Performance Evaluation of a Real-Time Phase Estimation Algorithm Applied to Intracortical Signals from Human Visual Cortex

Abstract

Cortical visual prostheses are a subgroup of visual prostheses which use electrical stimulation of the occipital cortex to evoke visual percepts in profoundly blind people. The stimulation approaches are usually open-loop, meaning that the stimulation is not controlled by any other factor. However, closed-loop approaches have shown advantages in many neural prosthesis. In the case of cortical visual prosthesis, the closed-loop approach can be based on the phase of local field potentials recorded by the electrodes. Indeed, previous studies have shown that it is easier to induce perception through stimulation at certain phases of brain oscillations.

Here, we evaluated the performance of a real-time phase estimator algorithm applied to local field potentials recorded with intracortical microelectrodes inserted in the occipital cortex of a blind human volunteer. Phase estimation was more accurate at certain phases than others. The error of the estimated phase was in the range ±20$^{\circ }$.

These results should be taken into account when implementing phase-locked stimulation approaches in cortical visual prosthesis. Indeed, the phase estimation accuracy represents the limitation of the closed-loop stimulation approach.