Orbitofrontal-ventral striatal coupling: A model of the neurobehavioral architecture of craving in substance use disorders.

Neuroimaging has identified cue-elicited activity in ventral striatum and orbitofrontal cortex (OFC) as a reliable signature of craving, yet activation magnitude in either region is an inconsistent predictor of clinical outcome. This theoretical review develops a model on which the clinically consequential variable is not activation but effective coupling: the cue-evoked gain-the factor by which an input is amplified-with which OFC valuation signals modulate ventral striatal action-selection activity. I formalize this gain as valuation-policy coupling (λ), which governs the probability that a drug-related appraisal recruits approach and consumption policies; it is a latent construct indexed by effective OFC-accumbens connectivity, approach bias under load, and pharmacological challenge. Two moderators-meta-awareness and elaborative versus nonelaborative engagement-determine whether the urge is re-entered or exited, and jointly specify a neurocognitive space in which craving occupies a high-λ, high-meta-awareness, elaborative vertex. The framework predicts an awareness-amplification effect-noticing an urge intensifies it under high λ and weakens it under low λ-specifies when mindfulness-based relapse prevention should attenuate craving and when it should not, and yields a discriminative prediction: λ-reducing interventions should reduce cue-modulated OFC-ventral-striatum coupling while leaving regional activation comparatively preserved. λ is a parameter of craving rather than of relapse-which also requires cognitive-control parameters-and a candidate transdiagnostic mechanism. Claims are labeled by evidential status; the λ-meta-awareness interaction is the model's principal hypothesis, and failure conditions are stated explicitly.