VSSI-L1N: EEG Extended Source Imaging Based on Variation Sparsity and Nuclear-norm Regularization

Accurately reconstructing the location and extent of cortical sources is crucial for understanding brain function and disease mechanisms. Source signals inherently possess low-rank structural characteristics and the existence of noise and background human body activities has a great influence on the source imaging results. This paper presents an Electroen-cephalography (EEG) source imaging (ESI) algorithm, VSSI-LIN. The algorithm employs a combined constraint of$L$21-norm and nuclear norm to improve the accuracy and resolution of cortical source reconstruction. VSSI-L1Nuses the$L$21-norm in the spatial variation domain to promote sparsity and extract the time-domain information, employs the nuclear norm to capture the low-rank structure of the signal, adjusting and reducing the interference of human body background activities. The solution of VSSI-L1Ncan be effectively derived through the alternating direction method of multipliers (ADMM) algorithm. The effectiveness of this method is demonstrated through simulation and experiments, and it is shown to have higher accuracy and better spatial resolution. The source code of VSSI-LIN is available at https://github.com/hangj-cache/VSSI-L1N.git.