Reward Valuation in Vision Language Models: Causal Mechanisms Underlying Anhedonia

2026-07-07

AI Summary

This study investigates whether vision-language models possess human-like reward valuation mechanisms and can emulate the causal underpinnings of anhedonia. Grounded in clinical anhedonia assessment tasks and the nucleus accumbens (NAc)–dopamine reward system framework from neuroscience, we identify reward anticipation units within the model and validate their causal role through targeted perturbations. Results show that perturbing NAc-selective units shifts model preferences toward low-effort, low-reward options, closely mirroring human anhedonic behavior while preserving baseline performance on non-reward tasks. This behavioral shift significantly correlates with clinical anhedonia scales (DARS and MAP-SR). Our work establishes, for the first time in an AI model, a causal mechanism aligned with human anhedonia, revealing a human-like reward valuation circuit and distinguishing specific reward-processing deficits from general cognitive decline.

Abstract

Recent Vision-Language Models capture increasingly complex aspects of human cognition. Here we ask whether this alignment extends to reward valuation, which we assess in a mechanistic framework built on clinical tests that were developed to evaluate anhedonia and motivational deficits in major depressive disorder. In the brain, anhedonia is frequently linked to dysregulation in the Nucleus Accumbens (NAc) and the broader dopaminergic reward system. While neuroimaging has localized these deficits, establishing a causal link between NAc activity and specific behavioral symptoms remains a challenge. We use these ideas from neuroscience to functionally identify reward-anticipatory units in vision language models, and test their causal role via targeted perturbations. Perturbing NAc-selective units induces behavioral effects that mirror human anhedonia: the model shifts toward low-effort, low-reward options in effort-based decision-making tasks. Crucially, our results reflect a specific deficit in reward valuation and anticipation rather than a loss of task capability: the perturbed model maintains baseline performance when reward-based choice is removed. This induced vulnerability further aligns with clinical anhedonia and motivation scales, including DARS and MAP-SR. Taken together, these results reveal reward valuation circuits in AI models that parallel those in humans.

Problem

• reward valuation
• anhedonia
• vision-language models
• motivational deficits
• Nucleus Accumbens

Innovation

• reward valuation
• vision-language models
• causal perturbation
• anhedonia
• nucleus accumbens