Disease Phases PROVEN
Phase 1 · Pre-clinical (≈20yr window)
Silent Aβ accumulation. PET amyloid+, cognitively normal. CSF Aβ42↓ is first biomarker change. Intervention window.
Jack et al. 2010 NIA-AA criteria · Bateman et al. 2012 NEJM (DIAN cohort: 25yr pre-symptom)
Phase 2 · MCI (3–5yr)
Tau spreading Braak I–IV. Mild episodic memory loss. CDR=0.5. PET tau+, hippocampal atrophy begins.
Braak & Braak 1991 Acta Neuropathol · Albert et al. 2011 NIA-AA MCI criteria
Phase 3 · Early AD
Hippocampal + entorhinal atrophy. ADL partially intact. EEG γ-sync measurable loss. CDR=1.
McKhann et al. 2011 NIA-AA dementia criteria
Phase 4 · Moderate AD
Cortical spread Braak V. Language + executive function fail. Cholinergic + glutamatergic depletion. CDR=2.
Braak staging (tau) · Thal phases (amyloid) well-established neuropathology
Phase 5 · Severe AD
Global atrophy Braak VI. ADL lost. Palliative focus. CDR=3.
Staging is descriptive. Progression rates highly variable — SORRY: no reliable individual prediction model.
Core Mechanism
Amyloid cascade hypothesis (Hardy & Higgins 1992 Science): Aβ accumulation is upstream cause → triggers tau hyperphosphorylation → neurodegeneration. STRONG but not complete — amyloid alone insufficient (cognitively normal amyloid+ exists).
Tau propagation: Spreads trans-synaptically in prion-like fashion. Braak stage predicts cognitive state better than amyloid load. PROVEN
APOE4: 1 allele = 3× risk, 2 alleles = 10× risk. Impairs Aβ clearance + accelerates tau spread + disrupts lipid metabolism. PROVEN
Approved Disease-Modifying
Lecanemab (Leqembi) FDA 2023
Anti-Aβ protofibrils mAb. Slows decline 27% (CDR-SB). Early MCI/mild AD only. ARIA risk 21% (mostly asymptomatic).
van Dyck et al. 2023 NEJM (CLARITY-AD: n=1795, 18mo)
Donanemab (Kisunla) FDA 2024
Anti-Aβ plaque mAb. Slows decline 35% (iADRS) in early AD. Stops when amyloid cleared. ARIA risk similar.
Sims et al. 2023 JAMA (TRAILBLAZER-ALZ 2: n=1736)
Approved Symptomatic
ChEIs (Donepezil/Rivastigmine/Galantamine)
Cholinesterase inhibitors. Modest cognitive benefit. All stages. Not disease-modifying.
Cochrane meta-analysis 2018 · standard of care
Memantine (NMDA antagonist)
Moderate–severe AD. Reduces excitotoxicity. Modest ADL benefit.
Reisberg et al. 2003 NEJM
Pipeline / Open Fronts
Tau-targeting NO APPROVED (2026)
BIIB080 (ASO, tau mRNA knockdown) Phase 2. JNJ-63733657 (anti-p-tau mAb) Phase 2. No tau drug approved yet.
Mummery et al. 2023 Nat Med (BIIB080) · SORRY: mechanism-to-efficacy gap unclear
40Hz γ-Entrainment (GENUS) Phase 2/3
40Hz light+sound stimulation. Restores γ-sync, activates microglia-mediated Aβ clearance. Iaccarino 2016 (mice: 50% Aβ↓). Human Phase 2/3 ongoing (Cognito Therapeutics).
Iaccarino et al. 2016 Nature · Chan et al. 2024 (human EEG response confirmed)
GLP-1 / Metabolic Trial results 2026
Semaglutide EVOKE trial (n=1840). Insulin resistance drives Aβ accumulation. Results expected 2025–2026.
EVOKE trial NCT04777396 · de la Monte 2012 (T3D "type 3 diabetes" hypothesis)
APOE4 Correction Phase 1
APOE4→APOE3 gene editing via AAV. Phase 1 trials beginning. 10yr horizon minimum.
