作者: Z.J. Koles
DOI: 10.1016/0013-4694(91)90163-X
关键词: Signal processing 、 Computer science 、 Qualitative analysis 、 Brain mapping 、 Electroencephalography 、 Pattern recognition 、 Artificial intelligence 、 Set (psychology) 、 Principal component analysis
摘要: A method is described which seems to be effective for extracting the abnormal components from clinical EEG. The approach involves use of a set spatial patterns are common recorded and 'normal' EEGs can account maximally different proportions combined variances in both EEGs. These factors used decompose EEG into orthogonal temporal wave forms judged by expert electroencephalographer abnormal, normal or artifactual origin. original then reconstructed using only principal component analysis present topography components. effectiveness discussed along with its value localization sources. It suggested, conclusion, that may optimal interpretation since it allows what best terms quantitative available qualitative analysis.