A Model for Assessing the Electromagnetic Safety of an Inductively Coupled, Modular Brain-Machine Interface.

Brain-Machine Interfaces (BMI) offer the potential to modulate dysfunctional neurological networks by electrically stimulating the cerebral cortex via chronically-implanted microelectrodes. Wireless transmitters worn by BMI recipients must operate within electromagnetic emission and tissue heating limits, such as those prescribed by the IEEE and International Commission on Non-Ionizing Radiation Protection (ICNIRP), to ensure that radiofrequency emissions of BMI systems are safe. Here, we describe an approach to generating pre-compliance safety data by simulating the Specific Absorption Rate (SAR) and tissue heating of a multi-layered human head model containing a system of wireless, modular BMIs powered and controlled by an externally worn telemetry unit. We explore a number of system configurations such that our approach can be utilized for similar BMI systems, and our results provide a benchmark for the electromagnetic emissions of similar telemetry units. Our results show that the volume-averaged SAR per 10g of tissue exposed to our telemetry field complies with ICNIRP and IEEE reference levels, and that the maximum temperature increase in tissues was within permissible limits. These results were unaffected by the number of implants in the system model, and therefore we conclude that the electromagnetic emissions our BMI in any configuration are safe.

Tissue response to a chronically implantable wireless intracortical visual prosthesis (Gennaris array).

OBJECTIVE: Cortical vision prostheses aim to restore visual percepts to those who have lost sight by delivering electrical stimulation to the visual cortex. These devices need to be implanted intracranially using subdural or intracortical microelectrodes, and should preferably dispense with the need of transcranial wiring. The risks of cortical tissue injury from mechanical trauma, material biocompatibility, heat generation, electrical stimulation and long-term immune responses need to be evaluated. In this paper, we investigate the biological response to a wireless cortical vision prosthesis (Gennaris array), by characterizing the histological changes that occur following chronic electrical stimulation. APPROACH: Ten arrays (7 active, 3 passive) were implanted in three sheep using a pneumatic insertor. Each device consisted of a wireless receiver and Application Specific Integrated Circuit encased in a ceramic box, and could deliver electrical stimulation through one of 43 electrodes. MAIN RESULTS: Stimulation was delivered through seven of these devices for up to 3 months and each device was treated as independent for further analysis. Cumulatively, over 2700 h of stimulation were achieved without any observable adverse health effects. Histology showed that the devices and implantation procedure were well tolerated by the brain with a similar tissue response to the more common Utah arrays. However, voltage transients across the stimulating electrodes were not measured so exact charge injection could not be verified. SIGNIFICANCE: This work represents one of the first long-term tests of a fully implantable cortical vision prosthesis. The results indicate that long-term stimulation through wireless arrays can be achieved without induction of widespread tissue damage.

CMOS stimulating chips capable of wirelessly driving 473 electrodes for a cortical vision prosthesis.

OBJECTIVE: Implantable neural stimulating and recording devices have the potential to restore capabilities such as vision or motor control to disabled patients, improving quality of life. Implants with a large number of stimulating electrodes typically utilize implanted batteries and/or subcutaneous wiring to deal with their high-power consumption and high data throughput needed to address all electrodes with low latency. The use of batteries places severe limitations on the implant's size, usable duty cycle, device longevity while subcutaneous wiring increases the risk of infection and mechanical damage due to device movement. APPROACH: To overcome these limitations, we have designed and implemented a system that supports up to 473 implanted stimulating microelectrodes, all wirelessly powered and individually controlled by micropower application specific integrated circuits (ASICs). MAIN RESULTS: Each ASIC controls 43 electrodes and draws 3.18 mW of power when stimulating through 24 channels. We measured the linearity of the digital-to-analog convertors (DACs) to be 0.21 LSB (integrated non-linearity) and the variability in timing of stimulation pulses across ASICs to be 172 ns. SIGNIFICANCE: This work demonstrates the feasibility of a new low power ASIC designed to be implanted in the visual cortex of humans. The fully implantable device will greatly reduce the risks of infection and damage due to mechanical issues.