KMAN Publication Inc. (KMANPUB) Journal of Personality and Psychosomatic Research (JPPR) 3041-8542 4 Serial Number 15 2026 07 01 Alexithymia, Interoceptive Awareness, and Autonomic Indices: A Machine-Learning Approach to Psychosomatic Vulnerability Alexithymia, Interoceptive Awareness, and Autonomic Indices: A Machine-Learning Approach to Psychosomatic Vulnerability 1 11 EN Journal Article 2026 04 10 <p>This study aims to predict psychosomatic vulnerability by integrating self-reported measures of alexithymia and interoceptive awareness with objective autonomic indices using a machine-learning approach. This cross-sectional observational study included healthy adult participants (aged to years) recruited from urban centers in Colombia. Self-reported psychological and somatic data were collected using the Toronto Alexithymia Scale (TAS- ), Multidimensional Assessment of Interoceptive Awareness (MAIA), and Somatic Symptom Scale (SSS). Continuous objective physiological data, specifically electrocardiogram and electrodermal activity, were acquired at a sampling rate of Hz during resting, stress induction, and recovery phases. Machine learning models, including Random Forest, Support Vector Machines, and Gradient Boosting, were developed using an /  train/test data split and ten-fold cross-validation, with SHapley Additive exPlanations (SHAP) utilized to determine feature importance. The Gradient Boosting model yielded the highest predictive performance, achieving an accuracy of and an Area Under the Curve (AUC) of on the test dataset. SHAP value analysis identified that the top predictors of psychosomatic vulnerability were the “difficulty identifying feelings” facet of the TAS- , the “not-worrying” dimension of the MAIA, baseline root mean square of successive differences (RMSSD), and the low-frequency to high-frequency (LF/HF) power ratio during stress. Psychosomatic vulnerability can be highly accurately predicted by synergistically integrating subjective psychological traits and objective physiological markers through advanced machine learning algorithms.</p> https://www.journals.kmanpub.com/index.php/jppr/article/download/5888/11269