KMAN Publication Inc. (KMANPUB) Journal of Assessment and Research in Applied Counseling (JARAC) 8 Serial Number 30 2026 07 01 Predicting Emotional Dysregulation Using Artificial Neural Networks Based on Childhood Trauma, Attachment Insecurity, Alexithymia, and Mindfulness Predicting Emotional Dysregulation Using Artificial Neural Networks Based on Childhood Trauma, Attachment Insecurity, Alexithymia, and Mindfulness 1 15 EN Journal Article 2026 02 20 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to predict emotional dysregulation using artificial neural networks based on childhood trauma, attachment insecurity, alexithymia, and mindfulness among adults in Armenia.</p> <p><strong>Methods and Materials:</strong> This quantitative, cross-sectional, predictive study was conducted on 842 adults from Armenia. Participants were selected through convenience sampling from universities, community centers, healthcare settings, and online research platforms. Data were collected using the Difficulties in Emotion Regulation Scale, Childhood Trauma Questionnaire-Short Form, Experiences in Close Relationships-Revised questionnaire, Toronto Alexithymia Scale, Five Facet Mindfulness Questionnaire-Short Form, and a demographic information form. After data screening, normalization, and reliability assessment, descriptive and correlational analyses were performed in SPSS version 29. Artificial neural network modeling was conducted in Python using TensorFlow and Keras. The dataset was divided into training, validation, and testing subsets, and model performance was evaluated using MAE, MSE, RMSE, MAPE, R², and five-fold cross-validation. SHAP analysis was used to interpret predictor importance.</p> <p><strong>Findings:</strong> Emotional dysregulation showed significant positive correlations with childhood trauma (r = .69, p &lt; .001), attachment insecurity (r = .64, p &lt; .001), and alexithymia (r = .72, p &lt; .001), and a significant negative correlation with mindfulness (r = −.67, p &lt; .001). The optimized artificial neural network demonstrated strong predictive performance, explaining 91.4% of variance in the training set, 90.1% in the validation set, and 89.6% in the independent testing set. The model produced low prediction errors, including MAE = 4.28, RMSE = 5.64, and MAPE = 5.81%. Five-fold cross-validation confirmed model stability, with a mean R² of .902. SHAP analysis identified alexithymia as the strongest predictor, followed by childhood trauma, attachment insecurity, and mindfulness.</p> <p><strong>Conclusion:</strong> Emotional dysregulation was accurately predicted through a nonlinear artificial neural network model based on developmental, relational, emotional-awareness, and mindfulness-related factors. The findings highlight alexithymia and childhood trauma as central risk indicators and mindfulness as an important protective factor.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jarac/article/download/5658/10979