DIrectional and SCalable (DISC) Microelectrode Array for Speech Decoding

摘要

Abstract Currently, the brain-computer interface (BCI) field has demonstrated two distinct device strategies - macroelectrodes (e.g., surface grids and depth) versus microelectrode arrays, and some are even pushing the field to smaller, higher density arrays hoping to address the general signal degradation. Both approaches have been in development for decades. However, BCI devices to treat aphasia, dysarthria, or locked-in syndrome also need to access deeper brain regions given the very large, parallel networks involved in speech. Consider that two-thirds of the cortex is buried beyond the reach of most state-of-the-art technologies. We have designed a novel approach to brain recordings to address the challenge of multi-scale recordings at any desired depth. Our team presents a novel device whose form is based on the proven safety and utility of the stereo-EEG (SEEG). We created a directional and scalable local field potential array (DISC) using the phenomenon of "substrate shielding". This is not the first combination micro/macro device but is the first to demonstrate stereo-local field potentials using a patent pending design. Our preliminary in vivo data demonstrates significant improvement when using DISC in many critical factors predictive of future BCI performance: (i) signal amplitude, (ii) signal-to-noise ratio, and (iii) source separation in classification tasks. This project will allow us to safely test word decoding performance both offline and online in epilepsy volunteers from speech regions. The project's first aim is to develop a robust DISC hybrid assembly with 128 or more recording channels per implant. Each implanted device will be a commercially available SEEG combined with microelectrodes without any modification to the clinical function of the device. Aim 1 will include verification, validation, biocompatibility, and electrical safety testing. Aim 1 will also include functional and safety studies in animals to complete our effort to provide a safe, reliable system prior to human feasibility studies. After all milestones are met, including receiving an FDA investigational device exemption, this novel recording system will demonstrate the effect size and variance of word and speech decoding in humans as compared with conventional ring electrodes. Typically, 12-20 depth arrays are used in epileptogenic monitoring and we will replace two depth electrodes with a DISC hybrid assembly in 8 experimental patients and compare decoding performance to the within-patient controls and with a separate 8 patients having SEEG electrodes only. Enrolled volunteers will conduct overt and covert speech tasks. Positive results will inform and enable a word and speech decoder for persons suffering from locked-in syndrome and eventually non-fluent aphasia.