This plugin provides the following functions:

1. Merge Primary with Each Secondary SDM
   Merge ONE design matrix containing the task predictors of one study (functional run) (see field Primary SDM) with MULTIPLE design matrices containing confound predictors for the same study (see field Secondary SDMs).
   These secondary SDMs can contain different kinds of predictors as for example motion estimates, temporal filter predictors or physiological noise predictors.You can select as many secondary SDMs as you wish, but the number of primary SDMs is restricted to one.
   Result: There will be as many output SDMs saved to the source folder as there are secondary SDMs available. The file names of the resulting SDMs can be freely chosen by the user or the suggested default file names can be confirmed. Each of the output SDMs will be a combination of the task predictors in the primary SDM extended with the confound predictors of one of the secondary SDMs.
   
2. All predictor correlations from Primary SDM (CSV-file):
   Computes the predictor correlations of the primary SDM and saves them in a CSV file (use the Start button). The resulting CSV file name is identical to the input SDM name.
   This option is important to check for a potential problem with multicollinearity in your design matrix. Multicollinearity occurs when two or more predictors are highly correlated and hence one predictor, in that case a task predictor, can be explained by one or a linear combination of the other predictors. In that case there would be a linear dependence of the regressors in the model and the results of the GLM would not be easy to interpret, leading to possible misinterpretations of certain activations.
   
3. Merge All Secondary SDMs
   Merge MULTIPLE design matrices containing confound predictors of the same study (functional run) into ONE single SDM file (see field Secondary SDMs). This provides for example the option to merge motion estimate predictors with physiological noise predictors.
   Result: There will be ONE output SDM saved to the source folder containing all confound predictors of the single secondary SDMs available. 
   

The following options apply to the functions described in 1. and 3.:

Z-transform Confounds:
After merging SDM files all confound predictors can be z-transformed. In general, normalization of predictors in the design matrix is performed to obtain the same scale for all predictors and assure the interpretability of the resulting beta weights. For unbalanced designs it can be advantageous to only z-transform the confound predictors of the model to avoid potential problems when z-normalizing model predictors based on shorter event length.

Include Baseline:
When no 'Constant' predictor is found in the resulting SDM file, one can be added automatically as the last predictor. If this option is not selected by the user, the constant is not included in the resulting SDM, i.e. the constant is removed when one is found in the input SDMs.

All Predictor Correlations of Resulting SDMs (CSV file):
For each resulting SDM file an additional CSV file, containing the predictor correlations, is saved to disk using the same name as the resulting SDM file.