Abstract

Purpose: The tendency to overuse the Internet is increasing with the advancement of digital technology. On the other hand, due to the availability of pornographic sites and their low protection against immature entry, children are more prone to porn addiction. Generally, diagnosing a child's addictive behavior, like pornography addiction, is made by a questionnaire, the accuracy of which is discussed by the scientific community. Consequently, the development of an automated system with electroencephalogram (EEG) signal analysis has been considered recently by researchers. We envisioned evaluating either signal dynamics and performing some executive functions on diagnosis rates. Methods: Some geometrical indices of EEG phase space were extracted for healthy and addicted children under different conditions, including closed eyes, memorizing, executive tasks, and recalling. Different machine learning algorithms were applied to evaluate their performances in a porn-addiction recognition problem. The proposed scheme was assessed using available Mendeley Data of 14 participants (seven porn-addicted and seven healthy) at ages 13 to 15. Results: The results revealed a noticeable increase in the features in most brain areas of porn addicts during recall. It also highlighted the role of task dependency, classification parameter settings, and k adjustment for K-fold cross-validation on porn detection rates. Totally, performing tasks, especially memorizing and recalling, compared to resting conditions could better highlight the difference in brain dynamics between the healthy and addicted groups. Conclusions: The EEG dynamical features were classified with the highest accuracy of 100%, which designated the scheme as a promising computer-aided diagnosis tool for porn addiction.

Keywords

Addiction, Children, Phase space, Dynamics, Electroencephalography (EEG);, Detection