I've been fiddling for some time now with the P300 demo, was able to run the classifier with an ok score. I decided to compare the LDA classifier with other ones, and since I'm more comfortable with writing code, I've decided to branch out of the designer and fiddle in a jupyter notebook.
To do so, I've taken the replay scenario of p300, and added a csv file writer box, the input links are the stimulations from the generic stream reader (which loads the file I've used to train the last classifier) and an epoched averaged signal.
I've successfully loaded the csv but I'm having trouble making sense of the data - there is an epoch column that from first glance correspond to the stimulation based epoching box.
This creates a new epoch after the OVTK_StimulationId_VisualStimulationStart is visible in the stream, it then does this for 0.6s.
The csv file contains these rows though
Code: Select all
74 23.626867 0 -61.872613 5.423805 5.336922 NaN NaN NaN
75 23.340117 1 122.415990 3.753208 3.470638 NaN NaN NaN
where we can see that the epoch has changed, but there are no stimulations tied to this time.
Here are the beginning rows that have a stimulation with them
Code: Select all
0 23.034867 0 -175.652395 2.798264 1.342050 32769:32777:33029:33036:32778:32773:32771:3328... 20.0460161110:20.0460662949:22.0655885558:22.0... 0.0000000000:0.0000000000:0.0000000000:0.00000...
25 23.234867 0 89.746432 -8.077858 -7.473044 32780 23.2392178809 0.0000000000
38 23.338867 0 122.415990 3.753208 3.470638 33286:33025:32779 23.3400682809:23.3401066861:23.3401165335 0.0000000000:0.0000000000:0.0000000000
62 23.530867 0 28.951199 2.703738 4.148936 32780 23.5359182237 0.0000000000
112 23.636117 1 -69.652697 1.132564 4.114039 33286:33032:32779 23.6380859825:23.6381306809:23.6381336607 0.0000000000:0.0000000000:0.0000000000
Kind regards,
Bartlomiej