A silk-based self-adaptive flexible opto-electro neural probe

Highly precise optogenetic and electrophysiological studies provide an approach to correlate the neural activity of specific cells to observed behaviors, which is important for the explorations of neural circuits in the brain and the pathologies of the nervous system. Embedded optical fibers and electrode probes are the most common invasive devices in the studies of these areas. Previous studies have shown that foreign body reactions due to chronic tissue damage lead to gradual encapsulation of the implanted probes, resulting in dysfunction of the devices. Over the past two decades, researchers have experimented with advanced materials to fabricate invasive electrodes and optical fiber devices that are mechanically compatible with soft brain tissue, thereby reducing the relative micromotion and damage between the device interface and brain tissue. Extensive efforts have been made in this area; however, challenges have yet to be overcome for practical applications. First, the assembly of the optical fiber and electrode(s) generally results in a significant increase in the volume of the implant, which produces additional tissue damage. Second, an alternative strategy is accommodating recording and waveguide channels in fibers through a thermally drawn process. However, this method has difficulty breaking the limit of the number of recorded sites (usually less than 30). Third, existing optical waveguide materials, such as polycarbonate (e.g., 2-2.4 GPa) and quartz (e.g., 77-85 GPa), have a Young's modulus far greater than that of brain tissue (e.g., 1-10 kPa). Probes made of softer biocompatible materials, on the other hand, are mechanically incapable of embedding themselves into brain tissue. Moreover, the lack of rigidity poses challenges in the implantation into living brains. A rigid shuttle device is usually required for the implantation of a flexible probe. However, the insertion and retraction of the shuttle device inevitably cause extra damage to the brain tissue. Natural silk has remarkable application prospects in biomedical and bio-optical devices due to its excellent physical and biomedical properties, such as transparency, mechanical strength and crucial bio-compatibility. Here, we present an optoelectronic probe, Silk-Optrode, with natural silk as an optical waveguide material, which can simultaneously achieve high-precision intracranial light stimulation and electrophysiological recording. Through the attachment of the flexible electrode array to the surface of the silk optical fiber, 128 recording channels can be integrated on a single probe. Meanwhile, the diameter of the probe does not notably increase. Compared with polylactic acid (PLA) and other polymer materials as previously reported for making implantable flexible optical fibers, silk has the advantage of being adjustable under different hydration conditions. The Silk-Optrode can actively transform from a hard brittle state to a flexible and high elongation state after implantation (bending stiffness decreased by ~11 times), thus simultaneously achieving the simplest implant operation and causing less chronic tissue damage. Low optical transmission loss makes silk-based optical fibers suitable for optogenetic studies as waveguides. In these studies, Silk-Optrode achieved the effective regulation of neural activity and behavior while recording more than 50 high-quality isolated unit activities within one experiment. We envision that this technology will provide new opportunities for the combination of multifunctional biomaterial devices with neurological disease research.

The paper situates its contribution within prior work on optogenetic stimulation and electrophysiological recording, noting that traditional assemblies of optical fibers with electrodes increase implant volume and tissue damage. Thermally drawn fiber approaches integrate waveguides and electrodes but are typically limited to fewer than ~30 recording sites. Conventional waveguide materials (polycarbonate, quartz) have moduli orders of magnitude higher than brain tissue, exacerbating micromotion-induced damage, while very soft polymers lack sufficient rigidity for insertion and often require traumatic shuttles. The authors propose silk as a biocompatible, transparent waveguide with hydration-tunable mechanics to overcome these limitations.

