Hybrid-BCI

Tutorial for data processing for hybrid BCI

The authors would be grateful if published reports of research using this code (or a modified version, maintaining a significant portion of the original code) would cite the following article:

Shin, Jaeyoung, et al. "Open Access Dataset for EEG+ NIRS Single-Trial Classification." IEEE Transactions on Neural Systems and Rehabilitation Engineering (2016), DOI: 10.1109/TNSRE.2016.2628057.

Caution

Data processing was performed using MATLAB 2013b. Please use MATLAB 2013b or older version. We also recommend you to download an old version of BBCI toolbox in the following link:

https://github.com/JaeyoungShin/hybrid-BCI/blob/master/bbci_public-master_TNSRE.zip

This zip file includes the updated version of 'proc_BeerLambert'. You may run into some errors if you use later version of BBCI toolbox or MATLAB.

Basic Data Structures

This document is from https://github.com/bbci/bbci_public/blob/master/doc/ToolboxData.markdown and partly modified. For more details, visit BBCI toolbox.

Table of Contents

• cnt - Data structure holding the continuous signals
• mrk - Marker structure defining certain events
• mnt - Montage structure defining the electrode layout for scalp and grid plots

cnt - Continuous signals (EEG)

The structure holding continuous (i.e., not epoched) EEG signals is denoted by cnt.

cnt	is a structure with the following fields:
.clab	channel labels (CELL {1 #channels}) there may be additional information in other fields, but these are all optional
.fs	sampling rate [samples per second]
.title	task type: 'MI' - motor imagery / 'MA' - mental arithmetic
.T	length of cnt.x
.yUnit	unit of cnt.x
.x	multichannel signals (DOUBLE [T #channels])

mrk - Event markers (EEG)

The structure holding marker (or event) information is denoted by mrk. Using this structure you can segment continuous EEG signals into epochs by the function proc_segmentation.

mrk	is a structure with the following fields:
.time	defines the time points of events in msec (DOUBLE [1 #events])
.event.desc	structure of further information; each field of mrk.event provides information that is specified for each event, given in arrays that index the events in their first dimension. This is required such that functions like mrk_selectEvents can work properly on those variables.
.y	class labels (DOUBLE [#classes #events])
.className	class names (CELL {1 #classes})

mnt - The electrode montage (EEG)

The electrode montage structure, denoted by mnt, holds the information of the spatial arrangement of the electrodes on the scalp (for plotting scalp topographies) and the arrangement of subplot axes for multi-channel plots.

mnt	is a structure with the following fields:
.x	x-position of the electrode for scalp maps (DOUBLE [1 +channels])
.y	y-position of the electrode for scalp maps (DOUBLE [1 +channels])
.pos_3d	3D coordinate (DOUBLE [3 #channels])
.clab	channel labels (CELL {1 #channels})

            | **further optional fields are required for multichannel plots:**

.box | positions of subplot axes for multichannel plots (DOUBLE [2 #channels] or [2 #channels+1]; the first row holds the horizontal, and the second row the vertical positions. The optional last column specifies the position of the legend .box_sz | size of subplot axes for multichannel plots (DOUBLE [2 #channels] or [2 #nchannels+1]), corresponding to .box. The first row holds the width, the second row the height .scale_box | position of subplot for the scale (DOUBLE [2 1]) .scale_box_sz | size of subplot for the scale (DOUBLE [2 1])

cnt - Continuous signals (NIRS)

The structure holding continuous (i.e., not epoched) NIRS signals is denoted by cnt.

cnt	is a structure with the following fields:
.fs	sampling rate [samples per second]
nSources	the number of sources
nDetectors	the number of detectors
multiplexing	the number of device used for data recording
.clab	channel labels (CELL {1 #channels}) there may be additional information in other fields, but these are all optional
.x	multichannel signals (DOUBLE [T #channels])
.wavelengths	wavelengths of near-infrared light (Double [1 #wavelengths])
.signal	type of signal
.title	task type: 'MI' - motor imagery / 'MA' - mental arithmetic
.yUnit	unit of cnt.x

mrk - Event markers (NIRS)

The structure holding marker (or event) information is denoted by mrk. Using this structure you can segment continuous NIRS signals into epochs by the function proc_segmentation.

The structure is same as above.

mnt - The electrode montage (NIRS)

The NIRS montage structure, denoted by mnt, holds the information of the spatial arrangement of sources, detectors, and physiological channels on the scalp (for plotting scalp topographies) and the arrangement of subplot axes for multi-channel plots.

mnt	is a structure with the following fields:
.source	holds the informaiton of the spatial arrangement of sources and channel labels
.detector	holds the informaiton of the spatial arrangement of sources and channels labels
.x	x-position of the physiological channel for scalp maps (DOUBLE [1 +channels])
.y	y-position of the physiological channel for scalp maps (DOUBLE [1 +channels])
.pos_3d	3D coordinate (DOUBLE [3 #channels])
.clab	channel labels (CELL {1 #channels}). e.g.) source : AF7, detector : Fp1 -> clab : AF7Fp1
sd	source - detector pair (Double [#channels 2])

       | **further optional fields are required for multichannel plots:**

.box | positions of subplot axes for multichannel plots (DOUBLE [2 #channels] or [2 #channels+1]; the first row holds the horizontal, and the second row the vertical positions. The optional last column specifies the position of the legend .box_sz | size of subplot axes for multichannel plots (DOUBLE [2 #channels] or [2 #nchannels+1]), corresponding to .box. The first row holds the width, the second row the height .scale_box | position of subplot for the scale (DOUBLE [2 1]) .scale_box_sz | size of subplot for the scale (DOUBLE [2 1])