The Layer 7 Cortical Interface: A Scalable and Minimally Invasive Brain–Computer Interface Platform

Abstract

Progress toward the development of brain-computer interfaces has signaled the potential to restore, replace, or augment lost or impaired neurological function in a variety of disease states. Existing brain-computer interfaces rely on invasive surgical procedures or brain-penetrating electrodes, which limit addressable applications of the technology and the number of eligible patients. Here we describe a novel approach to constructing a neural interface, comprising conformable thin-film electrode arrays and a minimally invasive surgical delivery system that together facilitate communication with large portions of the cortical surface in bidirectional fashion (enabling both recording and stimulation). We demonstrate the safety and feasibility of rapidly delivering reversible implants containing over 2,000 microelectrodes to multiple functional regions in both hemispheres of the Gottingen minipig brain simultaneously, without requiring a craniotomy, at an effective insertion rate faster than 40 ms per channel, without damaging the cortical surface. We further demonstrate the performance of this system for high-density neural recording, focal cortical stimulation, and accurate neural decoding. Such a system promises to accelerate efforts to better decode and encode neural signals, and to expand the patient population that could benefit from neural interface technology.

Direct laser writing of 3D electrodes on flexible substrates

“…Applications include bioelectric medicine via the stimulation of peripheral nerves [15, 16], visual prosthesis via retinal interfaces [17], and cortical interfaces to provide neuromodulation for mental disorders or enhance stroke recovery [18,19]. Multiple companies are now pursuing neural interfacing through microfabricated polymer electrodes, and these projects include future applications in neural stimulation [2] [20] [21]. However, as the charge-injecting surface area decreases, adequate currents for stimulation require higher voltages leading to accelerated delamination and degradation of stimulating electrodes [22–24].…”

Decoding multi-limb movements from low temporal resolution calcium imaging using deep learning

“…Non-invasive or minimally-invasive methods for decoding behaviorally-relevant variables are attractive for the long-term use and stability of BMIs and have recently received massive investment of resources from industry (e.g. Défossez et al, 2023; Ho et al, 2022). However, non-invasive neural recording methods such as electroencephalogram (EEG) and magnetoencephalogram (MEG) primarily reflect strong, synchronous input into a cortical area, lacking fine spatial resolution (Lopes da Silva, 2013).…”

Long-term unsupervised recalibration of cursor BCIs

“…This latter step ensures similar average drift between simulated and observed neural activity, but leaves unspecified the variance of this drift. We chose to simulate a low variance in the amount of day-to-day drift, motivated by progress in next-generation BCI systems [39][40][41] . As channel counts scale, iBCIs will likely acquire a degree of signal redundancy that provides some protection from nonstationarities (as noted in e.g.…”