Verfahren und Vorrichtung zur Entfernung von Störungen aus Magneto-/elektroenzephalographischen Mehrkanalsignalen

Details

Veröffentlichungs-Nr.

EP4821439

Anmeldetag

11. April 2025

Veröffentlichung

14. Oktober 2026

Rechtsraum

EP

IPC

A61B5/369, A61B5/00

Abstract

The present invention concerns a signal processing method for removal of artifacts from electroencephalography (EEG) and/or magnetoencephalography (MEG) signals by use of Generalized Eigenvalue Decomposition (GEVD), the method comprising the steps of- obtaining (1) a EEG/MEG multi-channel signal,- providing (2) a matrix factorization of said multi-channel signal by applying a generalized Eigenvalue decomposition to the signal by° determining (2.1) a reference covariance matrix (refCOV) to be used as an input value to the Eigenvalue decomposition of said multi-channel signal,° extracting (2.3) a data covariance matrix (dataCOV) from the multi-channel signal and° producing (2.4) a generalized Eigenvalue decomposition of said data covariance matrix (dataCOV) based on said reference covariance matrix (refCOV),- producing (3), by use of said matrix factorization of the multi-channel signal, an output signal corresponding to said multi-channel signal with artifacts in the signal being removed and non-artifact parts of the signal being retained, the step of- determining the reference covariance matrix (refCOV) being performed by estimating a theoretical reference covariance matrix (refCOV) based on principles of EEG/MEG signal generation, utilizing electrode locations and leadfield theory.The present invention is also related to a computer readable medium storing a computer program which causes a computer to implement this method as well as to a device for acquiring and/or for processing a EEG/MEG multi-channel signal, the device having means adapted to perform this method.

Anmelder

Firma

Université de Genève

Land

🇨🇭 Schweiz

Fachgebiete