The results of the experiment show that brain reactions of suspects can be analyzed using an algorithm in real time, with great accuracy, even when there is an effort on the part of the subject to hide the information. The researchers are sure that the new method was able to solve historical crime mysteries such as the Damianiuk affair
About 30 years ago, one of the most famous trials in the history of the State of Israel took place, the center of which was the examination of the identity of the accused - John Demaniuk. Demjaniuk, a family man and a worker who worked according to his employers as an example, immigrated to the USA from Europe after World War II in order to realize the American dream. Decades after the end of the war, following a chance meeting with a Holocaust survivor, the suspicion arose that Demjanjuk was not just another war refugee, but that he was actually "Ivan the Terrible", a warden notorious for his extraordinary cruelty from the Nazi extermination camp - Treblinka. Following the serious suspicions, Demjanjuk was extradited to the State of Israel, where his trial began.
While very serious accusations were leveled against him, supported by evidence and documents, Demjanjuk himself claimed that it was a case of mistaken identity, since during World War II he was actually a prisoner of war by the Nazi regime in Germany. In the end, Demjanjuk was convicted, but after an appeal he was acquitted by the Supreme Court due to the reasonable doubt that a mistake was indeed made in the identification. Since then, Demjanjuk has been convicted of war crimes in Germany as well, but died during the appeal, so the question of his true identity remains unresolved.
Hidden knowledge discovery test
The mystery could have been shed light, even before the trial, through the creative use of a hidden knowledge discovery test - CIT, which could have identified Demjanjuk's prior acquaintance with information related to Ivan the Terrible (for example, his personal number from the Nazi army) - this is what the psychiatrist claimed a few years ago The late David Lyken of the University of Minnesota. A new study carried out under the leadership of Dr. Nathalie Klein Sela and Chen Guetta, under the guidance of Prof. Emeritus Gershon Ben Shahar and Prof. Leon Deuel - all from the Hebrew University, in collaboration with Dr. Yuval Harpaz from Inner Eye and recently published in the journal Cortex, presented a similar argument. In the CIT test, the suspect is presented with prepared information about a crime, and if the suspect knows the information, it is possible to identify A typical physical and mental reaction, even if he tries hard to hide it. "This test is a worthy alternative to the polygraph lie detection test, which is a well-known but problematic method for many reasons, mainly legal. The CIT is a more reliable and objective method, which relies on the fact that only the real criminal will identify information related to the crime he committed," explains Dr. Nathalie Klein Sella.
Brain reactions are measured by tens of sensors and at a high rate; Artificial intelligence makes it possible to take control of the data
The new study examined whether the brain responses in the CIT test are the result of the attempt to hide the familiarity of the hidden information (inhibition), or the result of directing attention to the familiar item (orientation). This question has theoretical but also practical importance - The attempt to hide can have a special meaning in a criminal scenario. Also, the brain reactions are measured by dozens of sensors and at a high rate, and expertise is required to process them. Therefore, the study examined whether it is possible to use a multivariate pattern classification machine learning algorithm to process the information quickly and automatically while maintaining high accuracy of the CIT test and even improving it.
In order to test this, dozens of students who volunteered to participate in the study were examined. Each participant attended two visits to the laboratory. In the first visit, personal details were collected from each participant (eg name, city of residence and subject of study). Then, during two sessions, it was explained to the participants that they would be shown words on the screen, a small part of which would be related to their personal information (and therefore might cause attentional bias). The significant difference between the two meetings was that in one meeting the participants were asked to hide their personal information and in the second they were asked to reveal it. At the beginning of each session, electrodes were placed on the heads of the participants to measure their brain activity.
The brain reactions showed a significant difference between the reactions to the subjects' personal information, and the reactions to personal information that is not relevant to them. The brain responses allowed the researchers to discover with high reliability items such as the first name of the subjects - from a list of different names. However, no difference was found between the session in which the participants tried to hide the information, and the session in which they tried to reveal it. Hence, the measured brain response reflects attention being directed to an item that is meaningful to the subjects, and not the attempt to hide. These results were found both when the brain waves were analyzed using the traditional methods, and also by using a multidimensional machine learning algorithm. in fact, It was found that the analysis using the algorithm reaches greater accuracy than the traditional methods in revealing personal information.
The CIT test is currently performed only in Japan
This finding may have far-reaching consequences for the possibility of applying information discovery using brain waves in the field. Today, the CIT test is only routinely used in Japan, and even there brainwaves are not used, as this requires a lot of time and expertise, which is difficult to apply in the field. The results in this experiment show that it is possible to analyze the mental reactions of suspects using the algorithm in real time, and with great accuracy even when there is an attempt to hide the information. This technique may significantly facilitate the performance of CIT tests using a platform adapted to the field, which could make the use of brain waves in investigations a reality in the near future. The computerized algorithm used in this study was developed for other purposes in the laboratory of Prof. Leon Deuel at the Hebrew University in collaboration with Prof. Amir Geva at Ben Gurion University, and is currently being developed for additional purposes at an Israeli startup company called InnerEye.
"The results show that it is possible to discover hidden memories, both in research and in application, by using all the complexity of the event-dependent brain waves, without the need to rely on a specific component as in a polygraph. Together, these results enrich the theoretical knowledge underlying the hidden information test and pave the way for the development of algorithms with many applications in criminal investigations", concludes Dr. Nathalie Klein Sela.
More of the topic in Hayadan:
