The Relational Collision Model: Profile-Specific Interaction Dynamics, PAG-Mediated Conflict Cascades, and the Metabolic Tax of Attachment-Discordant Dyads

Background: Papers 0–8 of the ABM Blueprint Research Series have established the neurobiological architecture of individual attachment regulation, from the foundational Birth Pulse calibration (Paper 0) through molecular mechanisms (ECS–PAG axis, Paper 1) to intervention sequencing (Recalibration Protocol, Paper 4). However, the existing framework addresses the individual organism in relative isolation. What remains unaddressed is the interaction dynamics that emerge when two differently calibrated nervous systems enter sustained proximity—the dyadic collision problem. Objective: This paper introduces the Relational Collision Model (RCM), a PAG-mediated framework for predicting, classifying, and intervening in attachment-discordant dyadic conflict. The model extends the three-profile ABM taxonomy (Architect, Radar, Special Forces) to a five-profile system by differentiating the Disorganized attachment phenotype into Controlling-Punitive (CP) and Controlling-Caregiving (CC) subtypes, consistent with Cassidy and Marvin’s (1992) classification and subsequent validation by Moss et al. (2004, 2006). This expansion yields 15 distinct dyadic collision configurations, each with a specific PAG-column activation sequence, predictable conflict cascade, metabolic cost signature, and intervention requirement. Methods: The model integrates six convergent evidence streams: (1) Gottman’s physiological research on Diffuse Physiological Arousal (DPA) and the “Four Horsemen” mapped onto PAG column activation; (2) Porges’ Polyvagal Theory as the autonomic architecture of co-regulation and co-dysregulation; (3) hyperscanning research on Interpersonal Neural Synchrony (INS) as a biomarker of dyadic alignment; (4) Panksepp’s affective neuroscience framework (PANIC/GRIEF, RAGE, FEAR systems) as the subcortical engines of relational behavior; (5) allostatic load and neuroimmunological research on the metabolic cost of chronic relational stress; and (6) Cassidy and Marvin’s (1992) empirical differentiation of Disorganized attachment into Controlling-Punitive and Controlling-Caregiving subtypes, with their distinct externalization/internalization signatures. Results: The 15-configuration collision matrix reveals four clinically distinct tiers: Tier 1 (Acute Risk)—configurations involving CP or SFu profiles that carry violence risk and trauma bonding dynamics requiring safety planning before dyadic intervention; Tier 2 (Invisible Damage)—configurations involving CC profiles where internalization, somatization, and autoimmune consequences accumulate without visible conflict; Tier 3 (Classical Visible)—configurations involving Radar profiles with observable protest-pursuit dynamics; Tier 4 (Silent Erosion)—Architect-dominated configurations with gradual bond dissolution. Each configuration generates specific, testable predictions regarding conflict onset latency, escalation trajectory, metabolic cost distribution, somatization patterns, and intervention response. Conclusions: The Relational Collision Model provides the first profile-specific, neurobiologically grounded framework for understanding why specific attachment combinations produce specific conflict patterns. By treating dyadic conflict as a mechanical interaction between calibrated defense systems rather than a failure of communication or character, the model opens pathways for targeted, hardware-first couple intervention.