论文部分内容阅读
使用机器学习方法分类功能磁共振成像(f MRI)数据已经逐渐被应用到探索大脑认知的研究中。研究大脑功能区之间的相关性,对于理解大脑的工作方式乃至探究精神疾病的病理机制具有重要意义。使用不同眼睛状态和偏头痛疾病这两组静息态功能磁共振成像(rs-f MRI)数据,计算90个脑区间的相关性,构建出相应的脑功能相关系数矩阵。利用SVM-RFE特征选择方法筛选出具有显著特征性的脑区,并以相应的脑区功能相关系数矩阵为分类特征进行SVM分类。这两组实验均得到较高的分类准确率,并对提取出的特征进行分析,发现这些特征可以体现实验数据之间的大脑功能差异,并且在大脑认知上提供有益的探索。
The use of machine learning methods to classify functional magnetic resonance imaging (f MRI) data has gradually been applied to the exploration of brain cognition research. Studying the correlation between brain function areas is of great significance for understanding the working methods of the brain and even exploring the pathological mechanism of mental illness. The correlation between 90 brain regions was calculated using resting-state functional magnetic resonance imaging (rs-f MRI) data of different eye states and migraine headaches, and the corresponding matrix of the correlation coefficient of brain function was constructed. SVM-RFE feature selection method was used to screen the brain regions with significant features. SVM classification was carried out by using the corresponding functional matrixes of brain regions as the classification features. Both of the two experiments got higher classification accuracy, and analyzed the extracted features. These features can reflect the difference of brain function between experimental data and provide useful exploration in brain cognition.