She-ping TIAN(田社平), Cheng PAN(潘城), Guo-zheng YAN(颜国正)
Dept. of Information Measurement Technology and Instruments, Shan ghai Jiaotong University, Shanghai 200240, China
Abstract-Independent component analysis (ICA) is a widely used method for blind source separation (BSS). The mature ICA model has a restrictio n that the number of the sources must equal to that of the sensors used to colle ct data, which is hard to meet in most practical cases. In this paper, an overde termined ICA method is proposed and successfully used in the analysis of human c olonic pressure signals. Using principal component analysis (PCA), the method es timates the number of the sources firstly and reduces the dimensions of the obse rved signals to the same with that of the sources; and then, Fast-ICA is used t o estimate all the sources. From 26 groups of colonic pressure recordings, sever al colonic motor patterns are extracted, which not only prove the effectiveness of this method, but also greatly facilitate further medical researches.
Key words-medical signal processing; overdetermined ICA ; PCA; colonic motor pattern
Manuscript Number: 1674-8042(2010)04-0401-05
dio: 10.3969/j.issn.1674-8042.2010.04.22
References
[1]A. Neumann, M. Grosse-Wentrup, M. Buss, et al, 2008. The effect of mutual information on independent component analysis in EEG/MEG analysis: A simu lation study. International Journal of Neuroscience, 118(11): 1534-1546.
[2]D. Mantini, R. Franciotti, G. L. Romani, et al, 2008. Improving MEG source localizations: An automated method for complete artifact removal based on independent component analysis. NeuroImage, 40(1): 160-173.
[3]M. Joho, H. Mathis, R. H. Lambert, 2000. Overdetermined Blind Source Separation: Using More Sensors Than Source Signals in a Noisy Mixture. ICA2000, Helsinki, Finland, p. 81-89.
[4]M. Camilleri, A. E. Bharucha, C. Di Lorenzo, et al. 2008. American n eurogastroenterology and motility society consensus statement on intraluminal me asurement of gastrointestinal and colonic motility in clinical practice. Neurogastroenterology and Motility, 20(12): 1269-1282.
[5]D. Liu, Y. H. Song, F. Wang, 2005. Over-determined Blind Signal Sep aration Based on PCA and Infomax. Proceeding of 2005 Chinese Control and Decisio n Conference. Shenyang, p. 538-541.
[6]A. Hyvarinen, 1999. Fast and robust fixed-point algorithms for ind ependent component analysis. IEEE Transactions on Neural Networks, 10(3): 626-634.
[7]A. Hyvarinen, J. Karhunen, E. Oja, 2001. Independent Component Analy sis. John Wiley & Sons, Inc. New York.
[8]M. M. Schuster, 2002. Atlas of Gastrointestinal Motility in health a nd disease 2nd ed. et al Decker Inc..
[9]G. Bassotti, U. Germani, C. Calcara, 1996. Twenty-four hour manomet ric recording of colonic motility: A technique with both research and clinical i dentities. Techn Coloproctol, 4: 142-144.
[10]S. S. C. Rao, P. Sadeghi, J. Beaty, et al, 2001. Ambulatory 24-h c olonic manometry in healthy humans. American Journal of Physiology-Ga strointestinal and Liver Physiology, 280(4 43-4): G629-G639.
[11]P. G. Dinning, M. M. Szczesniak, I. J. Cook, 2008. Proximal colonic propagating pressure waves sequences and their relationship with movements of c ontent in the proximal human colon. Neurogastroenterology and Motility , 20(5):512-520.
[12]G. Bassotti, G. de Roberto, D. Castellani, et al. 2005. Normal aspe cts of colorectal motility and abnormalities in slow transit constipation. World Journal of Gastroenterology, 11(18), 2691-2696.
[13]B. A. Orkin, R. B. Hanson, K. A. Kelly. 1989. The rectal motor comp lex. Journal of Gastrointestinal Motility, 1(1): 5-8.
[full text view]
