The central obstacle has long been mechanical. Optical windows implanted through the skull have transformed neuroscience by letting researchers image the living brain over months, but those implants are stabilized by bone. Peripheral nerves have no such anchor. They are wrapped in soft, mobile muscle and connective tissue, and they stretch and shift with every step. Earlier attempts at chronic nerve windows relied on rigid frames that were prone to instability, misalignment, and tissue damage, and were easily dislodged by the animals’ grooming. The new approach solves the anchoring problem with a flexible implant cast from polydimethylsiloxane, or PDMS, a silicone elastomer whose elastic properties closely match those of soft tissue.

Surgeons split the biceps femoris muscle overlying the sciatic nerve, displace it with fine sutures, and slide the folded PDMS window into a subcutaneous pocket. Skin tension holds the implant’s outer brim in a circumferential groove, while regenerating fascia grows through holes in the subdermal brim to anchor the device long-term. The result is a fluid-filled optical port with an injection septum through which dyes can be delivered and debris can be cleared. The numbers are striking: the soft windows remained embedded for a median of about seven weeks, with 90 percent lasting at least 18 days and some exceeding 80 days. Rigid 3D-printed acrylic implants of comparable size, tested as a control, lasted a median of only about nine days.

Crucially, the implant does not appear to interfere with the nerve it exposes. Mice with windows walked without asymmetry on a grid-walking test, and muscle force and electromyography measurements of sciatic-innervated muscles three months after implantation showed no deficits, with compound muscle action potentials in the normal range. Histological sections of windowed nerves roughly 90 days after surgery revealed axon densities and calibers indistinguishable from the opposite, unoperated leg. The researchers report that fluorescent signals from the transgenic labels did not photobleach appreciably, allowing the same individual axons, identified by fiducial features such as bends and swellings, to be tracked repeatedly over at least 24 days.

The imaging itself combines multiple modalities. Confocal and two-photon microscopy captured fluorescently labeled axons and Schwann cells within roughly 30 to 100 micrometers of the nerve surface, while second-harmonic generation imaging revealed the collagen and elastin of the extracellular matrix without any label at all. Genetically encoded reporters did much of the heavy lifting: Thy1-YFP and Thy1-GCaMP mice label neurons, and S100-GFP mice label Schwann cells. A topical dye called Nile Red stains myelin sheaths and lipid droplets, and an injected tomato lectin conjugate outlines blood vessels, even allowing the team to film individual blood cells streaming through intraneural capillaries. Axon diameters measured through the window averaged about 4 micrometers, consistent with large myelinated fibers in the mouse sciatic nerve.

Then came the injuries, and the window delivered a front-row view. Within minutes of a partial crush injury, Schwann cells near the damage site swelled, retracted, and shifted position, while myelin and collagen outside the injury zone remained stationary. Within 20 to 30 minutes, the fluorescent signal from affected Schwann cells dropped sharply, and over the following hour the cells beaded up and pulled back their membranes, behavior consistent with conversion to the mobile repair phenotype that Schwann cells adopt to clear debris and guide regrowth. In axon-labeled mice, the distal portions of damaged fibers lost their fluorescence over one to two days as Wallerian degeneration set in, with visible fragmentation by day two.

Regeneration followed with cinematic clarity. By four days after injury, thin new axonal sprouts tipped with growth cones had emerged from the proximal nerve segments and pushed through the injury zone, often retracing the paths of the degenerated fibers. The researchers estimate regeneration rates of at least one millimeter per day, matching published values for the mouse sciatic nerve. Growth cones of this kind had previously been filmed only in zebrafish and in very superficial mouse nerves, never in a major mammalian nerve trunk. In compression injuries that preserved the endoneurial channels, the delicate collagenous tubes left behind by degenerated axons, regrowing fibers closely aligned with those residual channels, supporting the long-standing idea that the extracellular matrix provides physical guidance cues for pathfinding. In one striking case, an axon encountering a surgically deformed channel deviated from it and sprouted multiple growth tips into a small defect in the epineurial collagen.

The platform also captured the slowest process of all: repair across a complete transection. The team cut the sciatic nerve entirely and bridged the 2.5-millimeter gap with a hemitubular PDMS conduit, chosen over a closed tube to permit optical access and nutrient exchange. Over days, mobile fluorescent cells, likely immune and stromal cells, populated the saline-filled gap, peaking in density around day eleven. By day 25, bundled axonal sprouts had crossed the conduit along its axis, and their density and organization continued to increase through day 35, accompanied by the return of grasping ability and sensation. Histology at day 59 confirmed small, myelinated regenerative axons at the distal end. In most of the conduit-repair mice followed, tissue cables formed, fluorescent regrowth appeared, and functional recovery was observed.

The authors are candid about limitations. Tissue scattering restricted imaging largely to the outermost portion of the nerve, an estimated 5 to 30 percent of its volume, and a fibrotic membrane inevitably formed against the implant within two to three weeks, degrading image quality, though it could often be mechanically cleared through the injection port. Whether that clearing itself disturbs regeneration remains untested, and the functional assessments in the uninjured window model involved small sample sizes. The transgenic reporters, too, are not perfectly cell-type-specific, labeling some non-neuronal and non-Schwann cell populations. Future refinements may include anti-fibrotic coatings, adaptive optics, three-photon excitation, and wearable or implantable cameras that could image awake, moving animals.

Even so, the implications are broad. By imaging the same nerve region repeatedly within the same animal, the technique eliminates inter-animal variability as a confound and makes it possible to link early injury events to later regenerative outcomes, something discrete histological sampling cannot do. On timescales of days to weeks, researchers can now track angiogenesis, Schwann cell migration, formation of the regenerative bands of Büngner, and extracellular matrix remodeling; on timescales of milliseconds to minutes, calcium transients, neuropeptide release, mitochondrial transport, and immune cell trafficking come within reach. The method is particularly suited to evaluating biofabricated nerve grafts, and could allow surgeons and scientists to see early whether a repair conduit is filling with regenerative tissue or failing. Combined in the future with genetically encoded calcium indicators or optogenetic actuators, the nerve window could enable both sensing and control of peripheral nerve activity, opening a new chapter in the study and treatment of peripheral nerve disorders.

Subject of Research: Longitudinal in vivo imaging of peripheral nerve injury and regeneration in mice using a soft PDMS optical window implant

Article Title: Longitudinal Single‐Axon‐Resolution Imaging of Peripheral Nerve Injury Response in Mice Using an Optical Window Implant

Article References: Luzhansky, I. D., Anisman, E., Perez, R., Zhang, S., Hoffman, M., Hunter, D., Cherian, A., Bonner, J., Feria, E., Ahmed, A., Malik, M., Du, J., Sudlow, L. C., Brogan, D. M., Wood, M. D., & Berezin, M. Y. (2026). Longitudinal Single‐Axon‐Resolution Imaging of Peripheral Nerve Injury Response in Mice Using an Optical Window Implant. Advanced Science, 13(53), Article e76375. https://doi.org/10.1002/advs.76375

Image Credits: AI Generated

DOI: 10.1002/advs.76375

Keywords: peripheral nerve injury, sciatic nerve, PDMS optical window, axon regeneration, Schwann cells, two-photon microscopy, Wallerian degeneration, growth cones, nerve conduit repair, endoneurial channels, intravital imaging, Thy1-YFP mice