Dynamic Existence Threshold (DET)

The Dynamic Existence Threshold: Organizational Consciousness across Complex Systems

Nathan M. Thornhill — Independent Researcher, Fort Wayne, Indiana, USA

Overview

Analysis code, precomputed results, and publication figures for the Dynamic Existence Threshold (DET) framework. The DET introduces a two-dimensional coordinate system (Integration × Differentiation) that characterizes critical transitions across financial markets, space weather, and human EEG without parameter tuning between domains.

Key result: The coupling metric I = R×S achieves AUC 0.909 for discriminating wakefulness from deep sleep in 50-subject EEG, while simple summation of spectral power anti-predicts (AUC 0.416) — confirming the framework captures organizational structure rather than magnitude. Out-of-sample temporal validation confirms no overfitting (all metrics hold on held-out test periods). Component analysis reveals that total organizational information (Phi = I + D) is approximately conserved during critical transitions (within 1-14% across all domains), while components undergo inverse exchange — integration surges while differentiation drops in cascades, and the reverse during consciousness transitions.

Repository Structure

├── det-chaos.md          # Paper (Markdown, source of truth)
├── det-chaos.tex         # Paper (LaTeX, for journal submission)
├── compute/                                # Analysis pipelines
│   ├── phase7b_tests.py                    #   Financial + space weather (Tests 1-4)
│   ├── phase7b_eeg_50subj.py               #   EEG 50-subject analysis (Test 5)
│   ├── phase7b_eeg_test.py                 #   EEG 5-subject pilot
│   ├── revisions.py                        #   FDR, negative controls, L4 ablation, EWS baseline
│   ├── out_of_sample_validation.py         #   Temporal split OOS validation
│   ├── null_model_id_v2.py                 #   Null model for I-D classification
│   ├── delta_beta_headtohead.py            #   Delta/beta ratio vs R×S comparison
│   ├── eeg_loader_direct.py                #   EEG data loader (direct bandpass)
│   ├── layer_sensitivity_finspace.py       #   Layer count sensitivity (financial + space)
│   ├── layer_sensitivity_eeg.py            #   Layer count sensitivity (EEG)
│   ├── bootstrap_and_sweep.py              #   Bootstrap confidence intervals
│   ├── id_categories_test.py               #   I-D directional dynamics
│   ├── phi_conservation_test.py            #   Phi conservation analysis (Test 6)
│   ├── generate_paper_figures.py           #   Publication figures (300 DPI)
│   └── layer_sensitivity_combined_plot.py  #   Figure 6
├── results/                                # Precomputed results (JSON)
│   ├── phase7b_results.json                #   Financial + space weather
│   ├── phase7b_eeg_50subj.json             #   EEG 50-subject
│   ├── phase7b_eeg_results.json            #   EEG 5-subject pilot
│   ├── revisions.json                      #   FDR, negative controls, ablation, EWS
│   ├── out_of_sample_validation.json       #   Temporal split OOS results
│   ├── null_model_v2_results.json          #   Null model results
│   ├── delta_beta_results.json             #   Delta/beta comparison
│   ├── layer_sensitivity_finspace.json     #   Layer sensitivity (financial + space)
│   ├── layer_sensitivity_eeg.json          #   Layer sensitivity (EEG)
│   ├── bootstrap_results.json              #   Bootstrap CIs
│   ├── id_categories_results.json          #   Directional dynamics
│   └── phi_conservation_results.json       #   Phi conservation (Test 6)
└── figures/                                # Publication figures (300 DPI)
    ├── fig1_quadrant_model.png
    ├── fig2_early_warning.png
    ├── fig3_classifier_comparison.png
    ├── fig4_eeg_stages.png
    ├── fig5_residence.png
    └── fig6_layer_sensitivity.png

Data Sources

Domain	Records	Source
Financial markets	6,785 trading days	yfinance
Space weather	9,802 days	NASA OMNI2
EEG	50 subjects, 136,394 epochs	Sleep-EDF, PhysioNet

Setup

The compute scripts import modules from separate project directories (phi-financial-test, phi-space-test, phi-neural-test). Set these environment variables before running:

# Required for financial + space weather scripts
export PHI_FINANCIAL_DIR=./external/phi-financial-test
export PHI_SPACE_DIR=./external/phi-space-test

# Required for EEG scripts (layer_sensitivity_eeg.py, phase7b_eeg_test.py)
export PHI_NEURAL_DIR=./external/phi-neural-test

# Required for EEG scripts that load Sleep-EDF data directly
# (eeg_loader_direct.py, phase7b_eeg_50subj.py)
export SLEEP_EDF_DIR=./external/sleep-edf

If not set, scripts fall back to default paths (./external/) and exit with an informative error if those defaults do not exist.

Note: The phi-*-test directories contain shared data loading and metric computation functions used across related projects. These modules are not required to inspect precomputed results in results/. For full reproducibility, contact the author at research@nathanthornhill.com.

Reproducing

pip install -r requirements.txt

# Financial + space weather (Tests 1-4)
python compute/phase7b_tests.py

# EEG (requires ~8GB RAM, ~30 min)
python compute/phase7b_eeg_50subj.py

# Robustness checks
python compute/revisions.py

# Out-of-sample validation
python compute/out_of_sample_validation.py

# Null model + delta/beta comparison
python compute/null_model_id_v2.py
python compute/delta_beta_headtohead.py

# Layer sensitivity
python compute/layer_sensitivity_finspace.py
python compute/layer_sensitivity_eeg.py

# Phi conservation analysis (Test 6, requires ~8GB RAM)
python compute/phi_conservation_test.py

# Figures
python compute/generate_paper_figures.py
python compute/layer_sensitivity_combined_plot.py

Precomputed results are in results/ for verification without re-running.

Prior Work

1. The Existence Threshold (Thornhill, 2026a)
2. Information Loss at Dimensional Boundaries (Thornhill, 2026b)
3. The Dimensional Loss Theorem (Thornhill, 2026c)

License

This repository is licensed under CC BY 4.0. Reuse permitted with attribution — please cite the paper.

Citation

Thornhill, N. M. (2026). The Dynamic Existence Threshold: Organizational Consciousness across Complex Systems. Submitted to Chaos (AIP).