摘要
对国内外ERP测谎研究进行了总结和讨论。ERP测谎目前采用的编题方法有准绳问题测试法(CQT)和犯罪知识测试法(GKT);涉及的脑电成分主要为P300;研究中使用的统计指标有波幅、潜伏期、头皮分布等,但头皮分布能否成为一种新的测谎统计指标还没有定论;统计方法早先使用方差分析,但该方法难以进行单个人的鉴别,因此研究者现在多采用Bootstrap等来进行统计。最后,对目前ERP在测谎实践中所面临的困难、解决途径和应用前景等进行了论述。
The current work is a summary on ERP studies for detection of deception. The major paradigm used in this area are Control Question Test and Guilty Knowledge Test, and the major ERP component used is P300. Amplitude, latency and scalp profile are often used as statistical indexes in these reseaches. However, there are no verdict of profile as a novel index of the detection of deception. Because ANVOA, which used in early researches, can not do individual diagnostics, bootstrap method is utilized in latest researches. At last, the main obstacle, resolving methods and prospective view in practice are discussed.
出处
《临床医学工程》
2009年第6期105-108,共4页
Clinical Medicine & Engineering
基金
教育部人文社会科学重点研究基地重大项目(05JJDXLX003)
教育部百篇优秀博士论文作者专项资金项目(200708)
天津市人才发展资金项目
