A new method of discrimination between non-guilty and guilty subjects based on the autonomic response pattern in polygraph test

本研究探索了一种基于自主神经反应模式的测谎实验中区分无罪与有罪的新方法。以皮肤电活动、呼吸运动和体脉图作为测谎指标。然而，皮肤电活动趋于习惯性，体脉图不能准确反映交感神经系统的活动水平。我们将归一化脉搏体积(NPV)作为一种新的自主神经指标引入到多导生理测试中。关键问题的净现值显著低于非关键问题。结果证实NPV是反映周围血管张力的敏感指标。在后来的研究中，NPV被用作一个可靠和有效的自主神经指标。为了探讨自主神经反应的相关模式，我们对情绪诱发刺激的反应进行了因素分析。提取了三个因子：血压和外周血管因子，呼吸因子Am…更多的时间和心跳间期，和一个因素的皮肤电指数。这一分析表明，自主神经反应是由生理系统塑造的。我们还发现批判性问题中的特征反应发生了变化。我们将这些生理变化称为“关键问题模式”。有罪被试在批判性问题中的回答符合这一模式，而非有罪被试的回答不符合这一模式。结果表明，3种信息有助于区分非犯罪者和犯罪者；显著差异的数量、差异对应于关键问题模式的百分比、与外周血管阻力相关的回答百分比对应于关键问题模式。我们检验了潜伏性类别辨别法在测谎试验中的有效性。这种新的多变量方法根据给定的回答如何遵循每个生理系统中的关键问题模式来进行有罪和无罪的区分。该方法的性能优于其他传统的多元分析方法。较少

The present research investigated a new method for discriminating non-guilty persons from guilty ones in polygraph test based on the autonomic response pattern. Electrodermal activity, respiration movement, and plethysmogram have been used as indices in polygraph test. However, electrodermal activity tends to be habituated and plethysmogram can not indicate the activity level of sympathetic nervous system precisely. We introduced normalized pulse volume (NPV) into polygraph test as a new autonomic index. NPV showed significant decrement in the critical than in the non-critical questions. The results confirmed that NPV was a sensitive index of the peripheral vascular tone. NPV was used as a reliable and valid autonomic index in later studies. For investigating correlation patterns of autonomic responses, we conducted factor analysis on responses induced by emotion-evoking stimuli. Three factors were extracted ; a factor of blood pressure and peripheral vessel, a factor of respiration am … More plitude and inter heartbeat interval, and a factor of electrodermal indices. This analysis showed that autonomic responses were patterned by physiological systems. We also found characteristic response changes in the critical question. We named these physiological changes "the critical-question pattern". The responses during the critical question of guilty subjects corresponded to the pattern, while the responses of non-guilty subjects did not. The results indicated that three information were useful to discriminate non-guilty subjects from guilty ones ; the number of significant differences, the percentage of the differences corresponded to the critical-question pattern, and the percentage of responses related to peripheral vascular resistance corresponded to the critical- question pattern. We examined the availability of latent class discrimination method for polygraph test. This new multivariate method conducted guilty/non-guilty discrimination based on how a given response follows the critical-question pattern in each physiological system. The performance of this method was better than other conventional multivariate analyses. Less