For decades, disorders of gut-brain interaction (DGBI) have been ensnared in an epistemological bottleneck, clinically managed as diagnoses of exclusion despite the absence of structural pathology on conventional endoscopy. Traditional bottom-up models of visceral hypersensitivity fail to explain the profound subjective-objective symptom mismatches observed in clinical practice. This Perspective proposes a radical paradigm shift: leveraging the Predictive Processing (PP) framework to reconceptualize DGBI as a hierarchical computational dysfunction termed “interoceptive overfitting”. We postulate that rigid, high-precision threat priors force the salience network (dACC and aIns) to misallocate pathologically high precision weighting to baseline physiological noise, such as healthy 3-cycles-per-minute (cpm) gastric slow waves. This top-down failure synthesizes illusory pain and triggers genuine autonomic disruption via active inference, creating a self-fulfilling loop of GI micro-sabotage. We present a clinical roadmap utilizing high-resolution body surface gastric mapping (BSGM) and Ecological Momentary Assessment (EMA) to identify “Probabilistic Mismatch Points” within a multimodal diagnostic matrix that accounts for non-rhythmic peripheral modulators. To resolve therapeutic stagnation, we propose closed-loop digital therapeutics (DTx) designed to recalibrate the brain’s predictive engine through validation-correction loops, targeted extinction learning, and dual-stream telemetry. This computational framework provides a rigorously scientific blueprint to resolve therapeutic stagnation in DGBI.
Zeng et al. (Wed,) studied this question.