Kihyun Kim1, Kabmun Cha1, Chunkee Chung2, Hyunchool Shin1
(1. Dept. of Electronic Engineering, Soongsil University, Seoul 156-743, Korea;2. College of Medicine, Seoul National University, Seoul 110-744, Korea)
Abstract：A method of estimating hand and elbow movements using electrocorticogram (ECoG) signals is proposed. Using multiple channels, surface electromyogram (EMG) signals and ECoG signals were obtained from patients simultaneously. The estimated movements were those to close and then open the hand and those to bend the elbow inward. The patients were encouraged to perform the movements in accordance with their free will instead of after being induced by external stimuli. Surface EMG signals were used to find movement time points, and ECoG signals were used to estimate the movements. To extract the characteristics of the individual movements, the ECoG signals were divided into a total of six bands (the entire band and the δ,θ,α,β and γ bands) to obtain the information entropy, and the maximum likelihood estimation method was used to estimate the movements. The results of the experiment show that the performance averages 74% when the ECoG of γ band is used, which is higher than that when other bands are used, and higher estimation success rates are shown in the γ band than in other bands. The time of the movements is divided into three time sections based on movement time points, and the “Before” section, which includes the readiness potential, is compared with the “Onset” section. In the “Before” section and the “Onset” section, estimation success rates are 66% and 65%, respectively, and thus it is determined that the readiness potential could can be used.
Key words：electrocorticogram(ECoG)；γ band；entropy；maximum likelihood estimation；readiness potential
CLD number： TP911.7 Document code： A
Article ID： 1674-8042(2012)04-0357-05 doi： 10.3969/j.issn.1674-8042.2012.04.012
References
[1] Ince N F, Goksu F, Tewfik A H. An ECoG based brain computer interface with spatially adapted time frequency patterns. International Conference on Bio-inspired Systems and Signal Processing, Portugal, 2008.
[2] Miller K J, Shenoy P, den Nijs M, et al. Beyond the gamma band: the role of high-frequency features in movement classification. IEEE Trans. on Biomedical Engineering, 2008, 55:1634-1637.
[3] YUAN Yuan, XU An-bang, GUO Ping, et al. ECoG analysis with affinity propagation algorithm. Proc. of the 4th International Conference on Natural Computation, 2008, 5: 52-56.
[4] LI Ming-ai, YANG Jin-fu, HAO Dong-mei, et al. ECoG recognition of motor imagery based on SVM ensemble. Proc. of 2009 IEEE International Conference on Robotics and Biomimetics (ROBIO), 2009: 1967-1972.
[5] ZHAO Hai-bin, YU Chun-yang, LIU Chong, et al. ECoG-based brain-computer interface using relative wavelet energy and probabilistic neural network. Proc. of the 3rd International Conference on Biomedical Engineering and Informatics(BMEI), 2010, 2: 873-877.
[6] Crone N E, Miglioretti D L, Gordon B, et al. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Event-related synchronization in the gamma band. Oxford: Brain, 1998, 121: 2301-2315.
[7] Menon V, Freeman W J, Cutillo B A, et al. Spatio-temporal correlations in human gamma bands. Electroencephalography and Clinical Neurophysiology, 1996, 98: 89-102.
[8] You K J, Shin H C. Classifying finger flexing motions with surface EMG using entropy and the maximum likelihood method. The Institute Of Electronics Engineers of Korea, 2009, SC:(6):38-43.
[9] Bercher J F, Vignat C. Estimating the entropy of a signal with applications. IEEE Trans. on Signal Processing, 2000, 48(6): 1678-1694.