Huang et al. 2023 (APOE4 structure-function) · SORRY: delivery to brain at scale unsolved
Glymphatic Enhancement
Sleep-dependent Aβ clearance via AQP4 channels. Slow-wave sleep (deep sleep) = glymphatic peak. Interventions: sleep hygiene, slow-wave enhancement drugs.
Iliff et al. 2012 Sci Transl Med · Xie et al. 2013 Science (glymphatic discovery) · SORRY: human intervention data sparse
V12 · 2025-2026 Updates
Trontinemab (Brain Shuttle Ab) Phase 2 · 2025
Anti-Aβ antibody with transferrin receptor-mediated brain shuttle technology. ~10x better CNS delivery vs standard anti-amyloid mAbs. Phase 2 data 2025 shows rapid Aβ plaque clearance at lower doses. Could reduce ARIA risk. Roche pipeline.
Roche TRON Phase 2 · NEJM Evidence 2025
GLP-1 Semaglutide (EVOKE Trial) Results 2025-2026
EVOKE trial (n=1840) — semaglutide vs placebo in early AD. GLP-1R activation: neuroprotection, neuroinflammation reduction, Aβ clearance via insulin-signalling pathway. Results expected 2025-2026. Largest metabolic-AD trial ever.
EVOKE/EVOKE+ NCT04777396 · Luchsinger et al. metabolic hypothesis
▸ V12 Update · Day 91 · 2026-05-05
Treatment panel updated with latest 2025-2026 clinical data. γ₁ = 14.134725141734693
CSF Biomarkers PROVEN
| Aβ42↓ | Sequestered in plaques. First to change ~20yr pre-symptom. Sensitivity 85–95%. Blennow et al. 2010 |
| p-tau181↑ | Phosphorylated tau at T181. Tangle marker, spreads Braak I→VI. FDA-cleared. Hampel et al. 2020 |
| t-tau↑ | Total tau = neurodegeneration (non-specific). Combined with Aβ42 = AT(N) framework. |
| Aβ42/40 ratio | More stable than Aβ42 alone. Better threshold. Hansson et al. 2018 JAMA Neurol |
Blood Biomarkers EMERGING
| p-tau217 STRONG | Blood test. 90% accuracy vs CSF/PET. AUC 0.96 in TRIAD cohort. Game-changer for screening. Palmqvist et al. 2020 JAMA |
| p-tau181 blood | Good specificity for AD vs other dementia. AUC 0.89. Janelidze et al. 2020 Nat Med |
| GFAP↑ | Astrocyte activation. Blood GFAP precedes symptom onset. Oeckl et al. 2022 |
| NfL↑ | Neurofilament light = axonal damage. Non-specific, elevated in many neurodegen. conditions. |
| Aβ42/40 plasma | SORRY: plasma Aβ ratios highly variable, matrix effects, not yet clinical standard. |
Imaging PROVEN
| PET amyloid | Florbetapir/Florbetaben. Visual + quantitative. FDA-cleared 2012. Gold standard for amyloid load. |
| PET tau | Flortaucipir (AV-1451). Braak staging in vivo. Predicts progression rate better than amyloid PET. |
| MRI atrophy | Hippocampal volume. Entorhinal cortex thinning. Free tool: FreeSurfer. Standard in trials. |
| FDG-PET | Metabolic decline. Temporoparietal hypometabolism = AD signature. SORRY: differential diagnosis with LBD difficult |
Genetic Risk
| APOE4 PROVEN | 1 copy: 3× risk. 2 copies: 10× risk. ~25% of population carries ≥1 copy. Most important genetic risk factor. |
| APP/PSEN1/2 | Autosomal dominant FAD. <1% of all AD. 100% penetrance. Onset 40s–50s. Bateman 2012 NEJM DIAN |
| TREM2 R47H | Microglia AD risk variant. 3× risk. Impairs microglial Aβ clearance. Guerreiro et al. 2013 NEJM |
| CLU, BIN1, CR1 | GWAS hits. Small effect sizes (1.1–1.3× risk each). Mechanisms still being worked out. |
EEG Signatures in AD
| θ power ↑ (4–8Hz) | Hippocampal overload. Earliest EEG signal, appears MCI stage. Correlates with episodic memory loss. Jeong 2004 Clin Neurophysiol |
