EEG-Annotate

A tool to identify particular response in unknown continuous EEG signals

EEG-Annotate is freely available under the GNU General Public License.

Please cite the following publication if using:

EEG-Annotate: Automated identification and labeling of events in continuous signals with applications to EEG
Kyung-min Su, W. David Hairston, Kay Robbins Journal of Neuroscience Methods Volume 293(1), Jan 2018, Pages 359-374. http://www.sciencedirect.com/science/article/pii/S0165027017303655

The data used to test EEG-Annotate is described in the data note:

An 18-subject EEG data collection using a visual-oddball task, designed for benchmarking algorithms and headset performance comparisons Kay Robbins, Kyung-min Su, W. David Hairston Data-in-Brief http://www.sciencedirect.com/science/article/pii/S2352340917306285

The actual data is available in three formats in NITRC

Visually Evoked Potential EEG https://www.nitrc.org/projects/vep_eeg_raw

To run example scripts (should work with MATLAB versions 2014a+)

- add path to EEGLAB (should work with versions 13.5.4b+)
- run EEGLAB, if clean_rawdata plugin is not installed, install it (version 0.31)
- add path to this AnnotateTools with Subfolders
- update data paths in an example script
- run the example script

Example scripts (under the examples subdirectory):

runBatchClassifier.m  runs a specified classifier for specified  
                      classes for all test files in a directory  
                      to produce a file containing a scoreData structure.    
   
runBatchAnnotate.m    runs the annotation for all of the files  
                      in a specified directory and produces a  
                      corresponding file containing an annotData  
                      structure.  The input files should each contain  
                      an appropriate scoreData structure.  

runReport.m           reads a directory of files, each of which contain  
                      an annotData structure and computes the  
                      performance metrics and bootstraps for statistical  
                      significance.  

Data preparation:

Before running the annotation pipeline, you must prepare the data to have power features if you are using VEP as training. If you have your own data for training you can prepare both your training and test data using the same features.

Releases:

Version 1.0.6 Released 11/27/2017

• Modified the batch comparison naming

Version 1.0.5 Released 11/15/2017

• Added publication information to README
• Added additional documentation to various functions
• Added getSampleTiming in preparation for report refactor
• Added reportComparison to compare two different annotations

Version 1.0.4 Released 10/18/2017

• Revised the parameter names for computing power features
• Added covariance features

Version 1.0.3 Released 10/03/2017

• Added non-parametric bootstrap test for statistical significance
• Began verifying package works for versions later than 2014a

Version 1.0.2 Released 09/19/2017

• Fixed ARRLS to have correct parameter settings

Version 1.0.1 Released 08/31/2017

• Added getAnnotateVersion
• Added powerFeatures aned batchPowerFeatures for consistent process
• Cleaned up some of the header documentation

References:

• EEGLAB: https://sccn.ucsd.edu/eeglab/

• ESS: http://www.eegstudy.org/

• EEG-Clean-Tools: http://vislab.github.io/EEG-Clean-Tools/

• ARRLS:

Adaptation regularization: A general framework for transfer learning
M. Long, J. Wang, G. Ding, S. J. Pan, and P. S. Yu
IEEE Trans. Knowl. Data Eng., vol. 26, no. 5, pp. 1076-1089, May 2014

• ARRLSimb:

Adaptive thresholding and reweighting to improve domain transfer learning for unbalanced data: With applications to EEG imbalance
K. Su, W. D. Hairston, and K. A. Robbins
15th IEEE International Conference on Machine Learning and Applications, 2016

Support:

This research was sponsored by the Army Research Laboratory and was accomplished under Cooperative Agreement Number W911NF-10-2-0022. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.