Development and Validation of a Multimodal AI Framework for Assessing Craving Intensity and Vedic Personality Traits in Substance Use Disorders

Abstract

The clinical evaluation of craving in the context of Substance Use Disorders, is still heavily dependent on subjective self-reports, which are frequently tainted by patient anosognosia and social desirability bias, despite advancements in addiction science. Moreover, the qualitative internal states described by Indian Knowledge Systems (IKS) and the subtle neurocognitive antecedents of relapse outlined by the Binding and Retrieval in Action Control (BRAC) model are not captured by existing diagnostics. By creating a ground-breaking Multimodal AI Fusion Framework, this study fills this diagnostic gap. The goal of the study is to translate the abstract philosophical concepts of Triguna, more especially, the transition from Tamasic inertia to Sattvik balance, into objective, measurable bio-behavioral indicators. The study uses a unique,non-invasive web-based battery that synchronizes three different data streams to produce a composite Relapse Risk Index using a cross-sectional tool-development approach (N=50). First, computer vision oculometrics calculates gaze entropy using common webcams to understand the hypervigilant scanning behaviour typical in high craving states. Second, in order to identify the microscopic motor rigidities linked to maladaptive event file binding, kinematic analysis records mouse cursor movements during cognitive interference tasks, similar to an Alcohol stroop test, by examining trajectory curvature and velocity peaks. Third, speech recordings are analyzed by Acoustic and Semantic NLP to determine the prosodic signs (jitter/shimmer) of Sattvik clarity against Tamasic depression. This study goes beyond conventional psychometrics by creating a machine learning classifier to merge various physiological information. The Vedic Lifestyle Scale and the Obsessive Compulsive Drinking Scale (OCDS) are used to validate these algorithmic features. By developing a scalable, culturally based methodology, the study makes a significant addition to precision psychiatry. It makes the case that effective rehabilitation involves more than just quitting drugs; rather, it involves a quantifiable change in neuro-cognitive economy, which can now be identified, measured, and tracked thanks to the convergence of artificial intelligence and traditional psychology theory.