| α power ↓ (8–12Hz) | Posterior parietal + occipital. Correlates strongly with MMSE score. Slowing of peak α frequency. Babiloni et al. 2016 |
| β power ↓ (13–30Hz) | Frontal reduction in moderate AD. Executive function correlate. |
| γ sync ↓ (30–100Hz) | Network integration fails. GABAergic interneuron loss. The GENUS target. Iaccarino 2016 |
| EEG slowing | Overall dominant frequency shifts from α toward θ. Simple, reproducible, cheap biomarker. SORRY: specificity vs other dementias ~70% |
Thinkbeat Parallel LABR-THINKBEAT-001
| θ excess | ≡ SLOW_MONOTONE: δ decreasing slowly, loop working but too much load |
| α loss | ≡ OSCILLATING: integration breaking down, routing unstable |
| γ sync loss | ≡ CHAOTIC: network collapses, quarantine signal |
| GENUS 40Hz | ≡ FAST_MONOTONE restoration: entrainment → microglia clear Aβ → convergence restored |
| Pre-clinical | ≡ SLOW_MONOTONE: degradation underway, δ detectable 20yr before CHAOTIC |
NP-ALZ-008: EEG biomarker → thinkbeat bridge. The 4-state shape grammar runs in the brain 20yr before symptoms. SORRY: this mapping is structural/metaphorical, not quantitatively validated.
GENUS Protocol Phase 2/3
40Hz flickering LED + 40Hz auditory click. 1hr/day. Entrains cortical γ oscillation. In mice (Iaccarino 2016): 50% Aβ reduction in visual cortex after 1 week. Mechanism: γ drives interneuron-mediated microglial activation → phagocytosis of Aβ. Human data: EEG entrainment confirmed (Chan 2024), cognitive outcomes Phase 2/3 ongoing. SORRY: mouse-to-human translation of Aβ reduction not yet shown in humans.
Novel Patterns (9)
PROVENNP-ALZ-001 · Tau trans-synaptic seeding = prion-like propagation · Braak stage predicts cognition better than amyloid load
PROVENNP-ALZ-002 · APOE4 dual role: impairs Aβ clearance AND accelerates tau spread simultaneously
PROVENNP-ALZ-003 · Blood p-tau217: 90% accuracy vs CSF gold standard (AUC 0.96 TRIAD) — democratises screening
PROVENNP-ALZ-004 · 20-year pre-clinical window: Aβ accumulates silently while cognition is normal · intervention window exists
EMERGINGNP-ALZ-005 · 40Hz γ-entrainment (GENUS): interneuron-driven microglial Aβ clearance · mice confirmed · human outcomes pending
EMERGINGNP-ALZ-006 · GLP-1 / T3D hypothesis: insulin resistance in brain → impaired Aβ clearance → AD as metabolic disease
OPENNP-ALZ-007 · Glymphatic hypothesis: sleep-dependent CSF flow clears Aβ via AQP4 channels · sleep = nightly brain wash
OPENNP-ALZ-008 · EEG oscillation → thinkbeat bridge: θ/α/γ degradation maps to SLOW_MONOTONE/OSCILLATING/CHAOTIC · 4-state grammar shared across scales
SORRYNP-ALZ-009 · TDP-43 co-pathology (LATE): ~40% of AD cases have TDP-43 inclusions at autopsy · mechanism and treatment implications unclear · diagnostic criteria immature
Open Sorries (4)
SORRY-1: Amyloid cascade alone insufficient — cognitively normal amyloid+ individuals exist. Why some tolerate amyloid load is unknown.
SORRY-2: Individual progression rate prediction model does not exist. Staging is descriptive only.
SORRY-3: GENUS mouse→human Aβ reduction not replicated. EEG entrainment confirmed; clearance mechanism unproven in humans.
SORRY-4: APOE4 gene correction delivery to brain at scale — AAV vectors don't cross BBB efficiently at therapeutic doses.