인지과학 (Korean Journal of Cognitive Science)

한국인지과학회 (The Korean Society for Cognitive Science)
권호 표지
DOI QR코드

DOI QR Code

기계학습을 이용한 단일 관련자극 P300기반 숨김정보검사

One-probe P300 based concealed information test with machine learning

  • 김혁 (고려대학교 심리학과) ;
  • 김현택 (고려대학교 심리학과)
  • Hyuk Kim (Department of psychology, Korea University) ;
  • Hyun-Taek Kim (Department of psychology, Korea University)
  • 투고 : 2024.02.29
  • 심사 : 2024.03.14
  • 발행 : 2024.03.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

국내 형사소송절차에서 진술의 진위여부 확인을 위해 사용하는 도구는 폴리그래프검사, 진술타당도분석, P300 기반 숨김정보검사 등이 있고, 이 중에서 폴리그래프검사의 사용빈도가 다른 도구들에 비하여 높다. 하지만, 검사결과를 뒷받침해 줄 수 있는 근거의 부족으로 인하여 재판과정에서 증거채택 가능성이 낮다. 폴리그래프검사를 뒷받침해 줄 수 있는 방법으로, 사전연구가 풍부한 P300기반 숨김정보검사가 주목을 받아 왔지만, 기존의 검사기법은 두 가지 제한점이 있어 실제 사건에서의 활용도는 낮은 편이다. 첫째, 검사에 필요한 관련자극만 3개 또는 6개 등, 사전에 노출되지 않은 정보가 다수 필요하기 때문에 실제 사건에서 사용 가능성이 낮다. 둘째, 기존의 P300기반 숨김정보검사 프로토콜에서는 관련자극과 무관련자극에 대한 P300요소 전위값을 명확하게 구분하기 위하여 오드볼패러다임을 사용하기 때문에 무관련자극에 대한 P300요소 전위값이 과소 추정될 가능성이 있다. 본 연구에서는 검사의 사용 가능성을 높이기 위하여 사전에 노출되지 않은 정보가 단 하나만 있어도 검사가 가능한 단일 관련자극을 사용하는 수정된 P300기반 숨김정보검사 프로토콜을 탐색하였고, 오드볼패러다임 사용으로 인한 무관련자극에 대한 P300요소 전위값이 과소 추정되는 문제를 보완하기 위하여 다양한 기계학습의 분류 알고리즘을 비교하였다. 연구결과 단일 관련자극으로 여성과 남성의 얼굴자극을 사용할 경우, 자극은 400ms 지속시간으로 60회 제시하고, 절단값을 유죄집단은 90%로 무죄집단은 30%로 하여 정점-정점 방법으로 P300요소 전위값을 분석하는 것이 적합함을 확인하였다. 단어자극의 경우, 지속시간을 300ms로 60회 제시하고, P300요소 전위값 분석방법은 얼굴자극과 동일하게 시행하는 것이 적합하다는 것을 확인하였다. 또한 관련자극과 무관련자극에 대한 정점-정점 P300요소 전위값을 6가지 기계학습 분류 알고리즘을 사용하여 분석한 결과, 로지스틱 회귀(LR), 선형 판별 분석(LDA), K-최근접 이웃(KNN) 알고리즘이 관련자극과 무관련자극의 분류에 적합하다는 것을 확인하였다.

