The research consisted of a preliminary study consisted at the Neuropsychiatric Department of the Military Hospital of Legal Medicine of Verona, Italy. Data were collected from 170 soldiers. The participants were grouped into three subsamples. The first group was characterized by a strong tendency toward normalcy, the second by a strong tendency toward pathology, and the third by a great variety of expressions of psychological and social problems. The participants responded to a questionnaire designed according to Squashing Theory principles. The data were processed by an Artificial Neural Network created by Semeion in Rome, Italy, and were compared with standard clinical psychiatric assessment reports and with the results of psychodiagnostic tests. Results revealed that the instrument had a predictive capacity of more than 90 percent. A comparison with the participants clinical reports and an initial follow-up also revealed high agreement. The broad variation in answers from the third subsample also suggested further methodological issues on the contribution of Artificial Neural Networks and Squashing Theory to the study of deviance related to drug addiction and other areas. Figures, author biographies and photographs, and 26 references (Author abstract modified)