Data set I

Group information:

Department of Biomedical Engineering

Zhejiang University

Hangzhou, 310027, P. R. China

Team Member: Mao Dawei. (Email: maodawei@163.com, bmebrain@zju.edu.cn ). Ke Daguan, Xie Mingqiang, Ding Jichang, Zheng Kening, Zhou Jie, Murat.

Counselor: Tong Qinye. Zhang Hong. Huang Hai

Algorithm description

Feature: Data from every electrode was transformed to time-frequency domain by Hilbert-Huang Transform. Extract Standard deviation of every time-frequency window as feature.

Classification method: mahalanobis distance to class center.

Processing Flow:

Hilbert-Huang Transform of ECoG from all electrodes.

Extract Standard deviation of every time-frequency window (5Hz*0.2S) as feature.

Calculated error rate by leave-one-out use train data. Select 7 electrodes (29,30,31,38,39,40,46) that has lowest error rate with single time-frequency window.

Use Standard deviation of time-frequency domain from all 7 electrodes, single frequency band and 11 continuous time windows as feature. Calculated class center of class 1 and calss-1 from train data.

Use mahalanobis distance to classify the test data.

Data format

Format: matlab format(matlab 6.1)

Name: resultofdataI.mat