Polygraph examination, statement validity analysis and P300-based concealed information test are major three examination tools, which are use to determine a person's truthfulness and credibility in criminal procedure. Although polygraph examination is most common in criminal procedure, but it has little admissibility of evidence due to the weakness of scientific basis. In 1990s to support the weakness of scientific basis about polygraph, Farwell and Donchin proposed the P300-based concealed information test technique. The P300-based concealed information test has two strong points. First, the P300-based concealed information test is easy to conduct with polygraph. Second, the P300-based concealed information test has plentiful scientific basis. Nevertheless, the utilization of P300-based concealed information test is infrequent, because of the quantity of probe stimulus. The probe stimulus contains closed information that is relevant to the crime or other investigated situation. In tradition P300-based concealed information test protocol, three or more probe stimuli are necessarily needed. But it is hard to acquire three or more probe stimuli, because most of the crime relevant information is opened in investigative situation. In addition, P300-based concealed information test uses oddball paradigm, and oddball paradigm makes imbalance between the number of probe and irrelevant stimulus. Thus, there is a possibility that the unbalanced number of probe and irrelevant stimulus caused systematic underestimation of P300 amplitude of irrelevant stimuli. To overcome the these two limitation of P300-based concealed information test, one-probe P300-based concealed information test protocol is explored with various machine learning algorithms. According to this study, parameters of the modified one-probe protocol are as follows. In the condition of female and male face stimuli, the duration of stimuli are encouraged 400ms, the repetition of stimuli are encouraged 60 times, the analysis method of P300 amplitude is encouraged peak to peak method, the cut-off of guilty condition is encouraged 90% and the cut-off of innocent condition is encouraged 30%. In the condition of two-syllable word stimulus, the duration of stimulus is encouraged 300ms, the repetition of stimulus is encouraged 60 times, the analysis method of P300 amplitude is encouraged peak to peak method, the cut-off of guilty condition is encouraged 90% and the cut-off of innocent condition is encouraged 30%. It was also conformed that the logistic regression (LR), linear discriminant analysis (LDA), K Neighbors (KNN) algorithms were probable methods for analysis of P300 amplitude. The one-probe P300-based concealed information test with machine learning protocol is helpful to increase utilization of P300-based concealed information test, and supports to determine a person's truthfulness and credibility with the polygraph examination in criminal procedure.

