Abstract: Implantable prostheses have become an important treatment option for a multitude of diseases, such as chronic pain, blindness and damaged motor function. At USC, in collaboration with several other universities, researchers are developing a microchip-based neural prosthesis for the hippocampus, a region of the brain known to be responsible for the conversion from short-term to long-term memory. One major part in accomplishing such a system is to build a microchip implementation of neural system models small enough to be implanted, while consuming very low power, so not to heat up or damage surrounding tissue. As part of this effort, in this talk, I will start by outlining the requirements to realize such a system in hardware. I will then describe a model that can capture the nonlinear dynamics of the input-output transformation of a part of the hippocampus. Next, I will demonstrate how this signal processing is implemented using subthreshold CMOS hardware to achieve a system that consumes only a tens of nW of power. I will then extend this topic by detailing the challenges one has to face using this method and how they can be mitigated. To conclude, I will discuss the future direction of this work and the impact our collaborative effort could have in the future.

Loading more stuff…

Hmm…it looks like things are taking a while to load. Try again?

Loading videos…