A research corner · zebra.sharan.fun

ZEBRA

Notes on the illnesses that get missed — connective tissue, dysautonomia, arachnoid webs, and the mechanisms proposed to bind them.

“Medical zebra”: the rare diagnosis behind the ordinary hoofbeats. This corner collects candidate mechanisms and grades them as it goes — established, proposed, or speculative. The grading is the point.

Update № 1 · October 28, 2026 · Mechanism, proposed

The Clearance Hypothesis: Glymphatics, Autophagy, and the Loop

A perspective paper from the Stanford Genome Technology Center (Nemat-Gorgani, Jensen & Davis, International Journal of Molecular Sciences, 21 October 2026 — MDPI 27(20):9351) proposes that a large share of the ME/CFS and Long COVID picture — fatigue, cognitive dysfunction, sleep disturbance, orthostatic intolerance — may be the downstream signature of a waste-clearance failure with two cooperating halves:

The glymphatic–meningeal lymphatic system flushes metabolic waste out of brain tissue: cerebrospinal fluid moves through the parenchyma (gated by aquaporin-4 channels on astrocyte endfeet, polarized toward the vessels) and drains via meningeal lymphatics to the cervical nodes. It works mainly during sleep. → Autophagy is the intracellular half: damaged proteins and organelles — including the misfolded, amyloid-like aggregates reported in these illnesses — are engulfed and digested by lysosomes, subtype by subtype (mitophagy, aggrephagy, ER-phagy…), not as one undifferentiated “recycling.” → When either half lags, aggregates accumulate; astrocytes clog, AQP4 depolarizes, clearance falls further; neuroinflammation and autonomic dysfunction follow — and both degrade sleep, which degrades clearance. A loop, not a list.

Why this corner cares: the hypothesis is mechanical in the literal sense. It is about fluid moving through tissue and machinery clearing cargo — the same physical register as a cerebrospinal-fluid obstruction or a connective tissue that gives too much. And it binds, in one frame, threads this corner usually tracks separately: dysautonomia, post-viral illness, sleep architecture, and the long clinical habit of reading a multisystem failure as separate complaints in separate clinics.

Epistemic flag, kept loud: this is a mechanistic perspective — a synthesis of existing evidence into a proposed model, from authors who say plainly that patients are heterogeneous and the model needs validation. It demonstrates no causation and establishes no treatment. Grade: proposed, testable, unproven. The simulator below is a toy of the proposal — a way to see the loop's shape — not a model of any person's body, and nothing here is medical advice.

The simulator — run the loop

One variable matters: waste load (misfolded protein and damaged-organelle burden, arbitrary units). Production adds to it. Two clearance systems remove it: glymphatic flow (strong in sleep, scaled by AQP4 polarization) and autophagy (steady, saturating). Cross the coupling switch and accumulated waste feeds back — degrading effective sleep and clearance — which is the paper's central claim, made playable: past a threshold, the system stops recovering between nights and tips into the high-waste regime on its own.

Sim day
0.0
Waste load
0.20
Clearance now
—
Regime
clearing

Walkthrough — see the tip

  1. Start at “typical.” Waste oscillates with the day/night cycle — cleared at night, creeping by day — and holds a low line. The system copes.
  2. Drag production up a little (0.55 → 0.75). The line rises but still recovers each night — the healthy range is buffered, and the coupling has not engaged. Push on toward 0.9 and you cross into the zone where it has.
  3. Untick the coupling and push production to 1.2: waste climbs to a high plateau and stays linear — bad, but stable and reversible. Drop production back and it drains.
  4. Re-tick the coupling at the same settings. Now the trace runs away: waste degrades sleep, degraded sleep cuts clearance, and the loop locks into the high regime. Drop production back to 0.55 — it does not recover on its own. That hysteresis is the whole hypothesis in one gesture: the way in is not the way out.
  5. Find an exit: with coupling on, recovery needs a combined push — sleep quality near maximum and autophagy raised and production low, held for many sim-days. Single-lever fixes fail inside the loop. (In the paper's terms: why multimodal interventions are even on the table — proposed, not established.)

The model, exactly as run

dW/dt = P − [ ( G(t) + A ) · W/(W + 0.25) + 0.05·W ] G(t) = 2.2 · AQP4_eff · Sleep_eff · night(t) glymphatic, sleep-gated A = 0.9 · Autophagy_eff autophagy, saturating night(t) = 1 at night (8h), 0.12 by day the 0.05·W term is residual spillover clearance (minor pathways) with coupling ON, past a buffered range (u = max(0, W − 0.45)): Sleep_eff = SleepQuality / (1 + 2.2·u) AQP4_eff = AQP4 · (1 − 0.6 · u/(u + 0.8)) Autophagy_eff = Autophagy · (1 − 0.6 · u/(u + 0.8)) Units are arbitrary; constants are chosen so the three regimes (clearing / plateau / tipped) are all reachable, and probe-verified against the walkthrough below. W is clamped at 8 (off-chart = fully tipped). Euler integration, dt = 1/240 day. A cartoon of a proposed mechanism — the shape is the claim, not the numbers.

Static equivalent, for readers who skip simulators: three regimes. Clearing — production below nightly clearance; waste oscillates low. Plateau — production above clearance without feedback; waste parks high but tracks production reversibly. Tipped — with feedback, past a threshold the system's resting state itself moves: high waste persists even after the original production is restored, because waste is now suppressing the machinery that would remove it. The paper's question for the field: whether some patients live in the third regime, and whether sleep architecture, autonomic measures, proteostasis markers, and clearance imaging covary the way the loop predicts.