키워드

참고문헌

  1. 김혁 (2021). 연구단보: 사진과 문장이 동시에 제시되는 복합자극을 이용한 뇌파검사기법 연구. 과학수사학회지, 15(2), 159-163.
  2. 김형준, 김재휘, 백승경 (2005). 刑事節次에 있어서 虛僞陳述에 관한 實證的 硏究. 중앙법학, 7(1), 185-208.
  3. 문혜민, 조은경 (2021). 아동 목격자 진술에서 준거기반 내용분석(Criteria-Based Content Analysis)의 진위 판별 효과성: 메타분석. 한국경찰학회보, 23(5), 123-158.
  4. 박성민 (2021). 조사자 증언의 문제점과 개선방안. 성균관법학, 33(2), 237-266.
  5. 박용철 (2013). 거짓말탐지기 검사 결과의 활용에 대한 소고. 형사정책, 25(3), 339-360.
  6. 이미선 (2018). 성폭력 피해아동 진술신빙성 판단에 있어서 평가자간 신뢰도: 진술분석 전문가 집단을 대상으로. 한국심리학회지: 사회및성격, 32(2), 67-83.
  7. 이지영, 민수정, 이도준 (2011). 스트룹 간섭의 희석 및 중복 효과. 인지과학, 22(4), 469-494.
  8. 최정학. (2009). 거짓말탐지기 검사결과의 증거능력 - "과학적 증거"의 허용기준과 관련하여 -. 경 희법학, 44(1), 9-33.
  9. Abootalebi, V., Moradi, M. H., & Khalilzadeh, M. A. (2006). A comparison of methods for ERP assessment in a P300-based GKT. International journal of psychophysiology: official journal of the International Organization of Psychophysiology, 62(2), 309-320.
  10. Delorme, A., & Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1), 9-21.
  11. Donchin E. (1981). Presidential address, 1980. Surprise!...Surprise?. Psychophysiology, 18(5), 493-513.
  12. Farwell, L. A., & Donchin, E. (1991). The truth will out: interrogative polygraphy ("lie detection") with event-related brain potentials. Psychophysiology, 28(5), 531-547.
  13. Farwell L. A. (2012). P300-based concealed information test: a comprehensive tutorial review of detection of concealed information with event-related brain potentials. Cognitive neurodynamics, 6(2), 115-154.
  14. Fisch, B. J. (1999) 1. The source of the EEG. In Fisch and Spehlmann's EEG Primer Basic Principles of Digital and Analog EEG (3rd Edition). (Page 3-17). Elsevier.
  15. Gao, J., Tian, H., Yang, Y., Yu, X., Li, C., & Rao, N. (2014). A novel algorithm to enhance P300 in single trials: application to lie detection using F-score and SVM. PloS one, 9(11)
  16. Ilan, A. B., & Polich, J. (1999). P300 and response time from a manual Stroop task. Clinical neurophysiology: official journal of the International Federation of Clinical Neurophysiology, 110(2), 367-373.
  17. Haider, S. K., Jiang, A., Jamshed, M. A., Pervaiz, H., & Mumtaz, S. (2018). Performance enhancement in P300 ERP single trial by machine learning adaptive denoising mechanism. IEEE Networking Letters, 1(1), 26-29.
  18. Hillyard, S. A., Squires, K. C., Bauer, J. W., & Lindsay, P. H. (1971). Evoked Potential Correlates of Auditory Signal Detection. Science, 172(3990), 1357-1360.
  19. Jasper, H. H. (1958). The ten-twenty electrode system of the International Federation. Electroencephalography and clinical neurophysiology, 10, 371-375.
  20. Krapohl, D., Handler, M., Sturm, S.(2012) Terminology Reference for the Science of Psychophysiological Detection of Deception, Third Edition, American Polygraph Association, 64 page.
  21. Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: finding meaning in the N400 component of the event-related brain potential (ERP). Annual review of psychology, 62, 621-647.
  22. Li, Y., Ma, Z., Lu, W., & Li, Y. (2006). Automatic removal of the eye blink artifact from EEG using an ICA-based template matching approach. Physiological measurement, 27(4), 425-436.
  23. Luck, S. J., & Kappenman, E. S. (2012). ERP components and selective attention. In S. J. Luck & E. S. Kappenman (Eds.), The Oxford handbook of event-related potential components (pp. 295-327). Oxford University Press.
  24. Lui, M., & Rosenfeld, J. P. (2008). Detection of deception about multiple, concealed, mock crime items, based on a spatial-temporal analysis of ERP amplitude and scalp distribution. Psychophysiology, 45(5), 721-730.
  25. Mehrnam, A. H., Nasrabadi, A. M., Ghodousi, M., Mohammadian, A., & Torabi, S. (2017). A new approach to analyze data from EEG-based concealed face recognition system. International journal of psychophysiology : official journal of the International Organization of Psychophysiology, 116, 1-8.
  26. Nemrodov, D., Niemeier, M., Mok, J., & Nestor, A. (2016). The time course of individual face recognition: A pattern analysis of ERP signals. NeuroImage, 132, 469-476.
  27. Polich, J. (2012). Neuropsychology of P300. In S. J. Luck & E. S. Kappenman (Eds.), The Oxford handbook of event-related potential components (pp. 159-188). Oxford University Press.
  28. Rosenfeld JP, Shue E, Singer E. (2007). Single versus multiple probe blocks of P300-based concealed information tests for self-referring versus incidentally obtained information. Biological Psychology, Vol. 74(3), 396-404.
  29. Rosenfeld, J. P. (2011). P300 in detecting concealed information. In B. Verschuere, G. Ben-Shakhar, & E. Meijer (Eds.), Memory detection: Theory and application of the Concealed Information Test (pp. 63-89). Cambridge University Press.
  30. Rosenfeld, J. P., & Donchin, E. (2015). Resampling (bootstrapping) the mean: A definite do. Psychophysiology, 52(7), 969-972.
  31. Rossion, B., & Jacques, C. (2012). The N170: Understanding the time course of face perception in the human brain. In S. J. Luck & E. S. Kappenman (Eds.), The Oxford handbook of event-related potential components (pp. 115-141). Oxford University Press.
  32. Raschka, S., & Mirjalili, V. (2019). Python Machine Learning: Machine Learning and Deep Learning with Python, Scikit-Learn, and TensorFlow 2, 3rd Edition. Packt Publishing.
  33. Watts, F. N. (1979). Habituation model of systematic desensitization. Psychological Bulletin, 86(3), 627-637.