When previously acquired speech, motor, social, or adaptive skills are lost, the testing logic changes. This page serves as the dedicated regression branch within the testing suite.
See framework disclaimer belowRegression is treated here as a distinct biological state. The concern is not only that symptoms are present, but that previously maintained neural function has become unstable. That shifts testing from broad exploration toward rapid confirmation of the highest-risk interacting drivers.
Regression is not just another symptom cluster. It suggests a convergence state in which the learning-maintenance machinery of the system is under simultaneous stress. In that setting, the testing question changes from “which branch fits best?” to “which converging drivers are actively destabilizing function right now?”
When a child loses a skill they previously had — stops talking, loses words, stops making eye contact, loses motor milestones — the cascade has reached a level of severity where the CREB/BDNF learning pathway can no longer maintain previously encoded skills. This is not a separate condition from the upstream channels; it is their downstream consequence. The cascade did not cause the regression: it removed the biological machinery needed to sustain learned behavior.
Regression is the point at which the cascade has crossed the biological threshold at which SST-14 interneuron output can no longer be maintained. Both the IDO1 excitotoxic arm (quinolinic acid depleting mitochondrial ATP in SST-14 interneurons) and the NF-κB transcriptional suppression arm (cytokines suppressing CREB, reducing SST-14 gene expression) are driving simultaneous convergent silencing. Testing must identify which active drivers are converging rather than treating regression as an isolated symptom list.
A third convergence thread is now directly measured in human ASD cohorts: IDO1 activation produces two parallel downstream consequences that both contribute to regression. The first — quinolinic acid driving NMDA excitotoxicity — destabilizes the electrical activity required to maintain encoded skills. The second — serotonin depletion impairing 5-HT2 signaling in the hypothalamic paraventricular nucleus (PVN) — suppresses oxytocin release and maps to the social communication losses that families report most acutely. Launay et al. (2023, Translational Psychiatry, n=271) directly measured NAD⁺ deficit correlating with plasma oxytocin suppression in ASD individuals, establishing the hypothalamic arm as a documented human consequence. In regression, both IDO1 arms are active simultaneously alongside the NF-κB transcriptional and CD26/adenosine suppression — NMN/NR addresses the excitotoxic arm and the hypothalamic arm together.
This is the first-pass battery when regression is present. It does not replace clinical judgment, but it changes the order of what is most urgent to clarify. The goal is to map the two suppressive arms of the cascade while also checking for gut, immune, and mitochondrial triggers that may be actively sustaining the loss of function.
| Test | Why it moves up |
|---|---|
| Serum BDNF baseline | Urgent downstream baseline. Low BDNF supports suppression of the CREB-dependent learning-maintenance pathway and gives a tracking marker for recovery over time. |
| OAT with kynurenine and TCA attention | Checks both IDO1 / quinolinic pressure and mitochondrial stress in one high-yield starting test. |
| Kynurenine / Tryptophan ratio | Confirms whether tryptophan is being actively diverted through IDO1. |
| Salivary cortisol + IGF-1 | Maps functional SST / HPA load and helps show whether the CREB pathway is being suppressed from the stress-signaling side. |
| hsCRP + cytokines | Clarifies whether inflammatory signaling is actively sustaining the destabilized state. |
| Zonulin + LBP | Moves up immediately when regression followed a febrile illness, vaccine event, GI illness, stool change, food refusal, or other gut-linked worsening. |
| Lactate:pyruvate | Moves up especially when motor regression or low-energy crash patterns are part of the picture. |
Once regression is identified, the question is not only which markers are abnormal, but what those abnormalities mean together. This section translates the fast-lane battery into a decision framework that clarifies what is actively destabilizing function.
Protocol implication: regression supports a multi-arm suppression state rather than a single isolated problem. The IDO1 node contributes two independent threads — excitotoxic pressure from quinolinic acid and oxytocin suppression from serotonin depletion — both addressed by NMN/NR. The next steps should be sequenced around the confirmed active channels because single-layer approaches are less likely to stabilize function when the system is under converging pressure.
The most common question families ask about regressive autism is: why now? Why did a child develop apparently normally for 18–24 months and then regress? The threshold dynamics model answers this question mechanistically.
Compensatory mechanisms can maintain adequate SST-14 interneuron coordinating output while upstream cascade burden accumulates below the clinical threshold. These compensatory mechanisms include:
A second biological challenge — a febrile illness producing acute IDO1 activation, a streptococcal infection reintroducing streptokinase CD26 blockade, an antibiotic course disrupting gut barrier integrity and elevating LPS translocation, or a dietary event increasing opioid peptide load — tips the balance at the moment the remaining compensatory capacity is insufficient.
The regression is the moment the threshold was crossed, not the moment pathology began. The trigger event — the infection, the vaccine reaction, the GI event — is not the cause. It is the final increment in a cascade that had been accumulating for months to years. This is why removing the trigger does not produce recovery: the cascade continues operating independently of the event that made it clinically visible.
The clinical implication is that regression test results should be interpreted against this background. The acute trigger event produced acute biomarker changes — elevated hsCRP, spiking cytokines, elevated zonulin if a gut event was involved. The underlying chronic cascade produced the baseline biomarker pattern that was there before the trigger: K:T ratio already elevated, homocysteine already elevated, glutathione already depleted. Identifying both layers — the acute triggering event and the pre-existing chronic cascade — is the complete picture.
Pediatric Acute-onset Neuropsychiatric Syndrome (PANS) and its streptococcal-associated variant PANDAS (Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal infections) are not mechanistically separate from the immune-derived autism cascade. They represent the accelerated acute expression of the same autoimmune mechanism through a more acute and identifiable triggering event.
In PANDAS, streptococcal molecular mimicry drives rapid production of autoantibodies against basal ganglia neurons and SST-14 interneuron surface proteins — the same autoantibody mechanism that constitutes State 1 SST-14 surface receptor jamming in the chronic cascade. The difference is temporal: in PANDAS the autoantibody production is acute, rapid, and clinically dramatic; in immune-derived autism the same mechanism has accumulated more slowly.
The Fourie & Armstrong 2024 IMIG case report included both PANS and ASD patients. The PANS group showed a stronger mean response (+4.4/5 vs +2.9/5 for the ASD group) — which the authors interpreted as a timing effect. In PANS, the autoimmune trigger is recent and the underlying neural machinery is more intact. In ASD, neuroinflammation typically began in early development and structural changes may have accumulated. IMIG's mechanism of autoantibody clearance and cytokine suppression is the same in both conditions.
When a child presents with acute-onset neuropsychiatric symptoms following an infection, add to the standard regression battery:
In PANS/PANDAS, the urgency of IMIG evaluation is higher than in the chronic ASD presentation — the autoimmune mechanism is recent, the neural substrate is more intact, and the window for intervention before structural changes accumulate is narrower. The same State 1 intervention logic applies: IMIG/IVIG targets the autoantibody clearance and cytokine reduction that is the rate-limiting intervention in this state.
This page is the dedicated regression branch. The rest of the testing suite still supports it: