Search

The electrocorticogram (ECoG) is a kind of signal source that can be classified for making use of a human brain computer interface (BCI) field. The feature extraction is crucial for increasing classification accuracy rate. In this paper, Power Spectral Density is used for the selection of the optimal electrodes. Common spatial pattern (CSP) algorithm is used for feature extraction, and the nonlinear classification of motor imagery with support vector machines (SVM).The classification accuracy rate of 83% is achieved on Data set I of BCI Competition III.

The occurrence of acute hypotensive episodes (AHE) in intensive care units (ICU) seriously endanger the lives of patients, and are depended mainly on the expert experience of doctors to treat currently. How to detect and predict AHE in advance has become a clinical problem which is highly paid attention to by the medical world. In this paper, the theory of medical Informatics has been applied to achieve the prediction of the occurrence of AHE. In our study, the mean arterial blood pressure (MAP) signals of those who experienced AHE and those who not are both described on different scales by using wavelet transform. Through the extraction of the median and maximum from the wavelet coefficients for learning and training based on support vector machine (SVM), a predicting model with a predictive accuracy of 90% has developed. The experiment demonstrates that the proposed approach has better prediction accuracy comparing with the one which extracts statistical parameters from the MAP signals directly, and is beneficial to early prediction of AHE and intervention.