Graphic User Interface

Once MVPAlab is installed, the graphic user interface can be launched by typing the following command in the MATLAB command line:

>> mvpalab

The initial MVPAlab window should appear as shown in the following figure:

Using this interface, users can create new studies, open previously created projects or open the plotting utility.

1. Create new study:

Creating new studies in MVPAlab using the GUI is very intuitive. Users only need to specify (1) the type of analysis required from the dropdown menu and (2) select the study location folder. Results, configuration and other study-related files will be hierarchically stored in this directory. Once everything is selected, clicking the configuration button (3) will create the project folder structure and launch the analysis configuration window (4).

2. Configure the analysis:

General configuration:

A detailed description of the following parameters and techniques can be found in Analysis configuration section.

Before computing a decoding analysis, additional details of configuration are required. Users have to specify the locations of the epoched datasets (5) and label each condition with a condition identifier (6). All the relevant parameters of the decoding analysis are set to its default value and can be modified within this configuration window. These configuration parameters include a wide range of processes that can be executed during the decoding analysis, such as: (7) analysis timing, (8) feature extraction, (9) trial averaging, (10) balance datasets, (11) dimensionality reduction, (12) data smoothing, (13) data normalization or enable the computation of the temporal generalization matrix (14).

Classification model configuration:

A detailed description of the following parameters and techniques can be found in Analysis configuration section.

Additionally, the employed classification models can also be designed here. Users can choose between (15) different classification algorithms, (16) kernel functions, (17) cross-validation strategies, (18) model optimization, and (19) select several output performance metrics. You can also activate parallel computation (20).

Statistical analysis configuration:

A detailed description of the cluster-based permutation approach can be found in the Statistics section.

Finally, users can (21) enable and configure a non-parametric cluster-based permutation analysis, which is necessary to draw statistical inferences at the group level. The total number of permutations (at a group (22) and subject (23) level) as well as the above and below chance significance levels (for (24) performance and (25) cluster size) can be set in this configuration windows.

3. Compute the decoding analysis:

Once the configuration parameters are correctly specified, the computation of the multivariate analysis can be started by clicking the Start analysis button (26). Depending on the size of the dataset and the selected configuration, this process may be time-consuming and CPU/memory demanding. Anyhow, during the computation of the entire analysis pipeline, MVPAlab prompts in the MATLAB command window detailed information of the processes being executed (27).

4. Plot the result:

For the graphical representation of the results, MVPAlab also offers an intuitive plot utility that can be opened by clicking on (28) Open plot utility button. This tool enables users to open (29), plot (30), combine and compare (31) results of different analyses without dealing with cumbersome lines of MATLAB code. The most common configuration parameters such as (32) titles, (33) font size, (34) axes limits, (35) axes labels, or (36) color palettes, can be easily configured using this tool. Other parameters (37) such as line styles, transparencies, data smoothing or highlighting can be configured for time-resolved analysis, temporal generalization matrices, frequency contribution analyses, and others.