Peer-Reviewed Foundations

All frameworks powering oDIX8 products are published and open to scrutiny.

 

EMoT | arXiv:2603.24065 | 2026

Enhanced Mycelium of Thought: A Bio-Inspired Hierarchical Reasoning Architecture with Strategic Dormancy and Mnemonic Encoding

The first reasoning framework to combine hierarchical topology, strategic thought dormancy with reactivation, and mnemonic memory encoding. Strategic dormancy proved architecturally critical — removing it caused quality collapse from 4.2 to 1.0.

[Read on arXiv →] https://arxiv.org/abs/2603.24065

Design Patterns for Clinical AI | arXiv:2604.01661 | 2026

Ontology-Aware Design Patterns for Clinical AI Systems: Translating Reification Theory into Software Architecture

Seven Gang-of-Four style patterns for building clinical AI pipelines resilient to ontological distortion: Ontological Checkpoint, Dormancy-Aware Pipeline, Drift Sentinel, Dual-Ontology Layer, Reification Circuit Breaker, Terminology Version Gate, Regulatory Compliance Adapter.

[Read on arXiv →] https://arxiv.org/abs/2604.01661

Artificial Epidemiology | Cambridge Open Engage | 2026How Autonomous Documentation Agents Could Manufacture Disease Prevalence at Population Scale

AI documentation agents may amplify coding distortion by 17x to 21x over five iterations. Formalises the divergence D(C) between administrative and clinical prevalence. Pre-study on 11,475 synthetic patients reveals substantial baseline divergence even in algorithmically clean data.

[Read on Cambridge Open Engage →]

Preserve, Don't Prune | Research Square | 2026

Strategic Dormancy as a Governance-Relevant Architectural Property for Clinical AI

Pruning at 1% prevalence eliminates 46.9% of diagnostic vocabulary and affects 24.4% of patients. Of 145 pruned conditions, 34 are clinically actionable. 80% of rare conditions have diagnostic bridges that pruning severs.

[Read on Research Square →]

PRAXIS-AI | SSRN | 2026

An Overlay Implementation Framework for AI in EU Primary Care

Nineteen implementation science frameworks evaluated against seven AI-specific criteria. None adequate. PRAXIS-AI fills the gap as an overlay framework with seven domains across five implementation phases, grounded in EU AI Act, GDPR, DiGAV, and ISO 42001.

[Read on SSRN →] https://ssrn.com/abstract=6474960

The Symptom and the Billing Code | SSRN | 2026

How Three Competing Forces Construct Minimal Medical Data Sets

Health data is not recorded but constructed through negotiation among clinical, billing, and documentation pressure. Introduces documentary enactment and the reification feedback loop. Demonstrated in the German FDZ Gesundheit: one encounter, 28 tables, 75 million insured.

[Read on SSRN →] https://ssrn.com/abstract=6475038