patternrecognition / openbmi

An open software package dedicated for the development of Brain-Computer Interfaces with various advanced pattern recognition algorithms

OpenBMI ver 3.0 is released.

An open software package dedicated for the development of Brain-Computer Interfaces with various advanced pattern recognition algorithms

We provide detail information in each folder and every function.

1. 'Examples' : Basic example codes for MI, ERP, SSVEP are included.

-Folder : Examples :run_MI (example for the analysis of motor imagination) :run_ERP (example for the analysis of event related potential) :run_SSVEP (example for the analysis of SSVEP)

If you are interested in our code and want to know simple example code, 'Examples' will be helpful for your understanding.

2. 'BMI_modules' : Detail implementation functions are included.

-Evaluations : functions related to cross-validation -Functions : functions related to feature extraction -LoadData : functions related to data loading -Online : functions related to online connection -options : functions related to frame of data construction -Preprocessing : functions related to EEG pre processing -TCPIP : functions related to TCPIP connection -Training : LDA functions -Visulaization : functions related to visualizatoin -Developing : functions related to developing functions -etc

3. 'GUI module' : Graphic user interface functions are included.

-OpenMI2 : functions related to MI GUI module -OpenSSVEP : functions related to SSVEP GUI module -new_p300 : functions related to P300 GUI module

4. 'Paradigm' : Paradigm functions with Psychtoolbox are included:

-Artifact : functions related to artifact paradigm -MI : functions related to MI paradigm -SSVEP : functions related to SSVEP paradigm -Speller : functions related to p300 spller paradigm

5. 'GigaScience' : Our paper in GigaScience codes are included

6. 'PR_BCI_team' : Other BCI group's codes are included

7. 'VIS' : OpenBMI demo

8. 'Test_mat' : Test mat file

9. 'TNNLS' : Our proceeding TNNLS paper codes (paradigm and basic examples) are included

If you have any questions about our code, please leave any message. web site: http://openbmi.org

openbmi's Issues

PreProcessing: Common Average Reference

Rereferencing to a common reference

Arrange a folder: 'Options'

1. check the folder 'Options', delete unnecessary or duplicated file
2. make all files as regular format

Description:
Example Code:
Input:
Returns:

see the 'prep_addChannels.m' as example

regular form of functions

It needs to change all scripts to regular format as below:
delete one of function prep_removeClass or prep_removeClass2

prep_addChannels
prep_addFeatures
prep_addTrials
prep_average
prep_baseline
prep_cutDim
prep_defineClass
prep_deleteChannel
prep_erpMeans
prep_removeClass
prep_removeClass2
prep_resample
prep_segmentation
prep_selectChannels
prep_selectClass
prep_selectClasses
prep_selectTime
prep_selectTrials

example

example

opt_eegStruct

Block a variety of error messages by giving warning when:

1. variable 'field' is empty.
2. the variable 'field' has unacceptable value
   example:
   field=[] or
   field={'x','t','fs','y_dec','y_logic','y_class','class', 'chan', 'aaa', 'aefaef'};

consider other type of errors

opt_selectField

1. make the scripts into regular format as below
   Description:
   Example Code:
   Input:
   Returns:
2. preventing all case of error message

A

A

Emotive modules

The emotive modules for the offline analysis of Emotive data.
It has sub-modules of :

1. Dada Load, 2. Head Load, and 3. Marker Load.

Arranging a folder: 'funtions'

1. Check a folder 'Functions'
2. It needs to change all scripts to regular format as below:

Description:
Example Code:
Input:
Returns:

see the 'prep_addChannels.m' as example

Paradigms : Makeparadigm

Our Makeparadigm related to make MI paradigm need to change all scripts to regular format.

Also, in our version didn't upload make paradigm functions of SSVEP and P300.

Need to upload this function as soon as possible.

opt_eegStruct

1. make the scripts as regular format
2. delete the warning message "Warning: Fieldnames are not unique between structures. " at line 18
3. error message when the input variable 'field' is empty -> check all other cases

PreProcessing: Local Average Reference

Rereference all channels to local average

PreProcessing: Laplacian Filter

The surface laplacian is a spatial filter in which from the signal of each channel, the (spatially) averaged signals by subtracting neighbours channels

File name error

Hi,
the Gigascience/Analysis_ERP.m throws an error at the line 48 due to a call to erp_performance function which is located in the ERP_performance.m file. this typo is easily corrected by either changing the name of the function.

Matlab version: R2014a.

rsqu function

The function of rsqu.m in visualization folder need to modify or move to another folder such as etc.

I think that rsqu function looks like another toolbox's function. Need to write references of rsqu function.

Also, make the scripts into regular format.

Data description

Dear Developers!

I downloaded the data from http://gigadb.org/dataset/view/id/100542 but I miss some information about the .mat structure. Im investigating one of the *_EEG_MI.mat files:

• are the train and test sets recorded on the same day without removing the EEG system?
• what is the unit of x? (Which I supose is the raw eeg signal) Is it in Volt, miliVolt, microVolt or in an other region eg: 1e-7
• what is the EMG data? Are those EOG channels? If yes what are the positions?
• what are the following fileds: i_val, pre_test, post_test, rest_interval

Thanks in advance!
Csaba Köllőd

regular form of functions

It needs to change all scripts to regular format as below:

prep_commonAverageReference
prep_envelope
prep_filter
prep_laplacian
prep_localAverageReference
prep_movingAverage
prep_normalize
prep_rejectArtifactMAxMin
prep_spectrum

PreProcessing: Basic Data Structure Processing - selecting, appending, etc.

The basig EEG signal processing:

1. selecting channel/classes/trials/time interval
2. appending channel/classes/trials, and appeding each of continous EEG data

Revising the prep_spectrum function

Hi, YoungJin

If you do not require this function, you will remove this function. I think.
Please check the \etc\prep_spectrum function.

Pre-processing

Enhancement of functions in preprocessing module.

1. For 1-channel segmented data, don't show warning messages.
   functions: prep_baseline, prep_selectTime, prep_average
2. Make the function "prep_removeClass" work, when the number of classes is more than 3.
3. Error in using prep_average, when the dat.y_dec does not have 1. Fix it!

Arranging a folder: 'visualization'

Check a folder 'Functions'
It needs to change all scripts to regular format as below:
Description:
Example Code:
Input:
Returns:

see the 'prep_addChannels.m' as example

prep_selectClass

give an option 'all' for selecting all classes:
CNT=prep_selectClass(CNT,{'class',{'all'}});

give a description of the function as regular form; input, returns, etc.

PreProcessing: Common Median Reference

Rereference signals to common median reference

I want to the fold of SSVEP in cross validation?