. I validated these devices in both a rat model and intraoperatively in patients undergoing epilepsy surgery, demonstrating the advantages of circumferential micro-contacts over conventional ring contacts. For the micro-ECoG electrode, I collaborated with Dr. Shepard at Columbia University to integrate an application-specific integrated circuit (ASIC) with wireless recording capability into the LCP substrate. I worked closely with the fabrication company, Dyconex AG., and the Shepard Lab to optimize the device's electrical layout and match design requirements. To address fabrication constraints, I developed an automated wiring algorithm that maximizes the number of routable pads for dense chip footprints designed for flip-chip bonding. Additionally, I worked with Duke Neurosurgery clinicians to refine the device’s form factor for surgical compatibility, reducing size of the cables through folding to allow for suturing of the dura during the surgical closing procedure. Finally, I explored reinforced silicone as an alternative encapsulation material to LCP. In collaboration with MicroLeads Inc., I developed a micro-ECoG device with integrated electronics for electrode multiplexing. By implementing a 6:1 multiplexer, we reduced the number of output wires in the high-density grid, minimized the implantation footprint and lowered the infection risk. I evaluated the device in chronic rat studies, recording auditory cortex responses over 30 days to assess long-term reliability. The results of this dissertation advance methodologies for designing and constructing high-channel neural interface systems, enabling recordings with greater spatial coverage, higher resolution, and extended implantation durations. In particular, testing LCP as a substrate for neural interfacing demonstrates the material’s potential for high-density, biocompatible neural interfaces, providing a scalable platform for future device development. The integration of wireless recording capabilities in LCP-based devices represents a critical step toward clinical translation, reducing the need for transcutaneous connectors and improving long-term implantation viability. Future work will build on these approaches to develop next-generation devices with enhanced functionality for clinical and research applications.