A governed N-of-1 longitudinal research archive for studying how training, recovery, physiology, behavior, and context interact over time.
Current published release: v1.1.0 — 2026-09-14
Version DOI: 10.5281/zenodo.22759132
All-versions DOI: 10.5281/zenodo.20815611
Frozen release commit: 92126e1cc882c3822d9e03b30b11cfc1d30b4fbb
The Daniel Longitudinal Study follows one human system through repeated measurement, machine-readable data, contemporaneous observation, source-preserved testing, registered predictions, retrospective interpretation, correction history, and explicit archive governance.
The aim is not to prove a universal intervention. It is to preserve an inspectable record of what was measured, what was predicted, what happened, what changed, and what was corrected.
Scope: single-subject, longitudinal, observational. This is not a clinical trial, coaching product, or generalized intervention protocol.
Choose the entry point that matches what you want to do. These are alternatives, not a required reading sequence.
| Goal | Entry point |
|---|---|
| Understand the project in about five minutes | docs/START_HERE.md |
| See what the archive currently says | LATEST.md |
| Inspect the public structured evidence | data/DATA_COVERAGE.md |
| Audit claims against evidence and provenance | docs/OBSERVER_QUICKSTART.md |
| Browse the complete repository map | INDEX.md |
Specialized material such as model-error records, source provenance, testing artifacts, verification, and release documentation remains linked in the relevant sections below rather than competing for first-contact attention.
Plain-language summary: Week 41 closed with seven aerobic and six Load Integration sessions, a further prescribed pull-up increase, and another exposure to the expanded hinge/grip workload. Ordinary activity and training remained available by report, while selected overnight recovery readings were less favorable. Week 42 now observes whether the additional workload remains manageable across further sessions; Phase 2D remains undeclared.
The formal archive labels below preserve the precise phase/state vocabulary used by the reports and validators.
| Domain | Current state |
|---|---|
| Phase | Phase 2 — Load Integration |
| Weekly operating posture | Continued LI progression / recovery-cost and dose-repeatability observation |
| Broader Phase 2 substate | Consolidation / lock-in observation |
| Active weekly window | 2026-W42 |
| Most recent closed window | 2026-W41 |
| Installed training architecture | B1 + Load Integration |
| Open model-error records | 047–050: four prospective November predictions; 041–046 closed |
| Record 043 | closed / not supported — overall_improvement_not_met / over |
| August snapshot | source artifacts archived, reconciled, structured, checksum-protected |
| Formal Phase 2D declaration | none |
Current-state anchors retained for repository validation:
Active weekly window:
2026-W42
Most recent closed window:
2026-W41
In formal terms, Week 41 documented conditional repeatability: pull-ups advanced to 5×5 on the progression day and the expanded hinge/grip prescription was repeated. Standard, reduced-load and rest boundaries were preserved. The pull-up increase is distinguished from W41's original repeatability-only question. Reported ease does not establish zero cumulative recovery cost or declare Phase 2D.
Term definitions: docs/CONCEPTS.md
Authoritative current detail: LATEST.md
Current public daily, sleep, and completed-training coverage extends through 2026-10-18. The bounded context-event index ends on 2026-10-14 with a single event for the further pull-up increase; it is not a complete daily-activity diary.
| Dataset | Unit | Coverage |
|---|---|---|
data/daily_biomarkers_v1.csv |
one row per represented day | 252 continuous rows through 2026-10-18 |
data/sleep_longitudinal_v1.csv |
one governed wake-date row | 252 continuous rows through 2026-10-18 |
data/training_blocks_v1.csv |
one row per completed session/block | 409 rows through 2026-10-18 |
data/context_events_v1.csv |
one bounded context event | 53 rows through 2026-10-14 |
Training is session-indexed: a represented day with zero completed training sessions does not require a synthetic training row.
The daily/training/event layers are curated public extracts derived primarily from governed private Daniel_Dataset_v1.x source states. They are not raw provider exports.
Schema and provenance:
schemas/machine-readable-layer-v1.mdDATA_DICTIONARY.mddata/source_provenance/data/DATA_QUALITY_NOTES.md
Source Artifacts
↓
Curated Structured Data
↓
Contemporaneous Context
↓
Retrospective Reports
↓
Phase / State Interpretation
↓
Registered Prediction Evaluation
↓
Model Correction
A higher-level interpretation does not overwrite the evidence beneath it.
The repository is designed to preserve disagreement as well as agreement: favorable observations can coexist with failed predictions; recovery telemetry can diverge from demonstrated function; corrections can repair source transcription without repairing the narrative.
When records conflict, interpretation generally prioritizes:
- verified primary source artifact
- direct provider/device export
- contemporaneous structured transcription
- contemporaneous collection note
- closed retrospective report
- later evidence-grounded synthesis
- unsupported memory
Silent alteration, outcome-driven rewriting, unsupported reconstruction, and retrospective prediction revision are prohibited.
Source-backed corrections must remain narrow, traceable, and reviewed for downstream effects.
Collection occurred across 2026-08-17 and 2026-08-18:
2026-08-17
TruDiagnostic → DEXA → VO₂ max
2026-08-18
Bod Pod
Preserved public artifacts under snapshots/2026-08/ include DEXA, VO₂, Bod Pod/COSMED, TruAge, Advanced TruAge, and TruHealth outputs. Seven protected August artifacts are registered in the folder SHA-256 manifest.
