Algorithms Description:
1. Preprocessing signals with CAR(common average referenced) EEG re-referenceing method.
2. extracting involved rhythms with three linear band filters(7-13Hz, 11-17Hz, and 18-22Hz) and three continuous wavelet filters(DB2,  DB4, SYM4), totaly six filters.
3. Calculating the variances of the output signals from each filters
4. Selecting the optimal ones for classification features by GA(Genetic Algorithm)
5. Training a SVM classifier.
6. classifing the test data with a sliding window with length 4 seconds and step by 0.5 seconds.

Artificial Datasets are from:
subject c d f.

Result Data Format:
ASC II file (named 'Result_BCIC_IV_ds1*.txt') for each subject, for example, file with name 'Result_BCIC_IV_ds1a.txt') containing classifier outputs (integer number -1, 0, and 1) for each sample point of the evaluation signals of subject a, one value per line. 

Names and Affiliations:
CHEN Guangming [1] EMAIL TO: ctycheer@yahoo.com.cn
WU Jin [2]   EMAIL TO:wujin03@gmailcom
ZHANG Jiacai [1]   EMAIL TO: jiacai.zhang@bnu.edu.cn
[1]. College of Information Science and Technology, Beijing Normal Unviersity
[2]. National Key Laboratory for Cognitive Neuroscience and Learning, Beijing Normal Unviersity