Zifeng Wang, Guangdong Zhou, Yifan Liu, Yuchen Huang, Taoyu Wu, Shiqiu Meng, Chengyao Qian, Yan Sun
The Load Group showed significantly reduced Alcohol Urge Questionnaire (AUQ) scores post-task. While no significant differences emerged in early attentional bias, the Load Group showed suppressed late attentional bias toward alcohol cues, whereas the No Load Group showed a significant increase.
BACKGROUND: Internet gaming disorder (IGD) involves dysregulated incentive processing, yet precise neural substrates remain inconsistent. While the Monetary Incentive Delay (MID) task assesses reward systems, traditional analyses conflate reward valence (Value) and motivational salience, masking process-specific neural alterations. This study evaluated which neural features-from localized activity to large-scale connectivity-most effectively characterize IGD clinical phenotypes.
METHODS: We analyzed fMRI data from 90 regular gamers, including 45 participants in the IGD group and 45 matched recreational game users (RGU), who performed a MID task with five reward/loss levels. Applying the Orthogonal-Decoding (DeCoP) framework, we dissociated anticipation BOLD signals into independent Value and Salience components. We compared localized activation (GLM), multivariate spatial patterns (Searchlight MVPA), and task-modulated functional connectivity (gPPI). Connectome-based Predictive Modeling (CPM) identified neural signatures predicting symptom severity and craving. An exploratory subsample (n = 15; 3 IGD and 12 RGU) completed clinical assessments approximately 12 months later.
RESULTS: Localized activation and spatial patterns showed no significant group differences across either anticipation or feedback. However, positive Value-modulated connectivity during anticipation significantly predicted IGD symptom severity (66 stable edges, ρ = .317, p = .002) and craving (37 stable edges, ρ = .348, p < .001). The intersection of these positive connectome models revealed a shared Value network (16 stable edges) anchored in a medial frontal-cerebellar circuit. Exploratory follow-up analysis showed that baseline shared-network strength was associated with 12-month craving (ρ = .724, p = .002).
CONCLUSIONS: Our findings suggest the neurobiological core of IGD is linked to aberrant network-level coordination during reward anticipation, rather than regional hyper-reactivity or feedback processing deficits. We identify anticipatory valuation dysregulation as a key feature of IGD, offering a process-specific connectomic signature of IGD symptomatology.