Silk optical fiber fabrication and characterization: A silk solution was preconcentrated and extruded at a controlled rate through custom-diameter nozzles (target fiber diameters 200–500 µm) into methanol to induce alpha-helix to beta-sheet transitions, forming insoluble fibers. Products were cleaned and soaked in deionized water to remove residuals. Fibers were produced with lengths >200 mm and high uniformity. Optical propagation loss was measured in PBS (wet state) using the cut-back method, yielding 3.0 dB/cm. Mechanical testing (tensile-fracture) compared dehydrated vs hydrated states: elastic modulus decreased from 38.7 MPa (dry) to 3.53 MPa (wet), with corresponding bending stiffness reduction for 200 µm diameter from 3.05E-09 to 2.77E-10 N·m² and increased ductility in the hydrated state. Stability tests in PBS showed no apparent diameter change over days. Phantom insertion tests in 1.5% agarose assessed hydration effects: dry fibers inserted smoothly; progressive hydration increased bending/softening, and after ~15 min hydration fibers were too soft to insert, defining a practical implantation window. Finite element analysis (FEA): A 3D model simulated a probe inserted 2 mm into brain tissue with the probe backend tethered to skull and 50 µm lateral brain micromotion. Comparisons were made among silk, polycarbonate, silica, and steel probes. Outputs included strain fields, relative displacement between probe tip and brain, and maximum von Mises stress during micromotion. Flexible electrode array design and fabrication: Ultra-thin microfabricated arrays with four polyimide shanks (2.5 µm thick, ~105 µm wide, 5 mm long) hosted 128 gold microelectrodes (100 nm thick). Bending stiffness of the array was ~4.23×10⁻¹³ N·m², far lower than silicon probes and lower than silk fibers; the implanted array volume was 2.713×10⁵ µm³ vs 3.14×10⁷ µm³ for the silk fiber. Arrays were connected via substrate-supported pads to flexible printed circuits using flip-chip bonding. Assembly to silk fiber used a prefabricated mold for alignment; four shanks were radially attached to the fiber quadrants with PEG (MW 30,000) or silk protein glue, then air-dried to strengthen bonding. Prior to recording, PEDOT was electrodeposited on sites to increase effective surface area and reduce impedance (~10× at 1 kHz). In vivo testing: Adult mice received injections of AAV (AG26975 PAAV-CaMKIIa-hChR2(H134R)-mCherry) in left medial prefrontal cortex (mPFC). After ≥4 weeks expression, mice were implanted with Silk-Optrode probes and recorded weekly in awake, freely moving conditions. Neural signals were acquired from up to 128 channels; optogenetic stimulation was delivered via the silk waveguide. Spike sorting used principal component analysis (PCA). Behavioral paradigms included open field tests (OFT).

• Silk optical fibers achieved low optical propagation losses of 3.0 dB/cm (wet, PBS), sufficient for optogenetic stimulation.
• Hydration-tunable mechanics: elastic modulus decreased from 38.7 MPa (dehydrated) to 3.53 MPa (hydrated); bending stiffness for 200 µm fiber dropped from 3.05E-09 to 2.77E-10 N·m² (~11× reduction), four orders of magnitude lower than quartz of same diameter.
• In agarose phantom, dry fibers inserted readily; progressive hydration increased flexibility; after ~15 min hydration, fibers were too soft to insert, defining an implantation window that minimizes time in a rigid state.
• FEA indicated silk probes best matched brain mechanics: strain fields diminished along the implant with negligible strain over most of the implanted length; the probe tip moved in attached fashion with brain tissue, and maximum brain stress during 50 µm lateral micromotion was lower than for steel, silica, or PC probes.
• Flexible electrode array metrics: bending stiffness ~4.23×10⁻¹³ N·m²; implanted array volume 2.713×10⁵ µm³ versus 3.14×10⁷ µm³ for silk fiber. PEDOT coating reduced electrode impedance ~10× at 1 kHz.
• In vivo recordings: An example mouse had 109 functional channels (85% fabrication yield) in mPFC and 57 putative single units isolated during a 10 min open field test one week post-implant (0.52 units per channel). Multiple well-isolated units were observed on single channels.
• The Silk-Optrode supports synchronized intracranial optical stimulation and multichannel electrophysiological recording in freely moving mice, with reports of low optical loss and high single-unit yields surpassing similar probe types.
• Two months post-surgery, devices showed better implant–neural interfaces with reduced immunoreactive glial responses and tissue lesions relative to prior approaches (as noted by the authors).

By exploiting silk’s hydration-dependent mechanics, the Silk-Optrode provides initial rigidity for accurate, shuttle-free implantation and subsequently transitions to a compliant state in vivo, reducing micromotion-induced strain and stress at the brain–device interface. FEA and phantom tests support that silk probes move more coherently with brain tissue, which should mitigate chronic damage and glial responses, aligning with observed reductions in immunoreactivity and tissue lesions at two months. The integration of a 128-channel ultrathin electrode array onto the silk fiber minimally increases implant volume, addressing the trade-off between functionality and tissue damage associated with conventional fiber–electrode assemblies. Low optical loss confirms suitability for optogenetic stimulation, while PEDOT-modified microelectrodes enable low-impedance, low-noise recordings that yield high numbers of well-isolated units, outperforming thermally drawn multifunctional fibers typically limited to fewer than ~30 sites. Collectively, the device advances bidirectional neural interfacing by combining high-density recording with effective optical stimulation in a mechanically adaptive, biocompatible platform for chronic use.

The study introduces a silk-based opto-electro neural probe that self-adapts mechanically post-implantation, enabling accurate insertion without shuttles and improved chronic biocompatibility. The device integrates a low-loss silk waveguide with a 128-channel ultrathin electrode array, achieving high unit yields during simultaneous optogenetic stimulation and electrophysiological recording in freely moving mice. Modeling and experiments demonstrate reduced brain strain and stress compared to conventional materials, and preliminary chronic data indicate improved tissue responses. Future work could extend chronic evaluations over longer durations and across brain regions and species, further optimize optical transmission and mechanical tuning, integrate additional modalities (e.g., fluidics, sensing), and refine scalable manufacturing and packaging for broader adoption.