(7) A Novel Fully Implantable High Data Rate BCI for Speech Restoration

Introduction: Brain-computer interfaces (BCIs) have demonstrated their ability to provide functional control of computing devices and to restore lost communication abilities for individuals with profound motor and speech deficits due to neural injury or progressive neuromuscular disease. Recently, these systems have achieved a high level of performance sufficient to enable restoration of interactive, conversational communication [1], enabled by intracortical electrode arrays implanted in motor areas. Current speech BCI studies utilize research devices with percutaneous components limiting their potential use in a clinical setting. We present features of a fully implantable, intracortical BCI system designed for clinical adoption while simultaneously supporting higher electrode counts that could facilitate increased BCI performance with potential to provide a clinical benefit to people living with impaired communication.

Methods: The system is composed of a cortical module, lead, and an internal transceiver. A single cortical module has 421 intracortical microelectrodes (each 1.5mm in length, 40 µm in diameter mean shank width) that are amplified and digitized on the module to allow for neural signal transmission over a subcutaneous lead. The internal transceiver form factor and implant procedure follow an approach similar to that of an implantable pulse generator of a deep brain stimulator. The internal transceiver receives power for the system with a wireless inductive link from an external, wearable transceiver. Broadband data are telemetered via a wireless optical communications link and support extraction of spiking, spikeband, and local field potential (LFP) based features. Power delivery and optical communications bandwidth are designed to support up to four cortical modules, or 1684 intracortical microelectrodes.

Results: Zero serious device or procedure-related events have been observed across 31 research sheep procedures. In a research setting, signal stability has been observed out to 2 years. Preclinical testing designed to support an implantable medical device is underway for safety, biocompatibility, and functionality. Twenty-six week biocompatibility is currently supported in accordance with ISO10993. Chronic implantation of the cortical module in ovine auditory cortex yields neuronal spiking and LFP features, and demonstrates a high bit rate readout of brain state. These neural features are validated based upon their signal characteristics and their dynamics with respect to auditory stimuli.

Conclusion: Preclinical test results are intended to support initiating a First-In-Human clinical study, under IDE, to investigate the feasibility of speech restoration and associated, potential clinical benefit in a population suffering from impaired communication due to a progressive neuromuscular disease or a neural injury.