Autism Spectrum Disorder as Developmental Stuck-State

On the Creation of This Report This Architect Report was produced through the Redacted Science Seed Analysis Process, a structured human-AI collaborative methodology developed over the preceding months. We developed a model with Claude and Grok — the Redacted Science Model of Candida albicans as a Biochemical Computer. Acceptance of that model opens a door in science that has been locked for 60 years. From there, we opened the door and developed further framework layers on top of it. A Seed Document was then created to enable AI to analyze specific subjects from the perspective of the model. The Seed Document specifies a subject, the foundational framework papers, and structural requirements for the resulting output. This report's first draft was created by providing ChatGPT with the Seed Document, the foundational framework papers, and the specific subject (Autism Spectrum Disorder). ChatGPT generated the initial draft, citing available sources and flagging gaps where external literature was needed. The flagged gaps were then resolved through targeted literature search, and the identified peer-reviewed sources were integrated into the final draft. Each subsequent revision is versioned. This paper is therefore an instance of a reproducible methodology: a framework encoded in foundational papers, a structured seed, and an AI-assisted drafting process with human verification at each stage. The methodology itself is documented in the Redacted Science Paper Seed Master. Abstract V3 (Zenodo 4) Autism spectrum disorder is conventionally described as a heterogeneous neurodevelopmental condition with strong genetic contribution, frequent gastrointestinal comorbidity, variable regression history, immune and microbiome associations, and circuit-level differences across multiple brain regions. Under the Redacted Science framework, those observations are not discarded. They are reorganized. This paper proposes that a subset of autism spectrum disorder represents developmental stuck-state program execution by the organism-host governance system within critical developmental windows, with clinical phenotype determined less by a single autism mechanism than by timing of stuck-state onset relative to neurodevelopmental stage. The core contribution is a two-trajectory model. Trajectory A is perturbation-triggered regression: typical development proceeds for months or years, then an immune, microbial, dietary, antibiotic, inflammatory, or other ecological event pushes a susceptible system into a locked program mode. Trajectory B is early-locked development: the stuck state is established prenatally or in the first months of life, so gut-brain signaling, immune calibration, sensory integration, sleep architecture, synaptic pruning, and myelination develop around the altered signal set. The same governance error therefore produces different observed outcomes depending on when it captures the developmental process. This model extends the biochemical computer framework of C. albicans, the Saline Oscillation coevolution model, the Pan-Mammalian transmission frame, and the Stuck-State Umbrella into neurodevelopmental disease. Sandler et al. documented vancomycin-responsive regressive-onset autism in 2000 with relapse after discontinuation and an ecological signature centered on absent Peptostreptococcus species and anaerobic cocci. Qiao et al. later demonstrated a host-susceptibility-plus-activation-event architecture in Shank3 haploinsufficient mice, where LPS challenge produced behavioral regression through inflammatory microglial and synaptic mechanisms. V1.5 integrated six pediatric and developmental mycobiome sources that sharpened the fungal layer: ASD-associated bimodal Candida distribution (Nirmalkar, Bimodal Candida in ASD) , sibling-controlled C. albicans enrichment, antibiotic-linked fungal dysbiosis in neurodevelopmental disorder, vancomycin-associated fungal expansion in an ASD microbiota-transfer context, and C. albicans-mediated barrier disruption in ADHD. V2 adds MRI stratification and longitudinal imaging literature showing ASD-specific multiparameter signatures, individual-deviation structure by language-delay history, two-cluster connectivity stratification, stable childhood-to-adolescence phenotype-morphology relationships, and stronger age-stage than gender-stage classification signal. V3 adds two further layers. Reed et al. demonstrated that the fever effect, the temporary improvement of autism symptoms during febrile inflammation, is mediated in mouse models by IL-17a acting on cortical circuits; the framework reads this as transient governance disengagement, with fever and IL-17a both operating as antifungal pressures that the existing design does not separate from the organism it suppresses. Ng et al. provided prospective birth-cohort evidence that epigenetic state at birth shapes infant gut microbiome assembly, with specific methylation-microbe combinations correlating with ASD and ADHD signs at thirty-six months, supplying upstream support for the Trajectory B mechanism while leaving the fungal layer unmeasured. The framework reads these as clinical, cohort, imaging, and mechanistic support for the two-trajectory model. The claim is not that autism has one cause, that vaccines cause autism, or that antifungal therapy is a universal treatment. The claim is narrower and testable: autism heterogeneity should stratify by developmental timing, regression history, mycobiome burden, gastrointestinal severity, antibiotic exposure timing, barrier state, imaging subgroup, and response distribution. Measure the organism layer and the developmental architecture together, or keep calling architecture mystery.