The Neurophysiological Biomarker Toolbox (NBT)

Getting started with EEG analysis in NBT

Version: Oct 2017

Approximate time needed to complete the tutorial: 20-30 min.

The Neurophysiological Biomarker Toolbox (NBT) is a software package running on MATLAB, which will help you analyze EEG data efficiently.

Aim of tutorial

In this tutorial you will learn how to

  • Install and start NBT
  • Load an example EEG signal
  • View the raw EEG signal and compute a power spectrum

Start with downloading NBT from Canvas(you can skip if you have already done these steps):

  1. Go to Pages / Computer practical and download
  2. A window will pop up which asks you whether you want to save or open the file. Click on open. The program 7-zip, which should be already set as the default program, will open your downloads.
  3. Once 7-zip opens, click on the button extract and save it on your C-drive, by typing C:\Temp in the window that pops up (creates the folder path “C:\Temp\NBT-stable” which you can then find in “Computer\System (C:)\Temp”).

If you have any problems in downloading, please ask a tutor.

How to install NBT in Matlab

Installing NBT is very easy!

  1. Start Matlab (skip this step if you have Matlab already opened)
  2. Set your Current folder to NBT folder
  3. Type 'installNBT' on the command line and press Enter

From this moment Matlab will embed all NBT functions. If you start matlab again you will have to repeat these steps

How to start NBT

  1. In the Matlab Command Window write: NBT ; Press “Enter”
  2. The NBT GUI will pop-up (see picture below).

How to load an NBTsignal, and plot the raw signal

To practice this tutorial you can use the 1 minute recording from an Eyes Closed Rest experiment called “Tutorial.Signal” which you need to download from Canvas→Pages→Computer practical.

Just as you did earlier, unzip, extract and save both files in a new folder in your C:\Temp, we suggest to call this folder NBT_test_files.

How to load

  1. Go to the menu File|Load NBT Signal.
  2. Find the file you want to load (i.e., HN13.S068.20130606.ECR.mat); click OK.
  3. At this moment you only have a RawSignal in the file, later you will have other versions of your signal, e.g., tempICASignal . If you have other versions of the signal in the file, you will be prompted to select which signal versions you want to load.
  4. Wait for the signal to load. The signal is loaded when you see in the Matlab Command Window “NBT Signal and Info Object loaded”.
  5. The NBT GUI will look similar to this:

How to plot the raw signal

  1. Go the the menu Visualization|Plot Signals
  2. A window pops up: In the field “distance between signals” fill in “40” and click “OK”
  3. You now see a plot of the raw EEG signal from all channels and for the whole time range. The EEG signal is “raw” in the sense that it is not filtered or cleaned for artifacts yet.
  4. By using the zoom buttons in the menu bar you can select an area you want to zoom in to or back out of.
  5. Scroll through the signal using the scroll bars in the left bottom corner, or select channels to plot by using the Select channel(s) to plot button. If you cannot see the scroll buttons, try double-click the edge of the viewing window.
  6. After exploring these plotting tools your window might look like the figures below.
  7. Typically, the strongest signal in eyes-closed EEG is the alpha oscillations (8-13 Hz). Try to find a channel with a clear alpha signal. Question: What is the frequency of this subject's alpha oscillations? Did your neighbor arrive at the same value?
  8. For your report or poster you need to make screen dumps of EEG signals. Press Windows button Print Scrn on the keyboard, and an image of your screen is now on the clipboard. Go to a program like Paint, and Ctrl V (paste) the image and crop the part you want to keep.
  9. Alternatively, you may go to “File” in the Matlab window and export the figure in a format that can be edited in another software, like Adobe Illustrator.

You can read more about The NBT signal viewer.

How to plot a power spectrum

A power spectrum is used to quantify the power of different frequency components of a signal. You can use it to, e.g., identify the peak frequency or peak power in the alpha band (8-13 Hz).

  1. Go to Visualization|Plot time-frequency plot and power spectrum of one channel
  2. Click “OK”.

A window will pop up (see figure below). In the top you see a time-frequency plot for one channel. In the bottom left you see the power spectrum for that same channel. In the bottom right you can select other channels (either from the list or the scalp map) that you want to see a time-frequency plot and a power spectrum for.
Question: Can you identify the peak frequency of the alpha oscillations?
Question: Is it equal to the frequency you found when plotting the signal above?

Later in the course, or if you have time now, it is recommended to play with the other plotting parameters to see the possibilities for a nice visualization of your signal and its time-frequency characteristics. For example, choosing wavelet (w) instead of multitaper and decreasing the 'frequency step' size will make a more smooth TF plot.

This is a raw signal, which means it probably still contains artifacts (i.e. signals not from the brain) that will disturb the analysis. In the next tutorial you will learn how to remove these artifacts.

Go to the next tutorial How to remove bad channels and transient artifacts

tutorial/getting_started_with_eeglab_and_nbt.txt · Last modified: 2017/10/30 14:57 by Klaus Linkenkaer-Hansen
The NBTwiki platform - version 2.8 - 9 May 2013
Copyright (C) 2008-2015