Skip to content Skip to navigation

Design of a cell-type specific bi-directional electrical interface for artificial retinas

D. Muratore1,2, E. J. Chichilnisky3


1 Department of Electrical Engineering and Wu Tsai Neurosciences Institute, Stanford University, CA, USA
2 Department of Microelectronics, Delft University of Technology, The Netherlands
3 Department of Neurosurgery and Ophthalmology and the Hansen Experimental Physics Laboratory, Stanford University, CA, USA


Current retinal prostheses provide coarse communication with the target neural circuitry, because they fail to  respect its cellular and cell-type specificity. Instead, they indiscriminately activate many cells at the same time and provide only partial restoration of vision. A clear example of this problem is the simultaneous activation of ON and OFF cells in the same region of the retina.
Because diverse cell types are intermixed in the retinal circuitry, a future artificial retina will need to sort spikes coming from different cells, and sort the recorded cells into different cell types. This in turn will permit the device to stimulate each cell and cell type in a way that matches natural function. No neural interface has ever been developed that can achieve these goals; however, advances in circuit design as well as in our understanding of the retina now bring this goal within reach.
During this talk, I will present the ongoing efforts at Stanford University to build a bi-directional neural interface capable of adapting itself to the underlying biology and interact with neural circuitry at cellular and cell-type resolution. Finally, I will dive into the design choices that will allow us to process the massive amount of recorded data required to achieve cell type classification within the power and area constraints of an implantable device.


Financial disclosure: None

Go Back