-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathreadme.txt
More file actions
56 lines (45 loc) · 4.39 KB
/
readme.txt
File metadata and controls
56 lines (45 loc) · 4.39 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Help document for eBook Reader, BCI Practical 2014 project.
QuickStart
----------
Note: .bat files - Windows machines
.sh files - Linux/Mac machines
To run the eBook Reader:
1) Start a buffer by running: dataAcq/startBuffer.bat or buffer/startBuffer.sh
(optional)
(1.1) If you don't have a measurement system connected, start a simulated data source by running:
dataAcq/startSignalProxy.bat or dataAcq/startSignalProxy.sh
2) Start Matlab based signal processing script by running:
eReader/startSigProcBuffer.bat or eReader/startSigProcBuffer.sh
3) Start Matlab based experiment control & stimulus presentation script by running :
eReader/runReader.bat or runReader.sh
4) BCI Controller will pop-up with the following function keys
4.1 Subject Name -- Type in the subject's name in the experiment control window, and then run through each of the following experiment phases:
4.2 CapFitting -- Check for the quality of the data/brain activity being recorded by the electrodes. This will show a topographic plot of the head with the electrodes coloured from red=bad to green=good. Add additional gel/water or adjust ground until they are satisfactorily not red!
EEG -- Real-time EEG viewer to check electrode connection quality. This shows a topographic arrangement of the electrodes with the current (filtered) signal in each electrode. If you have a well connected set of electrodes you should be able to see eye-blinks in the more frontal electrodes, and muscle artifacts (such as jaw clenching) in all of the electrodes.
4.3 Test Run -- Practice the task to be used in the BCI. A red fixation cross cues the user to get ready and lasts for 1 second before presenting one of the following 5 cues for calibration
R - Right -> Right-hand movement -> navigate to next page
L - Left -> Left-hand movement -> navigate to the previous page
U - Tongue -> Tongue movement -> increase fontsize
D - Toes -> Both toe movement -> decrease fontsize
keep clam -> Do nothing -> to represent the reading state
4.4 Calibrate -- Get calibration data by performing the task as instructed in 2 blocks, each lasting about 4 minutes
4.5 Train Classifier-- Train a classifier using the calibration data. 3 windows will pop-up showing: Per-class ERsPs, per-class AUCs, and cross-validated classification performance. Close the classification performance window to continue to the next stage.
4.6 Testing -- test the trained classifier in the following two ways
a. Prompt test - Stimulus presentation is similar to the calibration phase but now, even the classifier feedback is given. This lasts for less than 3 minutes and gives one a general idea of how "efficient" the classifier is.
b. Reader - Presents a story with 20 pages giving feedback (changing the pages/fontsize) every second to simulate a reading experience.
File List
---------
General Setup/GUI Files:
configReader.m -- basic configuration variables for the reader to specify various parameters like stimulus presentation time, colour, size and what characters to present etc.
runReader.m runReader.sh runReader.bat -- Controller functions to run the reader stimulus and experiment control
controller.m controller.fig -- files to generate the GUI and function call-backs for the experiment control
startSigProcBuffer.m startSigProcBuffer.bat -- control functions to run the various signal-processing functions as requested by the runReader.m experiment controller.
eegMob_Ch10_10-20.png -- an image file for cap fitting reference for a 10 channel tmsi mobita setup.
cap_tmsi_mobita_reader10 -- text file with 10 electrodes setting for mobita
cap_tmsi_mobita_reader32 -- text file with 32 electrodes setting for mobita
Experiment Phase Files:
readerCalibrateStimulus.m -- generate the calibration phase stimulus, i.e. show fixation, cued targets etc.
readerEpochFeedbackStimulus.m -- generate stimulus for the testing phase, where it cues the user to perform an action and gives the classier feedback
readerEpochFeedbackSignals.m -- makes classifier prediction and gives appropriate feedback values for every trial
readerNeuroFeedbackStimulus.m -- generate the stimulus for the Reader feedback phase - makes a story with 20 pages for testing purposes.
readerContFeedbackSignals.m -- makes classifier prediction and gives feedback every second to the readerNeuroFeedbackStimulus.m