Flexible, Self-Attaching Substrates for Extra-Neural Cuff Interfaces based on Shape-Shifting Polymers

Electrical recording and stimulation in the peripheral nervous system (PNS) is used to treat neurological disease, control neuro-prosthetic devices, and for fundamental studies in neuroscience. The study of natural neural networks is of particular interest for applications in the fields of machine learning, automation, self-driving cars, robotics, and unmanned aerial vehicles. Extra-neural cuffs, which wrap the nerve bundles with sensing/stimulating electrodes, are often preferred over penetrating probes because nerve damage is minimized. However, state-of-the-art extra-neural cuffs have many limitations including large size, a limited number of electrical contacts, a reliance on suturing for attachment to the nerves, low compliance and flexibility, and the inability to adjust to changes in nerve diameter during, for example, the flexing of joints. A serious problem for stiff cuffs is that they must be larger than the nerve to avoid compression damage. A gap between cuff and nerve then compromises spatial resolution and signal-to-noise ratio in recording cuffs and requires higher current pulses for stimulation. A gap can also serve as a pocket for growth of scar tissue which will further degrade communication between the electrode and the nerve, and can eventually lead to failure of the interface. This thesis presents fabrication and characterization of substrates for extra-neural cuff interfaces (ECI) that overcome the limitations outlined above. As a result, cuff interfaces made with these substrates will feature high longevity and resolution over a wide range of applications. These substrates are extremely thin (≤ 2 μm), have programmable diameters (over the range of 100-500 μm), and, being self-wrapping, they can securely hold nerves without suturing. Moreover, since suturing or other permanent attachment methods are unnecessary, the cuffs detach from the nerves when faced with extreme strain (e.g. due to the motion of motile limbs). Since the cuffs are so thin, they are highly compliant, and can accommodate increases in nerve diameter with little compression. The cuffs are also extremely resilient, retaining their diameter even after 100 open-close cycles. The ECI substrates have been targeted for recording neural signals on the descending contralateral movement detector (DCMD) neuron in the locust, which provides a direct insight to the mechanism of biophysical processing of the LGMD neuron. This will serve as the first step toward developing cuffs for clinical applications in humans. Shape-shifting phenomena of thin polymer films are used to create and characterize substrates for the ECI. We have explored two approaches to shift the shape of 2D films to a 3D structure. The first approach involves implanting the surface of 2 μm thick, monoaxially textured polymer films of polycarbonate (Makrofol and Lexan) and polyethylene terephthalate (Mylar) to a depth of ~ 0.6 μm with energetic helium ions. The implant causes chain scissioning and carbonizes the implanted region, thus, leading to shrinkage. This shrinkage causes global buckling of the resulting bilayer film. A careful tailoring of initial conditions of implantation method yields cuffs with diameters in 100-500 μm range. Unfortunately, with just implantation, the diameter of these cuffs is sensitive to ambient moisture, making them unreliable. For the second approach, in addition to helium ion implantation of 2 μm thick, monoaxially-textured polycarbonate film to ~ 0.6 μm depth, heating above its glass transition temperature (~ 300 F) is done. The heating shrinks the un-implanted layer (~ 1.4 μm thick) of polycarbonate in the textured direction. The difference in strain, thickness, and elastic modulus between the implanted and the un-implanted polycarbonate causes the bilayer to bend according to the Timoshenko model. Significantly, with this approach, radius of curvature is not affected by moisture; it is even impervious to immersion in phosphate buffered saline, which is widely used to simulate cerebro-spinal fluid. Bulge testing, tensile force measurements, surface energy determinations, and Raman spectroscopy were used to character the bilayer materials. Finally, some limitations of the current process are discussed and future work for fabricating functional extra-neural cuff interfaces is suggested.