Electrophysiological Characterization of Prejudice in Actors of the Colombian Armed Conflict

ABSTRACT : Colombia has had the most prolonged armed conflict, with the highest number of victims in Latin America. Since the beginning of the century, multiple efforts have been made to de-escalate it, such as the peace agreements with the paramilitaries in 2002 and the FARC guerrillas in 2016. However, efforts to de-escalate the conflict have had mixed results due to persistent prejudices among former actors. Therefore, it is necessary to characterize this prejudice to design more effective psychosocial intervention strategies to promote reconciliation. In this thesis, we present the characterization of electrophysiological patterns associated with a psychological test that assesses prejudice among former actors in the Colombian armed conflict. This characterization was done in three stages: Analyzing electrophysiological signals in the time domain, analyzing them in the frequency domain using graph theory, and merging electrophysiological features with other domain features to train an interpretable machine learning model. In the first stage, we developed a novel methodology for EEG-ERP analysis based on massive univariate statistical methods and Bayesian inference hypothesis testing. This methodology was used to analyze ERP related to an IAT task designed to assess prejudice, and we found that participants with prejudice toward one (victims) group exhibited higher activity than participants without prejudice or with prejudice toward the other group (ex-combatants). In the second stage, we developed a novel methodology for EEG-based functional connectivity analyses that adopted the techniques currently at the forefront of engineering and incorporated Bayesian inference hypothesis testing. This methodology was used to detect and measure differences among the configuration of the brain networks of former conflict actors; as a result, we found that victims and ex-paramilitaries generate more prejudice against victims, and civilians and ex-guerrillas generate more prejudice against combatants. However, victims and ex-guerrillas regulate more the prejudice against victims, and ex-paramilitaries regulate more the prejudice against combatants. All of these results are consistent with the results of the IAT task. In the third stage, we developed a novel methodology based on global and local analysis of interpretable machine learning models to identify the most important features in the characterization of groups and to evaluate the convenience of reclassifying subjects individually. We used demographic, behavioral, and electrophysiological features to characterize the groups of the former actors. As a result, we found that five characteristics of the 128 evaluated are sufficient to discriminate between groups of actors in armed conflict and to determine whether or not a participant should be reclassified. This research has allowed us to comprehensively characterize the phenomenon of prejudice that persists among former actors of the Colombian armed conflict, which will allow psychologists to design more specific intervention strategies to promote reconciliation among them.

• Autores: Quiza Montealegre, Jhon Jair
• Tipo de recurso: Doctoral thesis
• Fecha de publicación: 2024
• Institución: Universidad de Antioquia
• Repositorio: Repositorio UdeA
• Idioma: eng
• OAI Identifier: oai:bibliotecadigital.udea.edu.co:10495/41900
• Acceso en línea: https://hdl.handle.net/10495/41900
• Palabra clave: Fenómenos Electrofisiológicos; Electrophysiological Phenomena; Conflicto armado; Prejuicios y antipatías; Prejudices and antipathies; Reconciliación; Reconciliation; Pruebas psicológicas; Psychological tests; Aprendizaje automático (inteligencia artificial); Machine learning; https://id.nlm.nih.gov/mesh/D055724
• Rights: openAccess
• License: https://creativecommons.org/licenses/by-nc-sa/4.0/