Neural interfaces are devices that are implanted into the body and interface electrically with the central (CNS) or peripheral nervous system. They are inserted surgically for short periods of time or chronically to diagnose or treat several neural diseases such as epilespy, deafness or chronic pain. The devices present in the clinic today are usually made of thick metallic disks embedded in a thick and rigid elastomeric carrier. The mechanical signature of these stiff implants is orders of magnitude away from the soft and dynamic tissues found in the CNS. This mismatch hinders a good interface between the implanted electrode and tissue leading to chronic scarring, low efficiency of stimulation or low spatial resolution recordings.

Advances in materials science and microengineering has enabled the creation of implantable soft bioelectronic interfaces. These devices integrate passive and active electronic functions in flexible and soft substrates. They can then interface intimately with tissues in numerous anatomical locations in the body. With manufacturing processes borrowed from the semi-conductor and MEMS industry, these implants can be fabricated with small footprints and include dozen of channels of transducers for stimulation, recording, drug-delivery, etc.

The main objective of this Thesis is to establish a reliable and scalable microfabrication process to create soft implants and use them for experiments in large animal models. By using laser-machined silicones and elastic thin films, fully stretchable devices could be manufactured easily. On the spinal cord, soft implants enabled the restoration of movement in non-human primates implanted chronically where the devices remained stable in the time-frame of 6 weeks. On the brain, microelectrode arrays recorded high signal amplitude signals and low noise, enabled by the close contact to the cortex. New crucial building blocks were developed such as a low-profile, flexible high-channel count connector or surgical tools for implantation onto the spine. Scaling-up the soft brain implants to human anatomy permitted their implantation in cadaveric models to validate their use in the surgery room. Ultimately, these results built the foundation for the translation of the soft technology to the clinic.

To further enhance the capabilities of the soft implants, thin silicon chips were integrated in the silicone substrate. The main challenge is the mechanical mismatch between the chip and soft substrate material that can lead to delamination at low strains. A rigid island concept was introduced and validated using simple silicon chips with embedded tracks. The integrated system could be stretched up to 30% elongation without electrical failure. Thin bare electronic chips were then integrated and showed great performance when in use. Further developments on the integration and hermetic packaging will bring new functionalities to the soft implant technology.

The soft neurotechnology presented here can be readily translated to the clinic enabled by the combination of the high-yield manufacturing process, in-vivo validation in large animal models and human anatomy, and the material transfer to medical grade materials. This will bring radical changes to existing treatments and enable new therapies such as large-scale brain machine interfaces, personalized spinal cord neuromodulation and countless others.