The contemporaneous repository record controls the canonical TruDiagnostic sample event: 2026-08-17 at 05:37 local. Provider-displayed administrative metadata remain preserved but do not overwrite that collection record. See data/source_provenance/2026-08-trudiagnostic-reconciliation.md.
Integrated snapshot values are represented in:
Recent protected records:
| Record | Domain | Outcome |
|---|---|---|
| 041 | recovery_capacity | supported |
| 042 | ambient_execution | not supported — continued adaptation |
| 043 | biological_translation | not supported — overall_improvement_not_met / over |
| 044 | protocol_governance | not supported — narrow snapshot-directed deviation |
| 045 | autonomic_reconvergence | supported |
| 046 | autonomic_unload_reload | failed_autonomic_recompression |
Record 043 remains fixed under its preregistered biological scoring boundary. Later favorable training or behavioral evidence does not reopen or rescue the outcome.
Four November 2026 predictions are now prospectively registered and open, not scored:
| Record | Forecast locked on 2026-10-19 | Type | Status |
|---|---|---|---|
| 047 | DunedinPACE = 0.76 | point | Open |
| 048 | SymphonyAge 40.0–44.0 years inclusive | range | Open |
| 049 | DEXA fat ≤36.0 lb and lean ≥183.0 lb | state (two conditions) | Open |
| 050 | Provider-measured relative VO₂max ≥52.0 mL/kg/min | state | Open |
The open records contain no November outcomes. Collection rules: methodology/2026-11-snapshot-collection-plan.md. Locked forecasts and scoring rules: methodology/open_prediction_evaluation_plan_047_050.md. The complete registry is data/model_error/model_error_gap_v1.csv. Numerical thresholds do not authorize training changes or a Phase 2D declaration.
Full adjudication: data/model_error/record_043_closure.md
UDI is an error/calibration framework, not a biological score. See docs/methodology/UDI_framework_v1.md.
This archive does not establish population-level causality, clinical efficacy, generalized intervention guidance, protocol safety for other people, or universal device/assay accuracy.
Known bounded data-quality restrictions remain visible rather than silently repaired. Examples include historical sleep-field reconciliation notes and DQ-011 for the recorded 2026-09-08 resting-heart-rate value pending source verification.
Some historical private source states also lack exact retained immutable files for SHA-256 registration. Those hashes remain explicitly unavailable rather than reconstructed.
See data/DATA_QUALITY_NOTES.md and methodology/anonymization.md.
Read-only validators:
python tools/validate_repository.py
python tools/validate_machine_readable.py
python tools/validate_weekly_rollover.py
python tools/validate_august_snapshot.py
python tools/validate_coherence.py
GitHub Actions runs repository validation on pushes to main and pull requests.
A validation PASS confirms implemented mechanical/governance checks. It does not establish biological causality, clinical validity, device accuracy, or correct scientific interpretation.
See VERIFICATION.md.
AI/LLM systems may assist with source-backed structuring, drafting, calculation, analysis, audit review, validator code, and formally governed prediction generation.
AI is not a source-evidence class and may not invent observations, fabricate provenance, override stronger source evidence, or rewrite registered predictions after outcome access.
Disclosure: docs/AI_ASSISTANCE.md
The externally published v1.1.0 release is fixed to:
GitHub tag: v1.1.0
Frozen release commit: 92126e1cc882c3822d9e03b30b11cfc1d30b4fbb
Release archive: daniel-longitudinal-study-v1.1.0-92126e1cc882.zip
Release archive SHA-256: f7bad6d466c28d85fc263128083fe37d9039cb116cff67fb1c8a35482d5ce0a9
Zenodo version DOI: 10.5281/zenodo.22759132
Zenodo all-versions DOI: 10.5281/zenodo.20815611
The GitHub v1.1.0 tag remains fixed to the Batch 5A frozen commit. Post-publication DOI reconciliation occurs only on later main; it does not move the tag or mutate the published release archive.
The Zenodo lineage contains an earlier immutable record labeled v1.1.0 with DOI 10.5281/zenodo.22759127. The finalized version-specific DOI for the frozen Batch 5A package is 10.5281/zenodo.22759132. The all-versions DOI 10.5281/zenodo.20815611 remains the stable cross-version entry point.
Release notes: docs/releases/v1.1.0.md
Publication reconciliation: docs/audits/2026-09-14-v1.1.0-publication-reconciliation.md
Versioning rules: VERSIONING.md
Citation metadata: CITATION.cff
License: LICENSE.md
- Current state →
LATEST.md - Start guide →
docs/START_HERE.md - Observer route →
docs/OBSERVER_QUICKSTART.md - Dataset overview →
DATASET_OVERVIEW.md - Full index →
INDEX.md - Data coverage →
data/DATA_COVERAGE.md - Data dictionary →
DATA_DICTIONARY.md - Private-source provenance →
data/source_provenance/ - Source exports →
data/source_exports/ - Weekly reports →
reports/ - Snapshots →
snapshots/ - Model error →
data/model_error/ - Governance →
GOVERNANCE.md - Methodology →
METHODOLOGY_AND_CONTROLS.md - Verification →
VERIFICATION.md - Audits →
docs/audits/ - Changelog →
CHANGELOG.md
The pre-v1.1.0 long-form README is preserved byte-for-byte at docs/archive/README_PRE_V1.1.0.txt.
Public longitudinal archive initiated 2026.