Rank estimation and tensor decomposition using physics-driven constraints for brain source localization

This paper deals with the tensor-based Brain Source Imaging (BSI) problem, say finding the precise location of distributed sources of interest by means of tensor decomposition. This requires to estimate accurately the rank of the considered tensor to be decomposed. Therefore, a two-step approach, named R-CPD-SISSY, is proposed including a rank estimation process and a source localization procedure. The first step consists in using a modified version of a recent method, which estimates both the rank and the loading matrices of a tensor following the canonical polyadic decomposition model. The second step uses a recent physics-driven tensor-based BSI method, named STS-SISSY, in order to localize the brain regions of interest. This second step uses the estimated rank during the first step. The performance of the R-CPD-SISSY algorithm is studied using realistic synthetic interictal epileptic recordings.

[1]  Xu Han.  Robust low-rank tensor approximations using group sparsity , 2019 .

[2]  Dayan A. Guimarães,et al.  Simple and Efficient Algorithm for Improving the MDL Estimator of the Number of Sources , 2014, Sensors.

[3]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[4]  Rasmus Bro,et al.  MULTI-WAY ANALYSIS IN THE FOOD INDUSTRY Models, Algorithms & Applications , 1998 .

[5]  Pablo A. Parrilo,et al.  Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization , 2007, SIAM Rev..

[6]  Wim Van Paesschen,et al.  Block term decomposition for modelling epileptic seizures , 2014, EURASIP J. Adv. Signal Process..

[7]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..

[8]  Stephen P. Boyd,et al.  Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers , 2011, Found. Trends Mach. Learn..

[9]  Rémi Gribonval,et al.  SISSY: An efficient and automatic algorithm for the analysis of EEG sources based on structured sparsity , 2017, NeuroImage.

[10]  Thomas Kailath,et al.  Detection of signals by information theoretic criteria , 1985, IEEE Trans. Acoust. Speech Signal Process..

[11]  Christopher J. Hillar,et al.  Most Tensor Problems Are NP-Hard , 2009, JACM.

[12]  Lars Kai Hansen,et al.  Parallel Factor Analysis as an exploratory tool for wavelet transformed event-related EEG , 2006, NeuroImage.

[13]  Xu Han,et al.  Low rank canonical polyadic decomposition of tensors based on group sparsity , 2017, 2017 25th European Signal Processing Conference (EUSIPCO).

[14]  Rasmus Bro,et al.  Multiway analysis of epilepsy tensors , 2007, ISMB/ECCB.

[15]  Lotfi Senhadji,et al.  Localization of spatially distributed brain sources after a tensor-based preprocessing of interictal epileptic EEG data , 2015, 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).