Machine learning for real-time single-trial EEG-analysis: from brain-computer interfacing to mental state monitoring
|Machine learning for real-time single-trial EEG-analysis: from brain-computer interfacing to mental state monitoring|
|Authors:||Klaus-Robert Müller, Michael Tangermann, Guido Dornhege, Matthias Krauledat, Gabriel Curio, Benjamin Blankertz|
|Citation:||Journal of Neuroscience Methods 167 (1): 82-90. 2008 January|
|Database(s):||Google Scholar cites PubMed (PMID/18031824)|
|Web:||Bing Google Yahoo! — Google PDF|
|Article:||BASE Google Scholar PubMed|
|Restricted:||DTU Digital Library|
|Format:||BibTeX Template from PMID|
|Extract:||Talairach coordinates from linked PDF: CSV-formated wiki-formated|
Machine learning for real-time single-trial EEG-analysis: from brain-computer interfacing to mental state monitoring reviews experiments with brain-computer interface and mental state monitoring with electroencephalography.
 Mental state monitoring
For mental state monitoring arousal is regarded by monitoring a subject performing a monotonuous task. The varying degree of errors the subject made through time was smoothed to an error index and then regarded as a proxy for "concentration". EEG trials were labeled into two classes depending on the error index, and the labels were then used in a linear discriminant analysis classifier with the processed EEG data as features. The processing used a filter for the alpha band and computed power features for 2 seconds windows for each of the 128 electrodes, - except for those electrodes excluded dues to base impedance.