On Recalibration Strategies for Brain-Computer Interfaces Based on the Detection of Motor Intentions

Abstract

Coupling motor intentions decoded from cortical activities with coherent proprioceptive feedback is of interest for the motor rehabilitation of neurological patients with lesions in the central nervous system. For these interventions to be effective, repeated sessions need to be carried out to achieve functional long-lasting plastic changes of cortical circuits. Electroencephalography-based Brain-Computer Interfaces typically show significant decreases in accuracy when used across multiple sessions with fixed parameters. Therefore, it is important to look for optimal strategies to recalibrate these classifiers. Here we compare different recalibration strategies for systems decoding motor intentions based on electroencephalographic data of neurological patients.