Full analysis with one script

This tutorial explains how to use the Brainstorm scripting interface to run a full analysis, from the raw recordings to the source reconstruction. It is based on a median nerve stimulation experiment recorded at the Montreal Neurological Institute in 2011 with a CTF MEG 275 system. The sample dataset contains 6 minutes of recordings at 1200Hz for one subject and includes 100 stimulations of each arm.

The tutorial follows the analysis steps detailed in the three advanced tutorials in the category Processing continuous recordings. You should read them before reading this tutorial, to have the explanations that go with the analysis steps.

Creating the analysis pipeline

Select the menu File > Create new protocol. Name it "TutorialScript" and select the options:

• "No, use individual anatomy",

• "Yes, use one channel file per subject".

To start building your analysis pipeline, just click on the "Run" button in the Process1 tab. Then add all the processes listed below. The output of each process is the input of the following one, this is why they cannot necessarily be shuffled.

Import anatomy > Import FreeSurfer folder

• Subject name: Subject01
• Folder to import: Select the folder sample_raw/Anatomy
• Number of vertices (cortex): 15000
• Fiducials: Copy what is indicated below. This is a reason it is usually easier to do this step in interactive mode, and then run only the script starting from the next step.
• Input: None; Output: None

Import recordings > Create link to raw file

Input: None; Output: Raw file

Pre-process > Sinusoid removal (notch)

Input: Raw file ; Output: Raw file (new)

Artifacts > Detect eye blinks

Input: Raw file ; Output: Raw file

Artifacts > Compute SSP: eye blinks

Input: Raw file ; Output: Raw file

Import recordings > Import MEG/EEG : Events

Input: Raw file ; Output: 199 epochs in 2 conditions

Pre-process > Remove DC offset

Input: 199 epochs ; Output: 199 epochs

Sources > Compute noise covariance

Since the epochs are currently selected and pre-processed: we can use them to estimate the noise covariance matrix before we move on with the calculation of the average.

Input: 199 epochs ; Output: 199 epochs

Average > Average files

Input: 199 epochs ; Output: 2 averages

File > Save snapshot: Sensors/MRI registration

Input: 2 averages ; Output: 2 averages

File > Save snapshot: Recordings time series

Input: 2 averages ; Output: 2 averages

Sources > Compute head model

Input: 2 averages ; Output: 2 averages

Sources > Compute sources

Input: 2 averages ; Output: all the source files (1 raw + 2 average + 199 epochs = 202 files)

Save the pipeline

Save in current workspace

Export as script

Process report