• Source Power Correlation analysis (SPoC, Dähne et al. 2014a)
• multimodal Source Power Correlation analysis (mSPoC, Dähne et al., 2013)
• canonical Source Power Correlation analysis (cSPoC, Dähne et al., 2014b)
• Spatio-Spectral Decomposition (SSD) for dimensionality reduction (Nikulin et al., 2011, Haufe et al. 2014b)

0. Code for mSPoC to be added soon!
1. Please make sure the util folder (and all of its subfolders) are on the Matlab path. Otherwise the optimization required for (m/c)SPoC will not work! Run the startup_spoc.m script to add folders to the path.
2. Please read the documentation of the matlab functions ssd.m, spoc.m, mspoc.m, cspoc.m and run / look at the respective examples. I have tried to explain everything that you need to know to use the functions. If there is unclarity, please let me know and I will try to improve the documentation.
3. It is highly recommened to use dimensionality reduction via SSD before applying (m/c)SPoC. Dimensionality reduction greatly increases the computational speed and improves the quality of the results! Below you find a snippet of matlab code that shows an example of how to use SSD for preprocessing.
4. EEGLAB plugins are on the way!

S. Dähne, F. Biessman, F. C. Meinecke, J. Mehnert, S. Fazli, K. R. Müller, "Integration of Multivariate Data Streams With Bandpower Signals", IEEE Transactions on Multimedia, 15(5):1001-1013, 2013

S. Dähne, F. C. Meinecke, S. Haufe, J. Höhne, M. Tangermann, K. R. Müller, V. V. Nikulin, "SPoC: a novel framework for relating the amplitude of neuronal oscillations to behaviorally relevant parameters", NeuroImage, 86:111-122, 2014

S. Dähne, V. V. Nikulin, D. Ramirez, P. J. Schreier, K. R. Müller, S. Haufe, "Finding brain oscillations with power dependencies in neuroimaging data", NeuroImage, 96:334-348, 2014

S. Haufe, F. Meinecke, K. Görgen, S. Dähne, J. Haynes, B. Blankertz, F. Biessmann, "On the interpretation of weight vectors of linear models in multivariate neuroimaging", NeuroImage, 87:96-110, 2014

S. Haufe, S. Dähne, V. V. Nikulin, "Dimensionality reduction for the analysis of brain oscillations", NeuroImage, 101:583-597, 2014

% We assume X contains the continous EEG/MEG data with 
% columns of X corresponding to channels and rows to time points, 
% i.e. size(X) = [n_samples, n_channels]. 
% If you want to apply cSPoC, X is a cell array with X{n} being
% the n'th dataset. The preprocessing has to be applied to each 
% dataset individually


% First we call the SSD function which returns the 
% original data bandpassed and projected onto the SSD components
% Also, in order to compute the spatial patterns of extracted 
% (m/c)SPoC sources,  we need the covariance matrix of the 
% bandpassed sensorspace data.
% See SSD code for more info on the paramters

[W_ssd, ~, ~, C, X_ssd]  = ssd(X, bands, ...); 


% Dimensionality reduction by choosing only the first few 
% SSD components

n_ssd_components = 15; 
X_ssd = X_ssd(:,1:n_ssd_components);
W_ssd = W_ssd(:,1:n_ssd_components);


% Apply any of the (m/c)SPoC methods to get the SPoC spatial filters.

W_spoc = spoc(X_ssd, z);

% Note that the spatial filters are now defined in SSD space and have 
% to be mapped to sensor space to be applicable to the original data.

W = W_ssd * W_spoc;

% The spatial patterns are computed using the covariance matrix of the
% bandpass filtered sensor space data

A = W * C;