Group only

Retinal Implant

Retinal prosthesis overview

Most progressive vision loss occurs when the first layer of the retina (the photoreceptors) is damaged. The photoreceptors convert the incoming light to electrical signals that will be delivered to optic nerve through the rest of the retina, bipolar and ganglion cells. The aim of retinal prosthesis implants is to recreate this functionality in patients suffering from photoreceptor damage, through the use of electrical stimulation.

Our approach involves the use of an implantable electrode to stimulate the optic nerve, which communicates with, and derives power from the outside world via a wireless interface. An implantable integrated circuit is used to mediate this communication and power delivery, as well as generate the stimulus for the electrodes.

We have recently joined scientists from USC and Doheny Eye Institute. This project is a collaborative effort with the support of NSF and USC's BMES-ERC (Biomimetic Microelectronic Systems - Engineering Research Center).

Neuro-Stimulator for Retinal Prosthesis
Team member: Manuel Monge

Stimulation waveform

After the data is captured and sent to the implanted chip, the retina’s cells, bipolar and ganglion cells, need to be stimulated. For this purpose, biphasic current stimulation is used to inject and remove charge from the retinal tissue, electrically stimulating the cells and therefore the optic nerve. Since the number of electrodes has to be large and the area available is small (1000 electrodes inside the retina), the stimulator driver has to be small, low power and withstand high output compliance voltages due to the high output impedance of the electrode-tissue interface. Some approaches for the stimulator driver use old processes to support the high compliance voltages at the output but this increases the area used by the stimulator and its power consumption.

Our approach uses new processes to reduce the size of the stimulator and the power consumption. A 65nm prototype is being developed including additional techniques to overcome the limitations of the technology for this application. Thus, shared self-healing circuits have been designed to match both phases of the biphasic current stimulation, minimizing the remained electric charge in the tissue to safety levels and reducing the effects of process and voltage variations. In addition, protecting circuits have been used to tolerate up to +2.3V/-2.3V output voltages with 1.2V low-voltage transistors and 2.5V high-voltage transistors in a 65nm process; and power management circuits to minimize the power consumption. Also, the biphasic current stimulating waveform is programmable and allows a controllable pulse shape.

Related Publications:

Process Variation Correction for Data Amplifier
Team member: Kaveh Hosseini

Related Publications:

Wireless Telemetry System
Team member: Laleh Rabierad

Related Publications:

Wireless Power Delivery System
Team member: Meisam Hornavar Nazari

Related Publications:

© California Institute of Technology


Mixed-mode Integrated Circuits and Systems California Institute of Technology