Each year, several thousand patients a year become locked-in. The 2007 movie The Diving Bell and the Butterfly raised awareness of the condition, which results in paralysis of all voluntary muscles in the body except the eyes. If these patients can communicate at all, it is often an arduous task using residual eye movements to select letters or words one by one. We are focusing on a more tractable solution of devising a communication prosthesis that decodes a finite set of words and letters from cortical signals.

A micro-electrocorticography (micro-ECoG) grid, attached to a US quarter. Implant takes up approx 1/10th of the size of the coin.
micro-electrocorticography (micro-ECoG) grid (Spencer Kellis)

To achieve this, we are using novel micro-electrocorticography (micro-ECoG) grids to interface with cortical speech areas in human patients. micro-ECoG grids consist of an array of microscale, high impedance electrodes which rest on the surface of the brain. Because of their non-penetrating nature, micro-ECoG grids can be safely placed over eloquent cortex – brain areas that control speech and language, vision, sensation, and movement. Since micro-ECoG grids have closely spaced microscale electrodes that directly contact the cortex, they can potentially record independent neural signals from multiple, small areas of the cerebral cortex that will be needed to control a communication prosthesis.

The goal of this project is to develop and validate the design of micro-ECoG grids for recording surface local field potentials (LFPs), determine the time and frequency domain features in LFPs that are useful for decoding speech, and implement a proof-of-concept, real-time communication prosthesis.