Power-CCA: Maximizing the correlation coefficient between the power of projections

This work presents a variation of canonical correlation analysis (CCA), where the correlation coefficient between the instantaneous power of the projections is maximized, rather than between the projections themselves. The resulting optimization problem is not convex, and we have to resort to a sub-optimal approach. Concretely, we propose a two-step solution consisting of the singular value decomposition (SVD) of a “coherence” matrix followed by a rank-one matrix approximation. This technique is applied to blindly recovering signals in a model that is motivated by the study of neuronal dynamics in humans using electroencephalography (EEG) and magnetoencephalography (MEG). A distinctive feature of this model is that it allows recovery of amplitude-amplitude coupling between neuronal processes.

[1]  M. Berger,et al.  High Gamma Power Is Phase-Locked to Theta Oscillations in Human Neocortex , 2006, Science.

[2]  R. Eckhorn,et al.  Phase correlation of cortical rhythms at different frequencies: higher-order spectral analysis of multiple-microelectrode recordings from cat and monkey visual cortex , 1997 .

[3]  Liqun Qi,et al.  A practical method for computing the largest M-eigenvalue of a fourth-order partially symmetric tensor , 2009, Numer. Linear Algebra Appl..

[4]  Hans Knutsson,et al.  Learning multidimensional signal processing , 1998, Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170).

[5]  Chen Ling,et al.  Biquadratic Optimization Over Unit Spheres and Semidefinite Programming Relaxations , 2009, SIAM J. Optim..

[6]  H. Hotelling.  Relations Between Two Sets of Variates , 1936 .

[7]  R. Eckhorn,et al.  Phase correlation among rhythms present at different frequencies: spectral methods, application to microelectrode recordings from visual cortex and functional implications. , 1997, International journal of psychophysiology : official journal of the International Organization of Psychophysiology.

[8]  J. Palva,et al.  Phase Synchrony among Neuronal Oscillations in the Human Cortex , 2005, The Journal of Neuroscience.

[9]  D. Tucker,et al.  EEG coherency. I: Statistics, reference electrode, volume conduction, Laplacians, cortical imaging, and interpretation at multiple scales. , 1997, Electroencephalography and clinical neurophysiology.

[10]  Jürgen Kurths,et al.  Detection of n:m Phase Locking from Noisy Data: Application to Magnetoencephalography , 1998 .

[11]  Zhi-Quan Luo,et al.  A Semidefinite Relaxation Scheme for Multivariate Quartic Polynomial Optimization with Quadratic Constraints , 2010, SIAM J. Optim..

[12]  Gabriel Curio,et al.  Cross-Frequency Decomposition: A novel technique for studying interactions between neuronal oscillations with different frequencies , 2012 .

[13]  Chen Ling,et al.  Standard bi-quadratic optimization problems and unconstrained polynomial reformulations , 2012, J. Glob. Optim..

[14]  M. Hallett,et al.  Identifying true brain interaction from EEG data using the imaginary part of coherency , 2004, Clinical Neurophysiology.