Filed results, pre-registrations, the evidence base, methodology, and open infrastructure. Everything a reviewer needs to audit the work.
Walk-forward analysis of ionospheric Total Electron Content (TEC) anomalies preceding M6.0+ seismic events across all five Sentinel fault zones, using 25 years of IONEX Global Ionosphere Map data (2001 to 2025). 45,651 feature rows processed via INSPIRE methodology (Pulinets et al., 2021). Pre-specified threshold: AUROC ≥ 0.52 in ≥ 2 of 5 systems. Results: Cascadia 0.5618 (CONFIRMED_UNSTABLE, CI: [0.44, 0.67]), Japan Trench 0.5010 (NULL), Central Chile 0.4758 (NULL), North Anatolian 0.4823 (NULL), Sumatra-Andaman 0.4846 (NULL). 1 of 5 systems met threshold; 2 required. H3 NULL. GitHub commit 0d4f77e.
Classified as a resolution-limited null. IONEX GIMs operate at 2.5°×5° spatial resolution with 2-hour temporal updates, too coarse to resolve the fault-zone-proximal ionospheric signatures the LAIC chain produces at 10 to 100 km scales. Individual folds produced high AUROC values (Sumatra-Andaman Fold 1: 0.8213, Japan Trench Fold 3: 0.7977), consistent with episodic signal too localized to survive IONEX averaging. This result establishes the IONEX measurement floor and motivates high-fidelity ionospheric data in the next phase.
The GPS deformation pipeline processed 2,746 TENV3 station files (48,858 daily rows). Overall result: null under the pre-specified threshold. The North Anatolian zone returned AUROC 0.714, the only zone above the 0.65 threshold. With five zones tested, a single-zone exceedance is classified as exploratory. It informs the single-region design of the next phase and is treated as a candidate for prospective testing.
Amendment 3 stratified analysis of the human anomalous cognition intake pipeline returned Bonferroni-significant pre-event temporal signals: flood at day −1, landslide at days −5 and −1, across five catalogued hazard types. Results committed to GitHub and filed to OSF under the pre-specified protocol. Classified as exploratory pending prospective replication.
Walk-forward convergence analysis returned null under the pre-specified threshold. Geomagnetic stream: Kp, Ap, F10.7, and Dst, 9,131 daily rows across five fault zones. A re-run is planned with high-fidelity ionospheric data.
Primary pre-registration for the multi-stream study spanning geophysical, environmental, and human anomalous cognition domains across six hazard types: earthquake, flood, volcanic, avalanche, landslide, and tsunami. Amendments are filed as the study evolves, including walk-forward methodology (1), expanded event catalog (2), stratified temporal analysis (3), and the H3 ionospheric result (11). The next-phase benchmark design is described in the Sentinel technical plan.
Pre-registered blind study conducted with the IRVA Research Unit, testing whether Remote Viewing, Map Dowsing, and Dream Analysis produce spatially convergent signals over an undisclosed search target. Three-participant convergence threshold, Bonferroni correction, and submit-first, disclose-second sequencing. Research Coordinator: J. Tulloss. Project Lead: A. Weatherston (Augur Management).
The record of anomalous human cognition spans cultures and centuries, and decades of controlled laboratory research. The SRI remote viewing program, the PEAR laboratory at Princeton, the Ganzfeld meta-analyses, and the Global Consciousness Project produced results that current theoretical frameworks cannot fully explain.
Their integration into mainstream science stalled for lack of infrastructure to test them at scale, under the same pre-registration standards as any other empirical domain. Global sensor networks, cloud computing, and AI now make that possible.
Effect sizes across these programs are small to medium (d ≈ 0.2 to 0.4). Statistical significance here reflects the scale of replication, independent of effect magnitude.
Utts (1995) peer-reviewed assessment of 26,000+ SRI/SAIC trials concluded: "Using the standards applied to any other area of science, psychic functioning has been well established." Results far beyond chance; methodology soundly vindicated.
Meta-analysis across 28 autoganzfeld studies: z = 6.6, p = 2.8 × 10⁻¹⁰. Replication has varied: Milton & Wiseman (1999) found weaker effects in a non-autoganzfeld subset, while Storm et al. (2010) found continued significance across a broader meta-analysis.
Princeton's engineering anomalies laboratory documented micro-PK and remote perception effects across nearly three decades of peer-reviewed experimentation.
Across 500 formally defined events over 20 years, odds against chance for the composite outcome exceeded one trillion to one. About 70 globally distributed REG devices show correlated deviation coincident with major world events. Mean effect per event is small (~0.3σ) and can only be assessed cumulatively.
The work proceeds without first resolving the hard problem of consciousness. Whether anomalous cognition arises from quantum coherence, non-local field effects, or something else entirely, the methodological question stays the same: does it produce detectable, reproducible, statistically characterizable signal, and can that signal translate into practical benefit?
Science can answer that question with pre-registered hypotheses, open infrastructure, null-result transparency, and convergence testing across independent data modalities. Each evidence tier is scored on its own standard and never pooled into a single score.
Pipeline code, analysis scripts, and documentation: HAC intake, GPS deformation, geomagnetic ingestion, and convergence scripts, with full commit history.
synexisproject-bit →GCP project synexis-project-sentinel. Four datasets: hac_intake, sentinel_features, sentinel_groundtruth, sentinel_analysis, with documented schema and Cloud Run ingestion.
Schema docs →Pre-registrations, amendments, and filed results are independently timestamped on the Open Science Framework, giving any reviewer verifiable hypothesis lock dates.
OSF Registry →