| Description: | Problem from BCI III competition, Dataset 1. Data set provided by University of Tubingen, Germany, Dept. of Computer Engineering (Prof. Rosenstiel) and Institute of Medical Psychology and Behavioral Neurobiology (Niels Birbaumer), and Max-Planck-Institute for Biological Cybernetics, Tubingen, Germany (Bernhard Scholkopf), and Universitat Bonn, Germany, Dept. of Epileptology (Prof. Elger).
The classes are imagined movements of either the left small finger or the tongue. The data are from an 8x8 ECoG platinum electrode grid which was placed on the contralateral (right) motor cortex. All recordings were performed with a sampling rate of 1000Hz. the recording intervals started 0.5 seconds after the visual cue had ended and lasted 3 seconds (series length 3000). There are 8x8=64 dimensions. The training data and test data were recorded from the same subject and with the same task, but on two different days with about 1 week in between. The train data (278 cases) is from the first session, the test data (100 cases) from the second session.
There were 27 entries. The median accuracy was 79%. Ther best submitted solution obtained 91% accuracy on the test data. Features: Combination of Bandpower, CSSD or Waveform Mean, Fisher Discriminant Analysis. Classification: linear SVM
Reference: Thomas Lal, Thilo Hinterberger, Guido Widman, Michael Schroder, Jeremy Hill, Wolfgang Rosenstiel, Christian Elger, Bernhard Scholkopf, Niels Birbaumer. Methods Towards Invasive Human Brain Computer Interfaces. Advances in Neural Information Processing Systems (NIPS), 2004 |
