General Linear Model: Multi-Study, Multi-Subject
Rules for automatic generation of useful GLM file names
The Multi-Subject, Multi-Study GLM dialog automatically creates a useful name for a calculated GLM depending on the selected options, including those from the Multi-Study GLM Options dialog. The name is updated on the fly when changing GLM options. Before running the GLM, the produced name can be adjusted, if necessary. Individual parts of the generated name are separated by underscores.
Experiment name. The string "AuditoryExperiment_N-10_RFX_PT_ITHR-100.glm" is an example of an automatically generated GLM file name. The first part "Auditory Experiment" is the base name, which will be extracted from the first functional file in the study list as the substring between the first underscore (separating usually a subject ID) and the second underscore (usually separating preprocessing and/or spatial normalization info, e.g. "_TAL"). If the extraction of an experiment name is not possible, the name will be set to "Model" as default.
Type of functional data. The next entry specifies the type of the functional data used to calculate the GLM. The string "STC" is used for FMR-STC multi-study GLMs, the string "VTC" is used for VMR-VTC multi-subject / multi-study GLMs and the string "MTC" is used for SRF-MTC cortex-based aligned multi-subject / multi-study projects.
Mesh hemisphere. Only in case of a "MTC" GLM, the program parses the file name of the current mesh in the surface module . If it finds the substring "_LH","_RH" or "_LHRH" in the mesh file name, it will append it to the GLM file name. This avoids that a GLM run for the left hemisphere will be overriden by a GLM run for the right hemisphere (or vice versa). In order to work correctly, you need to use the hemisphere identifier in your mesh file names (or use the default names, which should come with these labels).
Number of studies. The next entry "N-<value>" specifies the number of functional runs ("<value>") included in the GLM analysis. In a separate subject GLM (e.g. RFX-GLM), this number corresponds to the number of subjects times the number of studies per subject.
GLM type. The next substring codes for the type of GLM used: "RFX" appears in case that the RFX-GLM option is selected. This option implies a separate subjects GLM. If this option is not selected but the Separate subject predictors option, the substring "SPSB" will be shown. If the Separate study predictors option is selected, the substring "SPST" will be used. If neither of these three options is selected, the substring "FFX" will appear.
Time course normalization. The next substring indicates the selected type of time course normalization (if specified). If percent signal change normalization is selected, the substring "PT" will be shown. If z transformation is selected, the substring "ZT" will appear. If the Use only baseline segments for z-transformation has been selected in the Multi-Study GLM Options dialog, the substring "ZTB-<value>" will be shown, with "<value>" replaced by the chosen threshold value for definition of "baseline" time points.
Correction for serial correlations. If the Correct serial correlations option has been selected, the substring "AR-1" is shown next.
Spatial masking. If the Enable mask restriction option has been selectedin the Multi-Study GLM Options dialog, the substring "MSK-<file>" will be shown with "<file>" replaced by the name of the selected mask file.
Intensity masking. If a value greater than zero (default: "100") is provided in the Intensity threshold box in theMulti-Study GLM Options dialog, the substring "ITHR-<value>" is shown indicating that voxels with intensity values not surpassing this threshold in all runs are not included in the GLM analysis.