It has become a standard medical procedure to replace lost body functions with prostheses. Recently some form of visual sensation has been restored in blind patients with the help of retinal implants. Retinal implants use electronics to stimulate retinal g anglion cells based on information recorded with a video camera.
It has been known for 50 years that neural circuits in the retina significantly change the space-time properties of the image flow that falls on the photoreceptor cells. Therefore it is desirable to use retinal neural algorithms between the image captured by the video camera and the electronics that stimulate retinal cells.
An earlier multi-layer retinal model is far slower than real-time. Based on my experience with retinal modeling in my Ph D thesis lab as well as at the University of California Berkeley and Harvard University my goal is to develop novel spatio-temporal algorithms corresponding to different retinal channels that can be implemented real time.
These algorithms are based on state-of-the-art recordings from mammalian retinas. The algorithms use the analog-and-logic cellular wave-computing paradigm called Cellular Nonlinear Networks that was co-invented by my former PhD supervisor.
The collaboration with Seville, Spain and two Berkeley, CA laboratories led to existing silicon chips. These chips are integrated into an industrial high-speed camera computer, called Bi-I that was awarded the title `Product of the year at Vision 2003 in Stuttgart.
During my postdoctoral training I plan to learn neurobiological methodologies to complement my current engineering knowledge including patch clamping, imaging and confocal microscopy. The host neurobiology laboratory is uncovering the retinal circuit details: Friedrich Miescher Institute, Basel.
The project is truly interdisciplinary by its nature; it combines my advanced computer science and engineering background with the host laboratory expertise and facilities in neurobiology.
Call for proposal
See other projects for this call